<%BANNER%>

Analysis of in vitro and in vivo function of total knee replacements using dynamic contact models

University of Florida Institutional Repository
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20110403_AAAAHY INGEST_TIME 2011-04-04T01:11:11Z PACKAGE UFE0013600_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 30235 DFID F20110403_AACIVK ORIGIN DEPOSITOR PATH zhao_d_Page_36.QC.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
c39b805de3ff6863fcdbd41e66a278e2
SHA-1
464b50b4f4a9caaafab53aa29e1c811351575da1
6223 F20110403_AACILQ zhao_d_Page_40.pro
f4fca5e81a78e3ee529ee30b9834dab5
704deb7c68d1debdb0a45e28935a666398722b9f
8423998 F20110403_AACIGS zhao_d_Page_64.tif
e82ef1e52a18ce5c19f406bbef395958
a6e02bd894ae7a0e97f9fcba29a21567178fd039
1051986 F20110403_AACIQN zhao_d_Page_14.jp2
e1f37fcc5b64f0170603276a2ec56880
264625a7efd988da763d6e0b4d609d7598230ab0
34635 F20110403_AACIVL zhao_d_Page_62.QC.jpg
412028738921837e834d548b6d3e441f
2653d454db1f55343c6cd1351211b90bc34c2f72
F20110403_AACIGT zhao_d_Page_65.tif
cde63795e2bd47c967b03421841e0584
a362ac63c64aacca7d49b4b598e27bcfdd6d496a
851237 F20110403_AACIQO zhao_d_Page_15.jp2
54f73f6695a0204d8677404949549148
6d0762b96f0b7d348927d8e14a27101cacdfe29b
8109 F20110403_AACIVM zhao_d_Page_63thm.jpg
026b47365ebdd6d60c250fa984ab30cb
2e2ee5fccbcdd49240bb38c4c08ea15180182556
45638 F20110403_AACILR zhao_d_Page_41.pro
7ac17164694361715e71cc93521c8f83
dfa817b9fdac6d1845aa50589891ecda566cb9a8
F20110403_AACIGU zhao_d_Page_66.tif
7cd63b0c209e1c982282cb148e5e3121
7e9a95cef94bc982b4397383c35b10db190244ee
516057 F20110403_AACIQP zhao_d_Page_16.jp2
27cd0b9da8e2f959200254a81f5815f8
370ac3bab09e33e26ebf0cb280857e1a5ca05076
29594 F20110403_AACIVN zhao_d_Page_73.QC.jpg
dd8773715283fd7cb41dc0df26305d53
03c282c76dbaa1247f6b161a839e25b433e7bcaa
49887 F20110403_AACILS zhao_d_Page_42.pro
d8ef47d79561d1dd21c39a6bf2aeb17a
4c40e6b6fe9048f0862f94d39c2f4d6b96bc68e4
F20110403_AACIGV zhao_d_Page_67.tif
5ce71843ef677e4ff4feb20dbab8daac
0283dff9e7287a8040e709641516416ae7f1dfbf
1051949 F20110403_AACIQQ zhao_d_Page_17.jp2
891174b6243c205f75a652e8a424e36c
ac0045123dcab6628bb28a56a6849f7b96b55b62
7971 F20110403_AACIVO zhao_d_Page_68thm.jpg
90038fb99b195725d41a595807eca0de
5d214be2c45c97b16d549b4174f26fde37697c6d
49541 F20110403_AACILT zhao_d_Page_44.pro
1974f27537f07c8536c97be1a2abec7f
18a994f5ab21a4b2127f2cc7a936a804e7fffffb
F20110403_AACIGW zhao_d_Page_68.tif
53add3ff8fbb4ce84869811dc45bbd06
5f3c0752bb2cb5a928227e2d9cb547845fab0145
1051922 F20110403_AACIQR zhao_d_Page_18.jp2
26b5cbe88cb6172e2a8cfaeab532d306
1cf076248024a5e1f95bf0cc557b6e8d470b7d84
34669 F20110403_AACIVP zhao_d_Page_19.QC.jpg
0e938e01ad1bbe1828089a3e5b4332c2
1a1558230a862bab87cd7a9e7df20bf61696ac8e
35739 F20110403_AACILU zhao_d_Page_45.pro
6c9901dc7c2c71fac60480820e875aa3
0fe89e8a4cde907f86a75e46b22c8784699042a5
F20110403_AACIGX zhao_d_Page_69.tif
5a16091da9741f5af9ea199b67403a90
4e8b06fc1cc0fd0cda99dab0afef3c3dd767ed9f
1051978 F20110403_AACIQS zhao_d_Page_19.jp2
75f73159c5a5d1fdf389bb1b274bad9c
d42f126ccf71f9c0fae80c105113a2118a4bce3c
8066 F20110403_AACIVQ zhao_d_Page_60thm.jpg
cd69820ae726c585a80f039e5506c3a5
ec5624fa09edc4555273160795779c1dce59bd55
50782 F20110403_AACILV zhao_d_Page_46.pro
059644bcee790129f36d6c0bd00a30ba
34a3706f53579604ff6fe25f09cd9b2de80784a1
32404 F20110403_AACIEA zhao_d_Page_56.QC.jpg
a0bedb6b1866d4fbbf5acd1ac8445072
80dfa34e5efc5be65368e433051166471aeb11dd
F20110403_AACIGY zhao_d_Page_71.tif
f723e334fdfb26bd432d95593b19defe
de7ac117e6da3acb39d92dfde46596ed04ae8a4d
1051981 F20110403_AACIQT zhao_d_Page_20.jp2
3722d2d9a830f4a7d4ac29fda3fae725
c9a0e1da39cad976d06b912d7db2973e5215e02d
27891 F20110403_AACIVR zhao_d_Page_67.QC.jpg
e5c52f6ca5a2eca8b423f36767f4f577
3188c2cc60ed7746511e1c02861dc94077667b68
43437 F20110403_AACILW zhao_d_Page_47.pro
77c167fe01eb3f91f8c36f30a6b483a1
456fb9a5c1b96bb88f18c86dcad21a6bd7dc16de
7626 F20110403_AACIEB zhao_d_Page_67thm.jpg
ca229238c2e1165d3de5f254aea4cdfa
5990910f07b1ef48b28b7b00768d3f7614bcef4c
F20110403_AACIGZ zhao_d_Page_72.tif
60019b814fb93b0e571e2233fac2ad82
a2867c2089b1f77e903b4df41be04f6c604de06b
625644 F20110403_AACIQU zhao_d_Page_21.jp2
7b9f78290acbbc35a2ca0af2010633a3
5e2eaae460d03afeccafd7a67708fb6984c872a3
30812 F20110403_AACIVS zhao_d_Page_50.QC.jpg
68968a21c5b4fe773eaa91831a70e494
43c395641f9ff7d37fd171977a05034838aad07b
52248 F20110403_AACILX zhao_d_Page_48.pro
aba4ee12dc10a4bdfca3400102277cce
5a2195f83e466abbf1dea8c151948742fe93503d
104165 F20110403_AACIEC zhao_d_Page_29.jpg
183cd805c3ecd92a4c45b40d9e2b55d2
637b7bf7779b61c2475f831ca00024ebb3761fea
1051958 F20110403_AACIQV zhao_d_Page_22.jp2
15a290f60ce3fe8e92360cb2e339f035
6e9f6bf8dade21dba7a87edce922a545ab1a3d74
1445 F20110403_AACIVT zhao_d_Page_03.QC.jpg
6bcb94e4906a7f9039e84726850e0c99
59618d86fc64d5f1c15d50a5bdf3259644f9e222
41886 F20110403_AACILY zhao_d_Page_49.pro
9867212953be6d566eb3d0ca34f2837e
a4f2e2313884f5fe77eb89830be194f9a237d9d3
32543 F20110403_AACIED zhao_d_Page_42.QC.jpg
c6253a1852d62fb9b06f8c71b7467194
b94635e1968e752f4a883db8a6be13cbe77617d9
1999 F20110403_AACIJA zhao_d_Page_46.txt
e1d4a1853b6a9e935dfb4e2ca98382fd
03647bfef1545e8fd3cc9542dc86ac3a8dd97051
6626 F20110403_AACIVU zhao_d_Page_10thm.jpg
7a3796d8dd77d47857e5ced925acb551
6dd120f1f8d4552dcc0c2599b0f791694c344c91
52260 F20110403_AACILZ zhao_d_Page_50.pro
3161fe06759bd7978cfe65057f629f7c
472a1f49e534395f6f44f579548d9a3bb7f06a31
49414 F20110403_AACIEE zhao_d_Page_68.pro
6511a437a29f5241c81e122d26f562c8
fa0381e134cc1de555e8675d89236b828d18ddf4
914290 F20110403_AACIQW zhao_d_Page_23.jp2
191bf3f88c88c06ee4ce623979ff334c
dcd00275d605217c28516d98f4836a9b9d96c505
1946 F20110403_AACIJB zhao_d_Page_47.txt
1061bd131069109c5f6c00dd7ad27bf2
723c3c4ebe701f860ef4b92dab771d1cacd49882
35530 F20110403_AACIVV zhao_d_Page_78.QC.jpg
728fdd13ea14730d6b9d8d18be45fbb5
e55f5808e8e055c7175736d93091d31b1b2aeb34
7539 F20110403_AACIEF zhao_d_Page_31thm.jpg
e19c05775d397c5390ea851fe8907a98
8304d24f73ba69dd16d9e4c0d63b5c681013078d
896796 F20110403_AACIQX zhao_d_Page_24.jp2
327ab007ce2340b4c5bde3cb8813d7fa
ddaf67501099cd3dde479ace1e8fcb3341e99472
2597 F20110403_AACIJC zhao_d_Page_48.txt
8057c64b857c77ad3fd8fc6997a4fb62
f155b9e716853d9907a895d8e84a955e5b271573
8185 F20110403_AACIVW zhao_d_Page_22thm.jpg
5dea70bad0ef1e2428bdcb12327c1f9e
5d9b545a278ab2d52f7477f3dde20aac0c8bdb43
96033 F20110403_AACIEG UFE0013600_00001.mets FULL
ebece952239e17ac19846f8fc5a586d6
da71229574961026d74cb24ade7ef651171408f4
BROKEN_LINK
zhao_d.pdf
85120 F20110403_AACIOA zhao_d_Page_26.jpg
0e56dd5d0bcde6aaff13a1774f8f9937
339cb83705b698b8e5f0f160b879b568efcc2648
969220 F20110403_AACIQY zhao_d_Page_25.jp2
76c4d931255482ea7c6f1c307c474723
cef70bdb389010a5a824dc3e962b1155ec1ecce1
1904 F20110403_AACIJD zhao_d_Page_49.txt
39a3a3c5cc362378346408708f746e10
4e2d835be4b6f828477f773580ecd03881a1a410
8282 F20110403_AACIVX zhao_d_Page_52thm.jpg
6991c13d12fd135d79a3f491edd5a187
1eb8e14c6b21d6aa39887a73a145347c836987b9
86677 F20110403_AACIOB zhao_d_Page_27.jpg
5eb6f9242b37f5af558cc1e44db5fe6c
01e2c2dbe061207a9651338d3f55d9cd3b540858
925873 F20110403_AACIQZ zhao_d_Page_26.jp2
b3d8d34d87b005c47f04661421e801c4
7497e026d2b137a4333788cb8da656cf6ef2980b
2814 F20110403_AACIJE zhao_d_Page_51.txt
a2c9dbb8cd6473a94f2694ac96e5c8e8
1d15a64c44894cf8e0c8bdbfa0ca4f25d043c7da
7724 F20110403_AACIVY zhao_d_Page_75thm.jpg
1cc28b5828521db5f5edc379b5275887
bde4e9f69fa9d41c628f6be63c97e7972e2ad664
98297 F20110403_AACIOC zhao_d_Page_28.jpg
c96beea08d9812571a789a92e66b6195
42473df3f5ee45ee7ce9d63e01aa037dee44b3fb
2000 F20110403_AACIJF zhao_d_Page_52.txt
9b70ea360a50fd11d7c8deb885e1ab62
e1be5fb4b9dfe2055c8f392c00967c2cb8ade561
7805 F20110403_AACIVZ zhao_d_Page_25thm.jpg
d239b7dcc7ad87c711d0dee77dc41e5d
244eb81cd532845821f408dca2ff8eda6021049f
1051950 F20110403_AACITA zhao_d_Page_79.jp2
7a3af2a78e9a393504c123bc71709cfb
38894c56ac2c77863521392f0b23f7032071ed84
F20110403_AACIEJ zhao_d_Page_01.tif
13e5e9b455d4a9bf601c8689a86c69fe
f30245c60834d780f2e5ccdf5a28b9f0f1e708d8
87513 F20110403_AACIOD zhao_d_Page_30.jpg
0e175489c74d06bdcd3ebcd94e98a3e3
fd1c8d38494bcf8cc6f315ea5b24ccee3d453334
2043 F20110403_AACIJG zhao_d_Page_55.txt
1404eb30a7192755ba8b7c0fdbd6ecce
1a23f0ed5e246ea7969e5b9f722d65a044ffecb8
327229 F20110403_AACITB zhao_d_Page_80.jp2
1dd168ff9dd1ffd63cb941a0c66847e2
abddb9610d3af8767109f6d59cedd20bbdec2d0d
89663 F20110403_AACIOE zhao_d_Page_31.jpg
6fcc7f5fd01e6bfdf0b681bf3404af57
0c78ffedafa4d2f5e664743a62349946c48f2aba
2322 F20110403_AACIJH zhao_d_Page_56.txt
e9579cf04bf84e8d65ff3057eaf70a4c
9e625262688d9c178f6783ed046c77249ba85d7a
870795 F20110403_AACITC zhao_d_Page_81.jp2
324c43857342bc777c7cbdbb00c7b373
e2cf5cdf1550a062b98c93642643e2a4115b3682
F20110403_AACIEK zhao_d_Page_02.tif
f3113539fb0c6bdbec353aa37ea58c86
fbc445a7925c8ef80ad272da5d28a38f1996a526
77019 F20110403_AACIOF zhao_d_Page_32.jpg
c3d57d1ae70602179b27fd2354249550
951eebba844cd38aaf162962f451101b027dd99d
2554 F20110403_AACIJI zhao_d_Page_57.txt
6824a1121b1b58f50c054c7c835a5b40
f7ce80786619b9cdef18671b158144b1170b1957
4125 F20110403_AACIYA zhao_d_Page_16thm.jpg
1ce0a75213de1eced8149bcc826db8b5
b37e3d21727bfc55b756bfd68a22eb3351fc0223
6597 F20110403_AACITD zhao_d_Page_32thm.jpg
74cdfc3ed745e29ec9879014709ff99d
a570a6561b9b9fbde838361597a1abc5a924b9cd
F20110403_AACIEL zhao_d_Page_03.tif
726b99314942126d362193c0809fd2bf
d8be0aa8081e37143652b4fa29e0d6304f66c8c5
85565 F20110403_AACIOG zhao_d_Page_33.jpg
92162397df531cb127f1d0fc28e7bce4
2c029563928303921841cfee4be7edb812297ace
1680 F20110403_AACIJJ zhao_d_Page_58.txt
a6383da4a4824a8a255dcdc995d45ba0
a4e54f7d8b1fa662b8bf645405b14f272cdbb816
7837 F20110403_AACIYB zhao_d_Page_17thm.jpg
4b09b3650413f1d40ee124c73fa791ae
8604be9824906671ea19898771c631c6cfaa0079
25004 F20110403_AACITE zhao_d_Page_64.QC.jpg
f7aa81964e96d20fe5e14e0e510a48e0
5102aa8437f0ab8ce9800d1a721f0c91d84b99a4
F20110403_AACIEM zhao_d_Page_04.tif
4f89bba930ee02bbbe2825d9bc9583dc
b608a68ad9e8a123498788ec8503854086cdb360
87001 F20110403_AACIOH zhao_d_Page_34.jpg
1b38ebe12d7c1a6d5b324c36904f036e
738ed7db228a0a401037b9808ed3fa07eddbb66f
1847 F20110403_AACIJK zhao_d_Page_59.txt
fa42417a4dd3d35795ca8aac3e5f468c
5284c0da922cb19de39a17f44dc7c18c8141f35b
7582 F20110403_AACIYC zhao_d_Page_23thm.jpg
a8b1e608b8b349cfb12de77978af6bed
0d035689aaf3ece21da6d2bf073a1f39fc8094c3
8635 F20110403_AACITF zhao_d_Page_29thm.jpg
174c59d5bcb35ef88ac11fa3c297541f
696d4d7060ac5db8888299030149f6d7c1d26c87
F20110403_AACIEN zhao_d_Page_05.tif
1851b28471d2fff2d79e1cc1c32e6a58
36089a8c11ae96b6eecc3e8ebf80b34e2c0bb635
102764 F20110403_AACIOI zhao_d_Page_35.jpg
14799ca3bad24d577f8c01ba28d337ae
79733eb6d2ebfc6a2744536a1b65bdf4ce6f49a7
2005 F20110403_AACIJL zhao_d_Page_60.txt
c114e9aa991e672037704db1ce01154e
3e9d484d19d64fdd677b64475a579d891b916506
7861 F20110403_AACIYD zhao_d_Page_24thm.jpg
7c38614ea34c4098c7b6e59569825727
3d048e9c7d655b7e4aeca202e80cd9fda3110b54
8431 F20110403_AACITG zhao_d_Page_62thm.jpg
7d609afeee48edab32ad25024b0941c2
4c5b6826566566cf06370d61aa42e6648b34fc77
F20110403_AACIEO zhao_d_Page_06.tif
d38570c5731a7b11ae56538795c9175a
6641d3e6ec6e23287e8be8722657935eec18b8d2
93570 F20110403_AACIOJ zhao_d_Page_36.jpg
d448c598644f267212719fd300ec018c
fab5340ebcaffda65fde09ac67a388746cd8ea56
2075 F20110403_AACIJM zhao_d_Page_61.txt
1b6e22b49ada5524dff2f25fdcf3e0d7
886e2f9e44217915e0030f9c5ce0b7c1c49898eb
7392 F20110403_AACIYE zhao_d_Page_26thm.jpg
cec44b83018bf3d7275b40e13218165a
25dc66aa7ed1b94b87139cb0a5ef548bc8113569
3812 F20110403_AACITH zhao_d_Page_74.QC.jpg
0a2dfc8a372d9c49a958701d64dd9d8d
09e8df96ff35d0c36b879f9a2163c653a5839779
F20110403_AACIEP zhao_d_Page_07.tif
fe21ead641a65cf34b4a29c4bd204669
1e1e5c30605ef9458fff0a1fb74c80f1152399ca
103358 F20110403_AACIOK zhao_d_Page_37.jpg
3cd2ff944bcf8d4350b015826d4298ed
634b43c5a763ede36197e7fef78f3ea6dc8661aa
2015 F20110403_AACIJN zhao_d_Page_62.txt
166daa56d3995f3ef81457d5521bf8a1
103b1b0e9f9166c9d8033f5cedd67163c6940e8a
8009 F20110403_AACIYF zhao_d_Page_28thm.jpg
927c27270c49d3936b042188ba6f5c9f
cdb32d6f62f8660d3fcbd8ec48df04b579ed3033
18467 F20110403_AACITI zhao_d_Page_05.QC.jpg
3ceef84fe7b6c08ccdd817766552f3e3
9832a13cd71b8b8fede358fda50ecc6661af9d4f
F20110403_AACIEQ zhao_d_Page_08.tif
5401e5614e9053166da789e3fe3558e9
253bc12ab3410b5600659c666fd5878c302400cb
106212 F20110403_AACIOL zhao_d_Page_38.jpg
40e928c230255c8f5664468717649712
94338234b4cf8a17275ad60f840213a4f775bafa
1925 F20110403_AACIJO zhao_d_Page_63.txt
0492a28e8e2eb3df59614a99afc68f65
5ad9eea44ab6cdca604adb044a1d19f61c0ec2cd
7181 F20110403_AACIYG zhao_d_Page_33thm.jpg
177014df6315194b7b70f6d41c8ed9c6
3b09fc5b1a1b54f77ac9b518361f4ee465061b4c
27993 F20110403_AACITJ zhao_d_Page_34.QC.jpg
c2708458f3adf757eb507672feac69fe
766177cc920ca5ba0b52b977e73f3a720f6ebd38
114917 F20110403_AACIOM zhao_d_Page_39.jpg
d66583814e84c0be5a44b84205a35078
300cde069b5e60a838d0138cb07e7b3d692ba3f8
F20110403_AACIER zhao_d_Page_09.tif
f573a0d8543c28624ed58da152d016f2
62bd7567f660cd5bf2e2e911f06d6af1c1fa340a
7812 F20110403_AACIYH zhao_d_Page_34thm.jpg
15a4f643d6ab8efc1d63ae40ede528e2
597fb55f42f2ee2180e271d50eedd6062a0cc392
34537 F20110403_AACITK zhao_d_Page_38.QC.jpg
80f15d92ffad2c15289c424aacf4d267
0989c91106b8e9d97d86536c7b6db5639a79639c
15363 F20110403_AACION zhao_d_Page_40.jpg
153dc2955b75e1251f26e8245c6899a0
e20dc4b13b3a4316081ab0bb0cd115d3742f30c8
1287 F20110403_AACIJP zhao_d_Page_64.txt
172cd50aba1e2c6e1fbba59eddde4eee
8f06b2c9c020816183f80c8da12ee7074ae324d8
F20110403_AACIES zhao_d_Page_10.tif
4b192f00b27bd5c28cb9b69c27738ca8
875b7154ca329222d0765c96cd8a17eeed094176
8126 F20110403_AACIYI zhao_d_Page_35thm.jpg
2169d0198bca0c61cff92a49f322f865
7e80dc79a5afd242e0a6e961df9e56ed46d5ce9e
605 F20110403_AACITL zhao_d_Page_03thm.jpg
ebf94c7349b3a2e5e959d2d06066f8e2
1577d91919881935100c52431e99b31da76a3855
94098 F20110403_AACIOO zhao_d_Page_41.jpg
04a3a07aded70bcbfa8239fe09cc8dad
a9907406b4110757e7434fa19aad1bbe8f7fc609
2141 F20110403_AACIJQ zhao_d_Page_65.txt
72972edbbeccd8edc190589eaaa746ea
44297f82097a10dacd986f6e4316d4ea97b4ad55
F20110403_AACIET zhao_d_Page_11.tif
bd81f89ccd942a345cdc23ceb2b00d04
1cdeb947bb91a3aa48352c87515c02e6f9b9d573
8031 F20110403_AACIYJ zhao_d_Page_36thm.jpg
748a9b85d2eeaae6a6e19793d667ce05
c26a8d97948276cc57a3b9f1a9b86f6f12496adf
6576 F20110403_AACITM zhao_d_Page_09thm.jpg
b468d4997425d7fa93e5dfd3e50352bf
63f2038a7e0eef13722d61967942533092a4c25b
100279 F20110403_AACIOP zhao_d_Page_42.jpg
104a27e8d55e63e7c75134518a3fb6e3
359b211591a8c34339c17b4c0fdcc6c653405ea1
1473 F20110403_AACIJR zhao_d_Page_67.txt
cea5604f01bf415fb97f8a1cf51c6c73
3259ba9cb89ac94519b87e1aff772188818585fd
F20110403_AACIEU zhao_d_Page_12.tif
8fc848ca04068bf1bb8d07ada9546d43
bde1e02fcea3393a469d68e623745c8611ca6dcc
7538 F20110403_AACIYK zhao_d_Page_41thm.jpg
a1311603448497487d67cadacbee9ef0
929986f7dfc8fae8fcc45af9f87a7e3e6ff9917d
32524 F20110403_AACITN zhao_d_Page_44.QC.jpg
806cfe200abafc02df55af33c74b69fe
178df4c71d8eb84ea67846e3bb2505d64fa37aa3
101119 F20110403_AACIOQ zhao_d_Page_43.jpg
1aabe99d8d8b7c47b1f238c5aa160b81
7d75363ede847900bc784fb2d8ab848a6c5454b0
2500 F20110403_AACIJS zhao_d_Page_68.txt
5b4970fb234312af6b303372fea8fbf5
e63c8496140fb80c6b109f368877f9ac3977c16b
F20110403_AACIEV zhao_d_Page_13.tif
fe7dabaa94ffc169cb993494b668f8c3
e7d704d56961b1730b448a7174105ee97727e9c2
8072 F20110403_AACIYL zhao_d_Page_42thm.jpg
3e3aaaa59988587f97dea8d5d93c5f51
f89a69f92306411a2a6442fbd4e614187991343f
32131 F20110403_AACITO zhao_d_Page_69.QC.jpg
2f1599ed83d812545ba5d82a84070b1d
c49e0f96425a815b5dd5094de6909a7d955425a5
99369 F20110403_AACIOR zhao_d_Page_44.jpg
90e972542782cf471a174a9c89a1657c
59c21535ecd8d0b7e25897ee9c08c4e275589004
1940 F20110403_AACIJT zhao_d_Page_69.txt
24fb1c13ad948ca15fb45d21af71a376
4021deee89955012f7f775f39fef05a0391a5457
F20110403_AACIEW zhao_d_Page_14.tif
9c11e1bd8c0fd42510e27df766e4855b
7adf4e9bc28c7beaf7acdbfe282b56de6f5f3a46
8263 F20110403_AACIYM zhao_d_Page_43thm.jpg
25bc9b882f8b0d4f6488ede4e442f4bf
2c3a0e588ba9511652f9cb440804fab29d0dca38
7492 F20110403_AACITP zhao_d_Page_59thm.jpg
83c32ccfac796530fe52d713bbd62c2b
c637be3f9e1059237a36a335f9eb841282d684c1
91439 F20110403_AACIOS zhao_d_Page_45.jpg
32bf3f4f2ad8a7005c4d654e2071dfea
963a3a78bbea8ff663dcd3a2bae774d51c2566f7
2010 F20110403_AACIJU zhao_d_Page_70.txt
35131b65a8a7b357acb3fe311ca7ad68
d72ed2f669b3d39133cb6b6ed11a388e0cda5012
F20110403_AACIEX zhao_d_Page_15.tif
f54bb8345f931beebec76719e8e82b9c
81142a71399b6797e531ad9fd7f3c4a71ff02a64
8337 F20110403_AACIYN zhao_d_Page_51thm.jpg
b6e19420fb13599fe5cad08f67ac786e
d03786bb4d5e83753b72022a7a6ee67930990917
1553 F20110403_AACITQ zhao_d_Page_40thm.jpg
b1247105b4bd24744942a1f83a05c6dc
0c2c787a4ba716ffbb7e4841395dbcf39ee9f9f8
102541 F20110403_AACIOT zhao_d_Page_46.jpg
806e6eb6eed195725a71fe1dae4ab341
252432d72a4f9c4f01bbc318731e040bc1ff6d7a
2024 F20110403_AACIJV zhao_d_Page_71.txt
c5f0bdd20a3b34611e386be5a60dd130
fd0ec37d0917104dbf1b759bb64052cc39c68043
F20110403_AACIEY zhao_d_Page_16.tif
386065162802e9cf0b54c83b3298a968
b771bad0ea65dd08f01808cefa9347ddb9166be2
7176 F20110403_AACIYO zhao_d_Page_54thm.jpg
9eee4a947c830da70a033a2f69c7fbde
0ce1e808606598c5d4550dcd36e9d43d5e5d796c
28157 F20110403_AACITR zhao_d_Page_72.QC.jpg
453e1f24e0a5d17f11d337b6f2d7ccbf
a1190e39b412fd8e1f122ac983dda5c76a991619
1674 F20110403_AACIJW zhao_d_Page_72.txt
70d10168e8812ea23c98af8df4d7352d
721fbc82bebfbeedc4d5f6123f1ea85f4a0275e5
F20110403_AACIEZ zhao_d_Page_17.tif
8c2070046909be75cceab1823aa6dbc3
20e5585a20e4e3a6babf441e747d3282fe2b41cc
6638 F20110403_AACIYP zhao_d_Page_58thm.jpg
3ea35d30328db48a6a61301139416545
c0e9ddbdac40ffd23332eb6c50b89a992269c994
32347 F20110403_AACITS zhao_d_Page_28.QC.jpg
0188cc06c4688ce339db1a7c572d0537
ec7a8238b2846806d83eb3c9d5425277d2381739
96404 F20110403_AACIOU zhao_d_Page_48.jpg
d74784341df9f09fee787fde73984835
e521f8826ff09c4c9425e6357cf8c5a3e97582ce
1859 F20110403_AACIJX zhao_d_Page_73.txt
3bd2c8419de897ea56db599cc189f409
1fdff06d1e1c7bd90e5e2b8e39605a6a161d6be5
8143 F20110403_AACIYQ zhao_d_Page_61thm.jpg
94488a96a18e7de781e9874359aef21d
97ecff6e675747c7d062941e00279cd2dd087af4
7097 F20110403_AACITT zhao_d_Page_57thm.jpg
dc515160d9f935613a99133e6d3d54c9
bc86120ee069d982cee4952d6df272d185537f76
90219 F20110403_AACIOV zhao_d_Page_49.jpg
4b09736ef701aaced1dcfa492e9200bd
b31cb7783438e0f02f503589f30056190f83a83d
F20110403_AACIHA zhao_d_Page_73.tif
38257a2a64670b642ac5fd88084b10cb
13e9b45648d5705f0c6547c569f6676ff0bf36ed
192 F20110403_AACIJY zhao_d_Page_74.txt
8b954a574446f564caad8888a2f35aef
2b19fa497af0b98da84da409f53c988adfaea885
6770 F20110403_AACIYR zhao_d_Page_64thm.jpg
479e5ea783c5b7a5607548f09520d224
07005d88d4a7a80e4a52f62ed7e660cfd340b73b
21886 F20110403_AACITU zhao_d_Page_04.QC.jpg
feabdcd2f082365a7e4b1bc86858e02f
af7e758d601d9da421f908640dadd814d2eed841
98080 F20110403_AACIOW zhao_d_Page_50.jpg
d04500eb25e94195b189f10a1bbc4c98
b74c1c9f808d1d86e86c9a3e2f71b8a6e0737448
F20110403_AACIHB zhao_d_Page_74.tif
3213142ad40aa2d437b435fc918f05e8
b30b730bfd9ddd577b76371af971fbfbd170bca5
2219 F20110403_AACIJZ zhao_d_Page_75.txt
06ff5d4664ebbc01faf9a45994a259ee
a7f2a0bc7ba7e539243bfa197a047d842d854b58
7987 F20110403_AACIYS zhao_d_Page_69thm.jpg
bbd159fb06cc615d2b26094bb3ce5324
2e85845a355c49f85481293e33b767299a0ef7bf
8278 F20110403_AACITV zhao_d_Page_56thm.jpg
6887a4a4b371d9d1667e60db54a8e1f7
39aa7fd8efc44bf51ff38e8a7e030fee0f6ce0f1
101734 F20110403_AACIOX zhao_d_Page_52.jpg
b0f33f940bf6b55f7feaebec27590e26
eb6207b4a805a0e8c1621b094a3dfcfa8ab1b133
F20110403_AACIHC zhao_d_Page_75.tif
f158558279df7a44ba72e9157fe88934
0a356cdd4182aab4bf15fcb2cab14072126ddea1
8547 F20110403_AACIYT zhao_d_Page_76thm.jpg
e466b1a4c6c8f3992b8ecb4ae9eb666a
6c0b15a648dbf1732064e10a2c398ae0cf7831a8
8678 F20110403_AACITW zhao_d_Page_79thm.jpg
fb937318886a836a9c9dcb39a277183c
2ed1e13d85335511c0b393a35b31e65a74ff88a5
103006 F20110403_AACIOY zhao_d_Page_53.jpg
6fcadfcccc303208b17cdc74e0e51b96
7fb8707df10f98fd9206b94456af10b17c3ff850
F20110403_AACIHD zhao_d_Page_76.tif
4d54278bf161d38fac3d555a8f87dcca
9ed957345deed1dc226abce9f9873b124233239c
59112 F20110403_AACIMA zhao_d_Page_51.pro
47dcfaaadc540a365381bf3e37b26006
e4ca7cd344e935f16ed5410369388a5a3d3d9b8f
8318 F20110403_AACIYU zhao_d_Page_77thm.jpg
99de4a6aaf39cbdaa58bd0be45df7558
8507a97d9c192f197394d0ebf4edd64165f12ad7
8155 F20110403_AACITX zhao_d_Page_47thm.jpg
a087d3b1662408da732c792d48c71f2f
3f191bbd185fd8aad4b8f5cbb488a779257bb485
79839 F20110403_AACIOZ zhao_d_Page_54.jpg
352d6496175052d4dc4e3fcc12810c6c
ea46432a6b8bf541045d6fdfb7652b8229f2304f
F20110403_AACIHE zhao_d_Page_77.tif
89e22e8ee8e6ef429b5bc7c4a49ae0a5
26c458e635e8f382d66e54ce26c319692b57586e
50949 F20110403_AACIMB zhao_d_Page_52.pro
b6824dcbae479c1551bcae84fc5338c5
f783ae782d317ef3dd94096ceb0872b227615737
8924 F20110403_AACIYV zhao_d_Page_78thm.jpg
3a7f0d2896b742eba699bcf4fe193390
665fd30416a8b25a44057bc4f742809520298924
8439 F20110403_AACITY zhao_d_Page_37thm.jpg
d1f522eccaf2c4c53e70d52fc9e067cd
bfef996151139b88cb5e3b67d47bf95a8303c405
F20110403_AACIHF zhao_d_Page_79.tif
db7fd257e3ac4ad0628cd7ea6d7a0f33
92d4c892f726e84bdcc615aa367a6cb17207a194
F20110403_AACIMC zhao_d_Page_53.pro
e1244383e38a504e66431b261b11279c
3461470763a576b876edc51c4446ab0c18728a06
2530 F20110403_AACIYW zhao_d_Page_80thm.jpg
98168a2d4495dcd2eecf8d574624d3c2
8d3e7eb1206a1052fb185c8bb0bb2e78165f6ff0
F20110403_AACIHG zhao_d_Page_80.tif
4d914815364848b04ec63a7f935549ef
b9159bd34d34737fac74c923d41d7060f76113d7
934775 F20110403_AACIRA zhao_d_Page_27.jp2
e27a6b7221d35a5fb830e8ce31c2a09d
1d218231d4cfe348aadc57932ab37609da626195
33467 F20110403_AACIMD zhao_d_Page_54.pro
2ad3ebe6aff56af4d7b0bf6b30f2a796
112b62e7343fdfca077f62f2dc22c0f98cbd6375
F20110403_AACIYX zhao_d_Page_81thm.jpg
3245a98a542c9c45212c3e174662f85a
6debd665f5e9cbb6e552d08c8b7b5edcce00c20e
25736 F20110403_AACITZ zhao_d_Page_81.QC.jpg
d66ea59cbe0743eb6a2b472482aa335d
3b5c1d3496f98313de61f13ba7f138ca9e3d8246
F20110403_AACIHH zhao_d_Page_81.tif
ad653c4af58ea87767373691d8a8c24f
b15cb8b86d4a57090d9bc779a962c5853924a91b
F20110403_AACIRB zhao_d_Page_28.jp2
7a7665d19abcebcc91d64a2606d2169c
1ccb7b7b6f0d59a956e98d28c2541c8b95c8178f
51886 F20110403_AACIME zhao_d_Page_55.pro
920008a16db10199aa8b9ba177729e89
d1e9a690ce8d2269a94c7eaa9bd419392c29a207
463 F20110403_AACIHI zhao_d_Page_01.txt
357bc5671daf5f5ead5a492ace907df7
f134f4ca6f0071564adfb20052721dfdf1b8e2bb
1051944 F20110403_AACIRC zhao_d_Page_29.jp2
c18955ffe55bb335d30fe7bfcd0b137d
fc065b9a7748c151ff699a09e6d30d26e1d70e57
50774 F20110403_AACIMF zhao_d_Page_56.pro
b98b0a56f6eb7c9ec0c8b376a15c4fbc
52574826519960eef38f13cb41041846565f5b02
F20110403_AACIWA zhao_d_Page_46thm.jpg
36064f9d6785762dd495646f46ed6da2
56195bda0257d947fcd2df0fbad65e032ca940fb
105 F20110403_AACIHJ zhao_d_Page_02.txt
24cd902b4bbd7b8313793f642a123ef3
36d85f422696b2fbe1c0de3214de470bab3e0280
951143 F20110403_AACIRD zhao_d_Page_30.jp2
d07ae9043dc3dea5331ebe01833f5365
273f359dd6423c799eed4a1e6543f3f5ac684a18
52942 F20110403_AACIMG zhao_d_Page_57.pro
97f9f8e45d188aedd382d3739cb80252
179349d66f9d216bcdb724862da43dee6b1c62d3
3369 F20110403_AACIWB zhao_d_Page_06thm.jpg
400dc6e499d15fcaa36199b54978eb05
8024c0ed064c5a870b5a060e23e686857dc8ef10
110 F20110403_AACIHK zhao_d_Page_03.txt
3dda4d88f717bd9b89d596ed2cb19860
f5bcf238a27bd2ff2a6b2c14ef18c1f59c6819e4
944041 F20110403_AACIRE zhao_d_Page_31.jp2
2b2406a497006ab5766f9d2621d3ba4c
808fad8a1bb998f88413f5f87465cb5e891045ca
41170 F20110403_AACIMH zhao_d_Page_58.pro
3a312a6b081d00fe8d9d3fa1c99501f1
c67e87e4d5b0f403e959f4fd516e1286127de2a7
31269 F20110403_AACIWC zhao_d_Page_18.QC.jpg
0abae258560960501399c3da4c775428
fe97619ad7ae5fd066eff50f34e6a370957e1416
1260 F20110403_AACIHL zhao_d_Page_04.txt
f26bd8fa4d9345eceed161a75244cdc6
67b314812f103c943be31b2e7143951e2e2d0eba
793471 F20110403_AACIRF zhao_d_Page_32.jp2
64bea7594dd832a3be5077bc497ccb90
be6ea980a3f96670d48a972026a36816ccc9cf41
45772 F20110403_AACIMI zhao_d_Page_59.pro
38392237d6afd05290369bc33361363c
988f7481e637f0e9f6d2221e924ce3daef0ea87a
8074 F20110403_AACIWD zhao_d_Page_18thm.jpg
3f5e0e571928ed494a68fc25da5e2484
c98b8509826d5a4038b6ce746d10feca6fd8102c
2297 F20110403_AACIHM zhao_d_Page_06.txt
60e9fac5fe76f367c866dfae5c07985e
b9d8cc8564518ebd2c2c37f0fd9419365a338d35
881800 F20110403_AACIRG zhao_d_Page_33.jp2
162cbd87032eb1f74a83d3f02122d9b3
bdb03c5004d8b12afdbc876375a32714b3705086
50033 F20110403_AACIMJ zhao_d_Page_60.pro
71c42d3160dfde338c15a17b5a1cbfb5
6a14391be6980277d124aafae0329a18b747b0a2
7624 F20110403_AACIWE zhao_d_Page_49thm.jpg
9cff7f79391e42ecc4fd2441b7abd8b4
edfd1cc5f6a7a2e4ded255b56adb69474345728c
930357 F20110403_AACIRH zhao_d_Page_34.jp2
df3dfa7d7bb7d5fbe9a1d43eb4a90cb9
9908d741a396abda69459f29ca521d4727ab76e2
51916 F20110403_AACIMK zhao_d_Page_61.pro
9af5dfbf7f0312e8ae745d4deba72979
8598a2852ac05fbb1deeab200c8473f5c17a2761
33918 F20110403_AACIWF zhao_d_Page_35.QC.jpg
7a76ee55ff727deb62ab03c8241e270f
7a77b2f6d10aabd1448c5943ac4e80e8f1c40feb
1051948 F20110403_AACIRI zhao_d_Page_35.jp2
076481f7ae671bd3238df860e019b8d4
d0d39b6dd0d6efebe4e7fb0df528c318bdd810f5
48774 F20110403_AACIML zhao_d_Page_63.pro
936855181b2168b43d15092370f5114e
9086d33a11919d5e8873b63ec9761ccbae9b2e0b
2678 F20110403_AACIHN zhao_d_Page_07.txt
f45017ec90785a4d1fbf978dc29e849a
6b952356af5aecfdb8d3d06236802ba83f27d9a6
25837 F20110403_AACIWG zhao_d_Page_15.QC.jpg
fb08f2629d22f5a83fecd8bec0ecae55
c050f824fc8e494885f7d0c62aa3fffd41adec16
1000596 F20110403_AACIRJ zhao_d_Page_36.jp2
0604d75de951e9a39872ed7611db40fc
895a44821b5994088653e38bc2110b56f3280dbf
26343 F20110403_AACIMM zhao_d_Page_64.pro
32dd2f5a936b1bc7c4620b3db4c0c144
9533d32a80711f1738bf1777642b583c2e150974
570 F20110403_AACIHO zhao_d_Page_08.txt
1d621a901f2dcab85a10489e4fba7f5a
ebfb3d9b47ee210c1a8fa56837d9c1e25639a71e
16125 F20110403_AACIWH zhao_d_Page_16.QC.jpg
53818b2f6ef1a46aff68f557a7770f21
2ed4d07c6b00ed25d7077f3eaf3f5ee4fdf30259
1051960 F20110403_AACIRK zhao_d_Page_37.jp2
d5fb4a9ba802afef72178a1b672b15b2
f29b63d0523a220c7b97f1a9a28784d5f4a15a4d
50804 F20110403_AACIMN zhao_d_Page_65.pro
8aac0a1bb790fe860431fc40b6062d1e
3f6d7bfdc858ab4f9dba733c63ae6caa23738e1d
2621 F20110403_AACIHP zhao_d_Page_09.txt
b02f9ba2e6d8389f8c0914cb8da9f241
3293d0148e00246592cfdb979f2a22d6a294f84b
28488 F20110403_AACIWI zhao_d_Page_31.QC.jpg
39f1a4372bf741c7f4b8d0dc587bd2af
beec17e0f88592874f20a5d807f24a6ae0d89c80
1051882 F20110403_AACIRL zhao_d_Page_38.jp2
150186debc6856b2a653b499d9a0d389
05932513ae544c37cf2dd432297e6d1a18569334
55326 F20110403_AACIMO zhao_d_Page_66.pro
da7e504cafa100bb1e35b7b641fbe76b
c819894e9e031ad09e6364ac40763ad9b3bd15cb
2834 F20110403_AACIHQ zhao_d_Page_10.txt
ae5add712ba245732b1b839cf610aac8
5ba4ec24ef09ddaad3769f776f365dff2fc7855b
1425 F20110403_AACIWJ zhao_d_Page_02.QC.jpg
996d50c58f3490cabceed96ab3090086
5245470c9f6e2d1a73445a7bee87364d1b381597
1051936 F20110403_AACIRM zhao_d_Page_39.jp2
69f4385adba1d4aed40a67116dfffc4a
c0f2f9fd21d89a90d914e3b681ad2c82d68bb669
30738 F20110403_AACIMP zhao_d_Page_67.pro
0c2ad580f7d8958713e861973b50481f
368a2e315aaf907f632321b379382cc354d0a5ae
1742 F20110403_AACIHR zhao_d_Page_11.txt
be2ed3f7985f8983afb3165ec8e05031
b9f7a6b435cbf18f3910a261073c87c972825e55
27708 F20110403_AACIWK zhao_d_Page_57.QC.jpg
961c32368eaa600d640f49f3efb242ba
fa6283a16af2d5a503c18978ff0fef8056d4731e
141006 F20110403_AACIRN zhao_d_Page_40.jp2
e901573e783a413364ebe4c562b48ddf
56a944ca9117e36c247daa006fad446963f5c4ab
49140 F20110403_AACIMQ zhao_d_Page_69.pro
df13759312008f4da789b4506a39a076
2fb13a0cc24ee9bd7a2db9b34695455cf9f9603a
1119 F20110403_AACIHS zhao_d_Page_12.txt
916d7761bcf9b8abb2bb2c2e4cf0f51b
177d5217c085a948ee7bbb6b09bd2c55b44c13cf
33120 F20110403_AACIWL zhao_d_Page_71.QC.jpg
4bd8cf94aabe4de84b88a003e2c286e4
3e0c6940147b5f9e8538704b6888328211c0fd29
1020919 F20110403_AACIRO zhao_d_Page_41.jp2
2d91e6ecfedd21548bc0402d24683dfb
c24ab51efb7c95f7aedaf1e4eff302e2776c9188
51075 F20110403_AACIMR zhao_d_Page_70.pro
69c44787342325f9dd10abcf678bc335
d82c208ff376c3bd52c33e8bccdaa2118e89b188
1756 F20110403_AACIHT zhao_d_Page_13.txt
7d21aba3e7bf46079d65ee89090abafc
a48fae64e042de2658a4da3e53c90ef9a6d46c26
32308 F20110403_AACIWM zhao_d_Page_60.QC.jpg
a4974dc9217e4034cd14c79b9ba9af80
afee69c11f746fcf6e7b834d39dd4b408b7ef2b7
1051974 F20110403_AACIRP zhao_d_Page_42.jp2
adfa50cf8da10188f023a137a8959d08
d6254324c5d3eb8949306f06ed765ac72e8d2c85
1982 F20110403_AACIHU zhao_d_Page_14.txt
2ddf834a4bca62b719ee909b23b7ec4d
344ec7781430dc596a627cdbcf73aac098698b8a
30626 F20110403_AACIWN zhao_d_Page_59.QC.jpg
1a7ac448c6167a152ae0295a4d4ac07c
68bd57d06e52b8606408c8d4c427f3d5df96f123
1051983 F20110403_AACIRQ zhao_d_Page_43.jp2
054536364d60471692f065576825b34a
7125b993ba93f1b9e931bd3caa4fa8c486e9ac39
51352 F20110403_AACIMS zhao_d_Page_71.pro
e8b4d1a640818448d6210ecf5ffacfe4
51b09a6793262a55b05851182a14b1d50f0bc9f9
1591 F20110403_AACIHV zhao_d_Page_15.txt
25e3b4d969655de2d6571f1b16a113d4
63f7661168011eca97c9ea52e521c633c718c079
5577 F20110403_AACIWO zhao_d_Page_21thm.jpg
c1043ca13e678e54dc295f2ca12056f9
f04371ea16df1420c290ad453d317ae582bafb06
1051972 F20110403_AACIRR zhao_d_Page_44.jp2
34e36bfc567ab0bde0cda02f966d73a8
3a134cc9706a9cceae6ae627b4045b8eb6be7dcc
41174 F20110403_AACIMT zhao_d_Page_72.pro
39c03a9644198b7f8fb497605ee60c4a
35af5b354670f2f44dc9deca45e9274d9e8513b3
914 F20110403_AACIHW zhao_d_Page_16.txt
b5e56d48abd9f470d627f27b9e55ad19
6b911c3090362c2b64a6a96e4b4e29f522b64a7a
30336 F20110403_AACIWP zhao_d_Page_41.QC.jpg
46dffbb4d14eecbe01bb742057b67137
1e8f392e986aac066034dc83420753a3608a263d
957948 F20110403_AACIRS zhao_d_Page_45.jp2
a82b774bd3fac01e7d7bb2bcc029f3fa
8db500b6c85971c1dbd8536462600fa01a0e2feb
45089 F20110403_AACIMU zhao_d_Page_73.pro
dbb382321f5f69c3f52d38ff2555d871
d2d9d113826197743a6e449fafc5db400d793346
1849 F20110403_AACIHX zhao_d_Page_17.txt
cb92a1e2e2b0d70eeafef2584a3f1c77
f183810ed6caf06dc1097f609adcc5c77e2342a1
8561 F20110403_AACIWQ zhao_d_Page_65thm.jpg
20db0802895fb19ddc8115ec7e2d9ad2
9ebd5e478aa7ea0d4a1f38d8f10b98cf39fe3a1d
1051959 F20110403_AACIRT zhao_d_Page_46.jp2
f03a3aba303e9a017c993bde0f7b19c3
3af0640d876ced3a7921fc960f0859d7e9b363d1
3764 F20110403_AACIMV zhao_d_Page_74.pro
4bb4255fa4d3a0ed0356b28a40be3478
6f91486a462c413dd1b2677ed7a77eba11a4dc3f
F20110403_AACIFA zhao_d_Page_18.tif
4b44cf34c0805ff2c1a31f8d389d6217
3708e0f743214ad48ff127f292d7d35486c54391
1910 F20110403_AACIHY zhao_d_Page_18.txt
9fb8b5cbd97b1cae5f183aaf98db97b4
eb4cf2b1d7de7000b850db7daf2a120e1cf36bf7
34194 F20110403_AACIWR zhao_d_Page_66.QC.jpg
e820dddfda00bca87a9fc9150df6e3d0
85b66442afb23425d0bafc31de26ddd32f05b367
988577 F20110403_AACIRU zhao_d_Page_47.jp2
4121cbcd0b6edf22b52fc5ee5dde49d2
ddc8107d269a9c93129255a6e2f6b4961a5a9564
54687 F20110403_AACIMW zhao_d_Page_75.pro
69e2317e6d4a9df5236cafaf8980888f
93b77c951787f2a9a733cdf4d9504ee873464e62
F20110403_AACIFB zhao_d_Page_19.tif
f61c22cb70790fdfcd77d4309b1ee35e
55870fa086a9a085a822dd0a8a091197f1d7ce59
1891 F20110403_AACIHZ zhao_d_Page_19.txt
af163494fc61039583b618da58929a4f
bac37821cffde9fb04278707287b43919542a6ed
130358 F20110403_AACIWS UFE0013600_00001.xml
a17789f8108685b9b3ea679099fd681c
51208c590f6340748c8631cb288a2073babda189
zhao_d.pdf
1033688 F20110403_AACIRV zhao_d_Page_48.jp2
611fe4330de0b172e3422672cb1f2b6d
fa0485549a852d317064fe5b4c4c9a0e5c4c6f6f
63228 F20110403_AACIMX zhao_d_Page_76.pro
bccfe20984051e7c999ebe80641f47f4
dce10c3d8e66b2a1ba31a7094f587f9e8a303de0
F20110403_AACIFC zhao_d_Page_20.tif
1489e8437c10f5ef61a28f8436087dde
61a4c8b634797314f96fa7c314c95a4286a36787
14443 F20110403_AACIWT zhao_d_Page_06.QC.jpg
c8fa3222f65aaa293a4b485bd888b1f3
40fbe4acd1efc79165176c873fd50343a5aa7f69
942858 F20110403_AACIRW zhao_d_Page_49.jp2
ff081f10167496c95ad91367c05edda4
ad81118c9ef7245568dbf9a599f8380a60ab2c20
2559 F20110403_AACIKA zhao_d_Page_76.txt
9834f3ac679d6e206dea720ef1a7043c
379d07145e99ffe764e22669ffa318bbfa51a699
64464 F20110403_AACIMY zhao_d_Page_78.pro
c4ea0dd752dc493a392e0108e5404cf6
35c486b404b9123d0b798de17d7c5c7fbc52749a
F20110403_AACIFD zhao_d_Page_21.tif
5629f5c44ece05cfa348ede28ff88a52
3091d9f747f1bb2896d27bad3336adb6b85c702d
27086 F20110403_AACIWU zhao_d_Page_07.QC.jpg
1e7751449b7874cb419fb59c7595155c
c3b8d746556d1a6ef06225d7f729f3e8b4d96f70
2514 F20110403_AACIKB zhao_d_Page_77.txt
751b3a7cc82b3fb873734a5336459d98
b505ed40d0536d7ee67251072ab9f928d8d0e3d8
64761 F20110403_AACIMZ zhao_d_Page_79.pro
154ebe0dbab4a5caad8c4eb9b136e4b2
c552217051d8654f32732a83ed4ef91d32a5a5fb
F20110403_AACIFE zhao_d_Page_22.tif
4c59c2f18efeaf9acdfcf2fe78a621af
5279f635751191127c3b88f8763806a43882ccf1
6410 F20110403_AACIWV zhao_d_Page_08.QC.jpg
f0c6e78c537ff67aa325bfda14ba2a61
af7525be1328800c57257cfcd604b73a2232868c
2639 F20110403_AACIKC zhao_d_Page_78.txt
42f9a5e94aaa0bc3b41b4dd76559ef17
2f3d4c3d09a08c73ee40362731d898383d844390
F20110403_AACIFF zhao_d_Page_23.tif
8234c63aedb5595fedbbbbf52087feb3
6ed3558468abe18d31ffe3c9ed505746d5479c78
1046955 F20110403_AACIRX zhao_d_Page_50.jp2
db922deeeb5f1eb863638b8cae4b0c9c
43531cbd9a315ac34c72b7c3b6cc5a662220047b
27555 F20110403_AACIWW zhao_d_Page_09.QC.jpg
2c4d19563928614d3abc9dbd3c66e493
92f878f5ceb494f8b56a92fcb44404f498d92663
2629 F20110403_AACIKD zhao_d_Page_79.txt
05d7cf3cdfaa8b4b61e2f2e55c4fd393
df7e0794e0eece3c3937d9fb7801945db15ed809
F20110403_AACIFG zhao_d_Page_24.tif
318394ab4fa1e8dbcd1f980faf8cca6e
37806d3227d59f150faf83780409f715a55e376e
105584 F20110403_AACIPA zhao_d_Page_55.jpg
de990791593031ef85aede196151eb71
05f4c4f3e9e6dbc79ba2cc55babfaf36f3931ac0
1051898 F20110403_AACIRY zhao_d_Page_51.jp2
c6ee244d5ff7513ee8788ca831185961
32332e29b369acf9f4a117f71ca2093127f82455
25800 F20110403_AACIWX zhao_d_Page_11.QC.jpg
2ec82373b9a96644de2659bc4b99c2c5
e1ae43c0896621778bf855d0150e3f91310ceb92
641 F20110403_AACIKE zhao_d_Page_80.txt
fae5f94f26e9b351c0a9fe4132d9fbfe
84a2cfa7cd5c2274d710128e9c07ab37d205d209
F20110403_AACIFH zhao_d_Page_25.tif
64aff06cfc4c14c398c898136ae97803
3837a885005f97b53d60fbe02552548883c25163
97826 F20110403_AACIPB zhao_d_Page_56.jpg
2e1c4fa57fa5ae97978a1226a5f414e7
5f84c3091c914be450ac18c8105d6a3d3ff8931f
F20110403_AACIRZ zhao_d_Page_52.jp2
4703a8acf7991ac8c1bbe49117ab4352
db65cb5fa96253bfda65c4532b19161cc916be10
27314 F20110403_AACIWY zhao_d_Page_13.QC.jpg
d58a0efb374a64632c44611a94cf8e9f
19d90ee2c678e08427e1bf9eaec6598ddea35ea5
F20110403_AACIKF zhao_d_Page_81.txt
610fb9e3514142dd87b5ba0176ae38a1
947a0718a38b4cff5a6634bfb0636c726c4f0867
F20110403_AACIFI zhao_d_Page_26.tif
2b836b88340fa0c3851625dfe32bad1b
ad21cad039fd8b657dea378724ecc4b72826f80e
94973 F20110403_AACIPC zhao_d_Page_57.jpg
af531607e0eee4701b4c133389c023d6
598500011ccc0c39811a7874c24646a649779d4f
32529 F20110403_AACIWZ zhao_d_Page_14.QC.jpg
9b5c0a846f93d0e231482f4c8400fd39
418be42cc8422994f7283750c95469945fa5586a
6721 F20110403_AACIUA zhao_d_Page_15thm.jpg
5ccdaec01fd34d7cdb6dc560a527ea3a
4b755a3039639b28d76058218b575a21f26aae82
8568 F20110403_AACIKG zhao_d_Page_01.pro
bf6010d1154c3257dbdc4589cffa96e2
f6350afa4c8c30f36caa07a4be229ac12e87d42c
F20110403_AACIFJ zhao_d_Page_27.tif
027ed401e6527c7ff623e80b4c0cb553
706500284bb9dd87299061cf4cb35755530907be
84820 F20110403_AACIPD zhao_d_Page_58.jpg
6c7665747ade451b58f98f0a0832f067
0b71f16e47d28b9d96001de1e8088869ff5d0b9f
8401 F20110403_AACIUB zhao_d_Page_66thm.jpg
a2dd1f60694c041c67744abc3b80d952
9c7fbefb9e1f0c3ada81d3ff3f7fe9826ef62555
1015 F20110403_AACIKH zhao_d_Page_02.pro
180c06ec3f174718569815ab979add1b
d4a254b10cbef42793aeeaeb50dba86814518b14
F20110403_AACIFK zhao_d_Page_28.tif
df6175e7bcb419568e6c54d468154a5f
70b018b489aed6c1977d4de364ae30a1591df8a5
95087 F20110403_AACIPE zhao_d_Page_59.jpg
69de8defbdb9baf16aac54177efe7497
d2b4d01937c88e670b8a736941cfc14529b0e6b2
32276 F20110403_AACIUC zhao_d_Page_20.QC.jpg
9a420a7052f3c9af7cbaadd9747321f1
ecc57b5d71890a78b11f81fc3ca2e9246af9ec03
1453 F20110403_AACIKI zhao_d_Page_03.pro
e4e1c949d73d563fbc05d8ca80d3267b
0c9cc5c5f5d49165d22165abbd94a0449966a827
100533 F20110403_AACIPF zhao_d_Page_60.jpg
7adabd2fcc01b2a2f8e0890a7200a024
5d27d9cebda0869333b3eb00fe711392df9ba324
6254 F20110403_AACIUD zhao_d_Page_07thm.jpg
16c2cc61a22f799dd687cb1a2e362d3b
a155681fc6ed16cee5c6c8205320b32e59c38761
30227 F20110403_AACIKJ zhao_d_Page_04.pro
e89bc1a4fca7e487b5539428e7a5d5f1
7a2e61353b439cdae4d538c93e83819e2ed8820f
F20110403_AACIFL zhao_d_Page_29.tif
e73612cbd02a749b37c535e61cab21c3
1478463ae027f7425556e179169bae56ae76ff0c
104203 F20110403_AACIPG zhao_d_Page_61.jpg
364fbdb1350ec0129bbfc124b5d53753
4393d2c93da688b7a36ee7cd9d410d01497dcb6e
33588 F20110403_AACIUE zhao_d_Page_29.QC.jpg
37bac5414d968f4d8b9c694db5201717
ca3e56536faccbe0d6c3652c59e493989e0128db
F20110403_AACIFM zhao_d_Page_31.tif
aad54394ef65974882937d213f9c8da1
f3171c36d79ed6500aa1d89f30195796b53210cc
105127 F20110403_AACIPH zhao_d_Page_62.jpg
427728ade371d33a1b080e1d737e2cb9
522286fc4363921e6a9090b8315b6e108b8adbf8
70655 F20110403_AACIKK zhao_d_Page_05.pro
0a2916cb2f0f2cde4117e1729e61aee6
b8478aa41a97394d052e9a4098d72e3a64f66c70
8799 F20110403_AACIUF zhao_d_Page_39thm.jpg
92aa53807228d7d334231f094e5a0874
09c4468afddaff3fd0ee6d423fa0b3e604367f4e
F20110403_AACIFN zhao_d_Page_32.tif
05cd6bef6d9917749bd1398a3214045c
45a68a3c1f3e9cbb6e71d7d40f4b8637012121d6
99586 F20110403_AACIPI zhao_d_Page_63.jpg
9482f0e089268872213993ccdacc81a8
3b846025fb3fd518ce30b061c77b0cfcfc52fcbb
56294 F20110403_AACIKL zhao_d_Page_06.pro
4b547679ce31104fbcaef10de32ece71
ca183f48468f5db0fd260db383b969bfc9aa70eb
8463 F20110403_AACIUG zhao_d_Page_71thm.jpg
2a7c32608cbdf4ce57f1a5e9e0f1f684
2b4bddde23bc0e32687e146207a099c313018b3a
F20110403_AACIFO zhao_d_Page_33.tif
186d31d7313f493ff08aed20c8aec054
258a15743d4a6e81cea99bbedc982e7fd0860263
70274 F20110403_AACIPJ zhao_d_Page_64.jpg
1beb9ae39586863dd6a9ca65f0b61808
6754f76dbdb9fe323fed13afc6a95c110d4e66ea
65243 F20110403_AACIKM zhao_d_Page_07.pro
c46a648fe2a1ba3a2d2ebc64618b329f
907cd2b1589fea3931ff671b2221be16c3163212
8269 F20110403_AACIUH zhao_d_Page_53thm.jpg
460f3b6546ebe0e879a94676881b57b7
eed2be23d0bfec3be1fb316115dfd47dda5f98b6
F20110403_AACIFP zhao_d_Page_34.tif
832f3a36df126df05efc002cc30ca85e
f111b53178cd0892a4813835265b3d9d201744c2
104827 F20110403_AACIPK zhao_d_Page_65.jpg
fdbd154ec79d33b04a7e933662f50e62
4f7bcdd40387dc889b62bbbb01e837426533a42f
13858 F20110403_AACIKN zhao_d_Page_08.pro
bdf2981ee58f9d4d8881430bbd8c3926
dddca8991841e11f57b2457291b77cfa6eb96871
28915 F20110403_AACIUI zhao_d_Page_49.QC.jpg
a904f7c3e8531b6b6d5cb6601b8273d5
2e2b47700a1ae86509878ba123b70104a5ae4733
F20110403_AACIFQ zhao_d_Page_35.tif
6d358fe918dad8d13942a6c80bcda6d0
e33f4dcd6126a1fc61748879dfd64c9f301b997b
109008 F20110403_AACIPL zhao_d_Page_66.jpg
f16ac01ad080605321aaeb5c8826ec29
e584656f6fe6f2e05a1dfc01c75ff560d0c3d3cf
64268 F20110403_AACIKO zhao_d_Page_09.pro
179332e3b36400f34c5621bcae0d3e61
e2e39e944d9f35c32fa54e8a9d46bb5c43b7c3e9
35558 F20110403_AACIUJ zhao_d_Page_39.QC.jpg
888641ee0cc8c1636701e5a1131ba963
adbe80d6fb9e2fe9c9e46fc9cc8db3fd602b7ce9
F20110403_AACIFR zhao_d_Page_36.tif
89c40b587c765920f7672db3efe68898
3d27034e571cd6272cde6ec0ad7b4b512a92f2db
85497 F20110403_AACIPM zhao_d_Page_67.jpg
65595be4a6c852248a5df2efadea363e
9a636790c4e569d241720db8330ff654b08f1abc
69163 F20110403_AACIKP zhao_d_Page_10.pro
a57ee50f68f9bea1e9886c099befc7fc
0fe242dab9984d958fa468e91452f3e677713223
31641 F20110403_AACIUK zhao_d_Page_22.QC.jpg
29a6379a28f632c14c74551f87dbfb2d
f4984258a93b258827a1dee935e07ee234a9b99c
F20110403_AACIFS zhao_d_Page_37.tif
8c3c0f6b85f18df04989d240efa4353b
9ce282fe9614f51a52d47538c4751b93e76cabd2
91242 F20110403_AACIPN zhao_d_Page_68.jpg
d262d5b8bf04cd165e03bf681889efa6
10598c3d085535ff195f8074fd663e1fc3843056
1776 F20110403_AACIUL zhao_d_Page_08thm.jpg
ce230ccdfb68e6720cb40b22349ec47d
10d8163e406c0a65f5a8ad37c5496e2467c8c753
F20110403_AACIFT zhao_d_Page_38.tif
1198c8529ce08661ee9723bb7b87b9ba
610a56c6fd0cf027227cebafb1a6588cee6d2d7e
99603 F20110403_AACIPO zhao_d_Page_69.jpg
abb4658e643dfb413230fc80ffb6b1f9
5cd91734bbaa6ab9b4a299fbd3a46e8627351749
39611 F20110403_AACIKQ zhao_d_Page_11.pro
0e526687f5a198f94752c3dd792331bd
4e2e6d4fa8588e87caad175d7250bdbb10b6b759
28141 F20110403_AACIUM zhao_d_Page_27.QC.jpg
9ae44a4ebbe4f896429a6a7649164095
9b385804f7b3b43c10f69c14a46e31a0ca41ba04
F20110403_AACIFU zhao_d_Page_39.tif
98422ea9307ad90e75d7c221619e78eb
fa07b1d0adde8403575ef3f244227d0daff3eb96
101171 F20110403_AACIPP zhao_d_Page_70.jpg
eebac12f99654cbfddc2c2ecf997c74a
ffa2e2a625d7642009ae2c7c4c23143e7f0e5c90
28024 F20110403_AACIKR zhao_d_Page_12.pro
f1b7261d942fcd1df9f9c6cd868fa8dd
db853e55bb65af1b68c456326ed1ee34b684f355
1074 F20110403_AACIUN zhao_d_Page_74thm.jpg
81be2f2e7c19140c0fba63da616e829c
53e59a37abf17f73fadd1094e872b45f5bda7855
F20110403_AACIFV zhao_d_Page_40.tif
1e5333f6820ab71caf840697a9bad75a
1c8988c1fe121314c582421843f0d77d7ef8ddd7
85064 F20110403_AACIPQ zhao_d_Page_72.jpg
46684d6f6376ed7859c23c189f281b21
1313c8eda0f32f549d149cc8e8c7481e28cd65cc
41738 F20110403_AACIKS zhao_d_Page_13.pro
2f097584690cce9332659ec1afb90291
88a08964f8c0748058b7f77fdaaa93ff279e87fa
30071 F20110403_AACIUO zhao_d_Page_68.QC.jpg
9697141789c8b90a47dd4706efacec7c
c322399a3fbb7f37b14139c4ffcbfd30da2f90ae
F20110403_AACIFW zhao_d_Page_41.tif
99554646accc0a530b1443ef86e55491
de4a160a47e56c53c4ac3fd5158ec595c8ceeee1
92312 F20110403_AACIPR zhao_d_Page_73.jpg
951efa09e0712a907d94dc81e14a504f
0ab48fb5142d2469d3d53b089a3b7020015cb106
48659 F20110403_AACIKT zhao_d_Page_14.pro
6f6216f8a55e4217474ce0f9d1c0b862
58a0d3da72b5964a22f0c790b2d1bb15c4e17f6b
8783 F20110403_AACIUP zhao_d_Page_19thm.jpg
05fe96e7ad5d8d2258047df54c7b67ad
24526ff19e65ca8fcc5e68a9f08a0ff3f87b6530
F20110403_AACIFX zhao_d_Page_42.tif
a30e3075f2b28650269bdb2967cab846
26674f24b2900bc24ca7c4bcc498391b92d5c082
10150 F20110403_AACIPS zhao_d_Page_74.jpg
343ef320b008d78594374e39b4176c03
b901d8da538556c85b9b3d493074faff2154cfa1
37918 F20110403_AACIKU zhao_d_Page_15.pro
db8bbdf7441499cee56f90d97935dcca
e238efe66547e056cf7fb38288ec1e75996c7a9a
7304 F20110403_AACIUQ zhao_d_Page_73thm.jpg
25b0ce487150bcc386f46c2e97e312de
181eeac37a68fdebff0b77fd1388208e167b6c95
F20110403_AACIFY zhao_d_Page_43.tif
789ab6c7d6b2734ad1d3cf18bd55246c
6775564b5755b262a6fd42ed9d7e4c4e91d5bfab
105415 F20110403_AACIPT zhao_d_Page_75.jpg
13dc30e86a0e9a01166b87e1894a48dd
506cddcb06811047ccab6351eee640e155f62322
22820 F20110403_AACIKV zhao_d_Page_16.pro
0fc2741cec2ae0b894647c829c17de2d
86e6c9a8d7ae93d7da05f6a17736ce3da4fd8bd0
51460 F20110403_AACIDA zhao_d_Page_62.pro
f2be8f906e1215c30c87a833c7089985
e6aedb14271f207218b1addb57114f1219de324d
2226 F20110403_AACIUR zhao_d_Page_01thm.jpg
f12e41e8d08e84118c0d1eb3fe722ffd
becaa13e98accd1c0ecb9d6f92f8636cdc253bb3
F20110403_AACIFZ zhao_d_Page_44.tif
f017aafa4c85314d6cd7284b71ed7a9f
e22c7354d0776feaf00dfa048f6394aa0b38881a
126817 F20110403_AACIPU zhao_d_Page_76.jpg
fa82f4bfbdb1006aaede3344d68b9439
35b3955292837f470297d88bf7748e345eb99639
46030 F20110403_AACIKW zhao_d_Page_17.pro
2aea09982041f74e19d249027e57ff09
e812a4ed7bb5a38790bdaa952df32b3775d0eea2
93695 F20110403_AACIDB zhao_d_Page_47.jpg
a91cd9bbe49bbf1a5ef00e5a65f33410
1f35049e3cf24bc4400d97c67d1a214e20246b2b
19311 F20110403_AACIUS zhao_d_Page_12.QC.jpg
a096a9ba84fe9e9fa2ac76dcf093b589
44d8e0ec67f0802b48fe389f09bafd6189b36498
46362 F20110403_AACIKX zhao_d_Page_19.pro
6c04e645256d0ad8089b35b3c8530d95
5e6a010c65a7bf6dca366f4b233826f14d1e2649
F20110403_AACIDC zhao_d_Page_30.tif
18284309945a46601a54a9c23091615a
828a205f041444504c2795fa4f5fa41d4c4e3c24
31699 F20110403_AACIUT zhao_d_Page_48.QC.jpg
acb4dcd07959c76854378e7779578b2b
782cdbae163da3da78bcb7f02cbe516b57b5d367
121232 F20110403_AACIPV zhao_d_Page_77.jpg
164f5a90ac48ce578b882fea9e339345
f6c665ebebef6b57fb3b5edf5c2cc58e4ed8d186
1866 F20110403_AACIIA zhao_d_Page_20.txt
91b459b2507c86633c6520a2b41fd2ba
6f55604992580882364880e0fe155bc231a68ab4
46423 F20110403_AACIKY zhao_d_Page_20.pro
d45ee3dfb2606ec24975df333493246d
0fbbb9243dd0c67282699eac028d50e0d5854a3e
34686 F20110403_AACIDD zhao_d_Page_55.QC.jpg
a2623a39d61533af945a38f1b343e5cd
64075dfd227afab82f612a31a4c0716c39eb1ed4
5539 F20110403_AACIUU zhao_d_Page_04thm.jpg
da2201cb112ab9dea831d193db7bb55f
6b83fda085bc38da76d2e4a85576aecc8be7bcba
126153 F20110403_AACIPW zhao_d_Page_78.jpg
ca3bdb92457c5b9917e705d2945de966
f2efb18d2eefa29609eedcb2f31f817b9595d831
886 F20110403_AACIIB zhao_d_Page_21.txt
5365e25a8d369c991ca7b9e5ac2c29f4
93b0d233a1d131c30e5064e28cd797f724e98cb3
48919 F20110403_AACIKZ zhao_d_Page_22.pro
4b0bdba1090a9ea07012207387e8cae3
35e3e14738ff762caad83c514a1f2071bbce7a13
108441 F20110403_AACIDE zhao_d_Page_51.jpg
d486c1e8dbafe121ca04d292bae77507
5ec371ae0b1322f39eab513d00b04353b0f48664
28041 F20110403_AACIUV zhao_d_Page_58.QC.jpg
a55d9124a1c56c8511c287a3db173cdd
2fad905fa92466dde43b6d62fb1c29c56060814e
130947 F20110403_AACIPX zhao_d_Page_79.jpg
f9b1a75d52c9b14d1df419bca07d9b65
b6bcb0a4b88ea4398ea06ed9cdb91ed4e22792f0
2063 F20110403_AACIIC zhao_d_Page_22.txt
f6e43c749a897ef75d444c331fce3456
55869473477cea07e90a114dd4e3378e75b29f36
28913 F20110403_AACIDF zhao_d_Page_10.QC.jpg
166885e128f5eae43e769f3be37d6818
25c9b900bddce97da4e487137d8b29317286fee8
F20110403_AACIUW zhao_d_Page_20thm.jpg
7805539ab0b4918f6112808a3824b73b
e795742f19e30f6d4e982954a787a13aa25218ef
14696 F20110403_AACINA zhao_d_Page_80.pro
e3c4689916aa78c721a6af5d9a11fca5
06fa90610f18812eda8050f7623429efafd8c29f
34168 F20110403_AACIPY zhao_d_Page_80.jpg
15a9d4d948cfbf61e220dc9fec2b3737
97612f066be1929fca03e9d3ee410c41bd86c292
1894 F20110403_AACIID zhao_d_Page_23.txt
5d61c28780e70745377569377527e9bf
ac046a40889a48f220aa335fb28670ce1a2a0b34
2586 F20110403_AACIDG zhao_d_Page_50.txt
6f6a4da60e6a74c4e5cdc4b300f55283
710ed0ec17627b39a342e9782227fbda8cf8f728
26962 F20110403_AACIUX zhao_d_Page_33.QC.jpg
70c1ee8bae9f2e44e33adfee314c9501
98b2631b608eb4e4b9999c99d827d579decb57bb
38191 F20110403_AACINB zhao_d_Page_81.pro
f5e15a04cdd76cde407f0a7538771a47
39466ca7c49bd08ed329c02b6921843c4c1c4655
80224 F20110403_AACIPZ zhao_d_Page_81.jpg
760e369186d7e4a20ab264ab656d5cf6
f51e93bb6bf9a0fbc7cea0b13ddc2baf7e2ecf18
1934 F20110403_AACIIE zhao_d_Page_24.txt
2be3dd8e9d6bd8e12ad58b080a0e6e46
11ac1f25ad2490ff2fdcec7e66a131011a2523a8
F20110403_AACIDH zhao_d_Page_50.tif
ba2c4d1cd008a0c7d162c0764735c828
367ac8752e7220ef92b397c8ab9b619eb8371377
29964 F20110403_AACIUY zhao_d_Page_75.QC.jpg
f248803965c0451d7c4175e272806592
eff378cfda66620af5504a8f120fc95773a282c8
27056 F20110403_AACINC zhao_d_Page_01.jpg
493742be3965e157f04b3e46361ead59
057256073eed0b585eac26e3ced0be3695f93a67
1712 F20110403_AACIIF zhao_d_Page_25.txt
a5b7912facd9e1e8cd02757a6e50d14d
b099ae6adb4b2aa35a6936afc19fcb7292c402bc
47580 F20110403_AACIDI zhao_d_Page_18.pro
ce5ca41e42aa738a5d8b437d89f9dcc9
1b7d2baf5fbdf2962a19a98d31a064e90dfba898
8308 F20110403_AACIUZ zhao_d_Page_55thm.jpg
7525668b721d38f13ad9fbe49d286cb9
6245a6b791086cb28b8504b411702b0a120ba6be
4387 F20110403_AACIND zhao_d_Page_02.jpg
9311c0072d6134f6f22a0b1a9dee6a9c
8eaef09fc1b0ebdc1f5c9d08d050616a17112f05
1804 F20110403_AACIIG zhao_d_Page_26.txt
0c937aa0057d71991f9c4ac06a543fb7
b050e88750d5b36161aa3c3d4389798cd1a7d912
1051977 F20110403_AACISA zhao_d_Page_53.jp2
a2aca9304f7bd526d1d4526a4320a48b
1d37c77c4e1473e36afbd7722b2ebd0984e234b9
5439 F20110403_AACINE zhao_d_Page_03.jpg
f5a409154b1dba3d7541b435ef277834
ac74c9c3602dd0ee2db3c05576870d204e2112ea
1798 F20110403_AACIIH zhao_d_Page_27.txt
2badfd61c5cf3097fdc7fa12810436a4
0eaf5957abc6f8a90e508bfb70ba63fb05866514
793869 F20110403_AACISB zhao_d_Page_54.jp2
148cc1a43f082cd78cd580cd66031b55
b89c05ea5ecb50bebbbba17ee6b023deb412b2d3
1997 F20110403_AACIDJ zhao_d_Page_53.txt
1b86a38da08a8336f224b7d15ff31b52
d2b765a9e6a15b78eeea110c266c0196093be606
67178 F20110403_AACINF zhao_d_Page_04.jpg
186625421e5f20250f02bb9e9ceb1bae
14b99aada9df6b030d6a60fbce9d918d1ac9c76d
1884 F20110403_AACIII zhao_d_Page_28.txt
b6f7d2c0d00159237f60e5a981c26521
1acdcb1dce38d5aa1869ae76f1b3e3e663cb3f0f
1051957 F20110403_AACISC zhao_d_Page_55.jp2
ae2e84bed63e4ceb6c4d28f0804af5a3
ba4885c8ee39a6a48150887aa97e2ce52e47b579
102256 F20110403_AACIDK zhao_d_Page_71.jpg
b926fa4a7eb418ed731e43444f8171a9
1ee8cfe05bb9093bc842b675f4193c3b708ea5a5
31581 F20110403_AACIXA zhao_d_Page_17.QC.jpg
4797cbbba096d88639235cede3fa019d
2501f7436320cb8d5ec183eab42a197be18390bc
1967 F20110403_AACIIJ zhao_d_Page_29.txt
b764ae0fe375e6e168836c51c3a9bca2
3ddc1b2e0b95e54866e3b4977940d4d56fe9bf71
1037924 F20110403_AACISD zhao_d_Page_56.jp2
21e2c2a6b3aca00544e09fa18980d8b7
ee34d89666c833f4c10bc76f10516b2301bfc437
F20110403_AACIDL zhao_d_Page_78.tif
6b5ccc315c278b3a3e1b725dfb01b135
a00ab0bca1229cb15113cae577f192e91e93fcc3
79187 F20110403_AACING zhao_d_Page_05.jpg
6b5699d9da81f14db0774f68293e928b
3a10792634f62e9c8b72e55c88850d49b3777124
20573 F20110403_AACIXB zhao_d_Page_21.QC.jpg
78d9a25894867da9e158b8b620b48410
1b091a06058e49ea3940949cab93bd74b57d522c
1822 F20110403_AACIIK zhao_d_Page_30.txt
013a3b339cefd8e4b82c51454ceea6f7
e82543bcfc52bc2b4930d8c42b803d5e128693b3
1002416 F20110403_AACISE zhao_d_Page_57.jp2
8180a62a4f17220c83964c2c48bf9763
5cefb838ba6dbc8a668e6841cc2310bb93e049cb
51006 F20110403_AACIDM zhao_d_Page_43.pro
c6b1fbdd65917d61c27b4a29d1ea55ef
103644305fe8345bab90586af4484d80df989415
61490 F20110403_AACINH zhao_d_Page_06.jpg
6494a8c3bccca558865e16d9a5da0f0e
b94f07fa517a260420c908c9884a976df5a553d7
27453 F20110403_AACIXC zhao_d_Page_23.QC.jpg
3ad8eff915389e6285072a9df565c9dc
2b87cdb78436bb3897e09b48d0cffee0bd45a6c2
2261 F20110403_AACIIL zhao_d_Page_31.txt
7e3738b41c84a6fe4d668c0abb178e3f
d4057aadd8b59408f36cb44a1319eefd28e6a1dd
931112 F20110403_AACISF zhao_d_Page_58.jp2
c9cf5d8e0dd332b6fd0c397329dfb939
b8ecfa03f48ef60ad53f6d1ca1345f03b4462f59
7290 F20110403_AACIDN zhao_d_Page_27thm.jpg
9a99906db5fc31b8a653c06377e7be3b
a1d3c80b797209aa29c88797fa036b638d80a7a3
102836 F20110403_AACINI zhao_d_Page_07.jpg
7a38830d10bdde69945261e06628ce4c
07bbb9acfb82be0cd25908a25ecb4c7cf53a79dd
29692 F20110403_AACIXD zhao_d_Page_24.QC.jpg
ae7eee41c3002386dff0d7cce13dcef3
6edad8bae48e2200987a2da6c7744765675b5c6d
1689 F20110403_AACIIM zhao_d_Page_32.txt
85d287d87c3a759bae9536bb27a5b999
3d087188d4f7007eb8bc7e2ec48f7cfe569239f2
1035624 F20110403_AACISG zhao_d_Page_59.jp2
8144d63a3777e536cfff81644673d439
ce78a3d9a26202bc5963e11080844de0686e718e
F20110403_AACIDO zhao_d_Page_70.tif
209721f371b1a6faf36ccd37307d75fd
65568140edfb87c7529cc05e941d7022c1f4f638
26265 F20110403_AACINJ zhao_d_Page_08.jpg
15e148ef7852dea6e710149683536fb6
35ac66d4e246fdee70942f184065d5eea834de55
29909 F20110403_AACIXE zhao_d_Page_25.QC.jpg
f00bbc39723f248aa57d169fd3189674
72a57af56f4242fd59dc5485591326158684b184
F20110403_AACIIN zhao_d_Page_33.txt
6edaf8696a75e450f799f9ecdbdda7aa
17e31ce8ef4a4ad73721300633f3349657145c77
F20110403_AACISH zhao_d_Page_60.jp2
8385f361862f576650dc0d055a011177
09e4b3f5aad2700cfce2ab843764cdb3f689a79d
8077 F20110403_AACIDP zhao_d_Page_48thm.jpg
ad4683551dcd094f2a17b79960ec2642
3499e5e2f6884661c265a600c5d2cd3ab740f067
105855 F20110403_AACINK zhao_d_Page_09.jpg
071adabaaabc0729113e3b581b6221d1
f5c87126dfc82041bb9733436d762a0494a5d42d
27887 F20110403_AACIXF zhao_d_Page_30.QC.jpg
927507c2d12e446cc2e0d04a5865aa01
34649f1af077069fa0d53389a4fe9f56c3f4a468
1051902 F20110403_AACISI zhao_d_Page_61.jp2
937604306991b3e9453347c4f5442c16
c3b759b274f237d9e7c3dbdf519a76521b8002f0
52679 F20110403_AACIDQ zhao_d_Page_38.pro
a74369c0921c9f3a8821da42cdf06998
b70861330403dd230576152c93f39f1ee7c12f8c
108369 F20110403_AACINL zhao_d_Page_10.jpg
bf183fe7039d42706262c1cc604c3d93
32bd358007a3a579bebb5f033a6acdc48f1dfa4d
33644 F20110403_AACIXG zhao_d_Page_37.QC.jpg
8821aa7c65cd6c5cd078f90f7d60b5e0
df88c9c343feb951b1462464d3cf1fed398f8582
2011 F20110403_AACIIO zhao_d_Page_34.txt
28beb112b6476396f39e9d5df09a0561
3c0bcf3a2eb29f86bb3030c7061eea04018a6da8
1051979 F20110403_AACISJ zhao_d_Page_62.jp2
e5a55f3c4a1bb65e9117997447f93635
1752ace3b41e910d9e1d6a9da057c3a49d9c785e
2841 F20110403_AACIDR zhao_d_Page_05.txt
59c10df4623bc6bf80e93626b49d68ca
f2c7ac85d381d2e7db73080124d20a626a94f4a9
85217 F20110403_AACINM zhao_d_Page_11.jpg
ee4e577fc4691c083609d6f1c53a0476
4d8c29c4e690d0c96e614044080a3910fa31f4a0
5571 F20110403_AACIXH zhao_d_Page_40.QC.jpg
6d4808271b51441700775761adc67d9f
69ed667fd9df733579fc7e95e63205d96e6aaefd
2026 F20110403_AACIIP zhao_d_Page_35.txt
7ceabec61f6d213b63b48dc465be327d
7254dd0ea3d0f796e0daf715cb15e5c997acbf7b
1051984 F20110403_AACISK zhao_d_Page_63.jp2
91940bcc32d81ef1685b44060f99145e
731c0421e63e31b337cbf17f683ba801df9b4cff
61566 F20110403_AACIDS zhao_d_Page_77.pro
f4c4663e3e31a555ba0d9c2252fadb6c
3fc94c810d5e9fc96ba199feb6e84b059705d14b
58630 F20110403_AACINN zhao_d_Page_12.jpg
7709b0899ace67e32699492353a28fff
ecf0d56b3dbf0733d9b926adc894f24c66bcccde
33558 F20110403_AACIXI zhao_d_Page_43.QC.jpg
2f7cd421d9dc863d3a0c13f72242a016
eb6c491bb40ee1267ddec2d8c0514f2346d0c81b
1669 F20110403_AACIIQ zhao_d_Page_36.txt
2562d209117931412e61629e7d99ef4b
90a01b1b3b96f1ffe199e1c70af64dd1b41e2e7c
717378 F20110403_AACISL zhao_d_Page_64.jp2
7b4585ec803622987e25878941d95c5e
004025f93d55dfa4f7189385fad925221fb6c2f7
109106 F20110403_AACIDT zhao_d_Page_19.jpg
3f95221964caa034195c7f8ef174abcc
adc15e25168079e13dfb0cd914f69d81bc6c359c
87434 F20110403_AACINO zhao_d_Page_13.jpg
112cb6dc453d6ad1ebb78ee6c8d36bb2
fcf4a6f829a0bd0a8d4b13fea68dbe585910e7ac
30172 F20110403_AACIXJ zhao_d_Page_45.QC.jpg
5559da8b0c0b1b0fd1c2141933b7a11e
1189c64961d27da683774eb46b6693e8664e392c
1051931 F20110403_AACISM zhao_d_Page_65.jp2
5d9ab806b12d5e9cef3ba59e593e6aa4
8b9c3c5de714f80ad11f98cacd29b0289f6f62b5
2020 F20110403_AACIIR zhao_d_Page_37.txt
f68829e4dfa39c0d387830271bf56281
ebdeeac9874c8dfcf2f566727463e98c6073e15b
6696 F20110403_AACIDU zhao_d_Page_72thm.jpg
3ac4d692c2ae6158442ead41563ec937
947aeae2582dacc677e64fe31cd8341083f2ab80
97100 F20110403_AACINP zhao_d_Page_14.jpg
20fabbefe7bd50e950a36abd27b51e14
790da91084839838292511bb82fe0b7d71de8d06
32805 F20110403_AACIXK zhao_d_Page_46.QC.jpg
1b1d310308283a8ce0479724ea34968a
46b2c77ccf09553d3a5d45fad49a3753b3deba17
1051963 F20110403_AACISN zhao_d_Page_66.jp2
0f1a20d2368a55b9266f1b10c142c8c8
23b008aa12c6995bc22ef608fd90905f1210eb23
2060 F20110403_AACIIS zhao_d_Page_38.txt
51ef6df6f330ea5f2c065ba5db708cb1
03e3416159e6c02dad4e62de10381126b1e70b25
8112 F20110403_AACIDV zhao_d_Page_45thm.jpg
2699a8f3792f405a79db228cef5de603
12455c382f7c23308d2cbc4c1aab33e7f9358e29
78281 F20110403_AACINQ zhao_d_Page_15.jpg
da44e9aabb51c8a8de4bc5f5fec4231f
07e7f7cc7f7dbfb20bdb2fd5247bff06a47bd3a7
33458 F20110403_AACIXL zhao_d_Page_51.QC.jpg
f94863b8e189939cbf17ca2b64743f27
a2afe07bfdeec16980bc1d66bfaa83b192445f1c
843003 F20110403_AACISO zhao_d_Page_67.jp2
11aa1ba74531da73b6cb1f0f0add5b6e
db9fd5aab3bf0885c19ae357f00b5b90e10cf342
2392 F20110403_AACIIT zhao_d_Page_39.txt
20103f7517664a8b7cb145aa4bd25b2a
8d4204b47020248bf73d70b304ff9bbbbfac52d2
1566 F20110403_AACIDW zhao_d_Page_54.txt
e123bff346c9c25d239c44de4e2be5bf
f65f10d41bab3a0fc771b280ee798456a9c8bc8b
49084 F20110403_AACINR zhao_d_Page_16.jpg
3dfb33c712c103ccde39434a8658b246
43d1f7e4f694c47363888bda47636bb5deff897d
32957 F20110403_AACIXM zhao_d_Page_52.QC.jpg
59816fce194bea111bbfc3ebc46395f5
6a85c7bf1272fd75d59165a6555bae4d4c64639c
942063 F20110403_AACISP zhao_d_Page_68.jp2
681e9ec790db8a2d2413ea25767f1fff
50c91f8bbdb25d116a4519f1d1d9f06481cef6ea
290 F20110403_AACIIU zhao_d_Page_40.txt
dd61491cae8971af47d8ec30b6d001d9
478e966736b86cf10f4213819f2ad8728e460d2b
2374 F20110403_AACIDX zhao_d_Page_66.txt
a91280c67017e198cca47ddb7af3c500
e790a970a3f91ff04f31f3f008b3354ee76807a5
98750 F20110403_AACINS zhao_d_Page_17.jpg
e5af9d35ca78796ef3661455a79eb86a
15c8ec01109f6bd4adff2e79304ae3081651c16c
33832 F20110403_AACIXN zhao_d_Page_53.QC.jpg
4fe603f8296718193b0409c54dd9943b
3034b34c1de01b9c8ad62f73cbf7a460ba7748b2
F20110403_AACISQ zhao_d_Page_69.jp2
f74c30311137ff9a28d04659f9a38371
6071e8cf31bc9608f1fd4901433f4f3c887b8be5
1864 F20110403_AACIIV zhao_d_Page_41.txt
fb2af7b597c9076d37699f31562ed183
60ab831e00f5daab41cf22ecc783267c9906b783
22434 F20110403_AACIDY zhao_d_Page_21.pro
83f5a4bd4848ec89cf0ada675f408805
b97250de633fafff610fbc86fd39c04c1c3bf30f
26690 F20110403_AACIXO zhao_d_Page_54.QC.jpg
4b373f0d46ddda9210edb09a1246673a
c10665edf0907f0bad1e56119430c2e484c76efe
1051956 F20110403_AACISR zhao_d_Page_70.jp2
2f593a68ee2a0605a2fd9713bd9bbd1c
744ef13ed0381a44c7a088db9aec228beed2311a
F20110403_AACIIW zhao_d_Page_42.txt
6686700a02d2da97d6bde20fb6e9cd2d
dec07ca57a2397262a1f83466fa5b4186f4a4eb1
7972 F20110403_AACIDZ zhao_d_Page_44thm.jpg
fa463bb04d1eb039c44b98ad55b87817
f8e3f5c2f4a36cd2188ae001bf3b37feee1539fb
95683 F20110403_AACINT zhao_d_Page_18.jpg
0032b4a41dbde17bfb8a948f58fb4c77
a17237d42a96813f65479c24fe09ef8faae2ba14
34254 F20110403_AACIXP zhao_d_Page_61.QC.jpg
8c8d093900d2bd98ad055504e689c8dd
4d372aae2c5ec74d61f867b781854ba8b039acf9
F20110403_AACISS zhao_d_Page_71.jp2
cbf9b9c43a892454f360312e38a35c07
1e7fa7dc765ad162052f9e44d71583e763493be0
2045 F20110403_AACIIX zhao_d_Page_43.txt
038796e4250348f708c4e5bcdf5661bf
1fcd5a4fe587945d62e48c1cb7d3000c814e010c
95890 F20110403_AACINU zhao_d_Page_20.jpg
71044970783a5ffc0e2248e8709e898f
92db85602ebbc0112655e0a6ae16e99e13af11a7
32392 F20110403_AACIXQ zhao_d_Page_63.QC.jpg
bed0b14940a86154e7bf5679b5b377cf
ffa1d46fec8a0caf1de87da45ccf3e0a6b4e663c
934526 F20110403_AACIST zhao_d_Page_72.jp2
44d02128479741dc5317212e9b54fafd
bb4e4ce7be0d7387b6428a7bbb9e2b9060a055b1
1952 F20110403_AACIIY zhao_d_Page_44.txt
8ca2e32c085d4072cf72128e60cf95bd
d7547b2c4e68b93c5d4822d23fbc0b0b640402e0
60497 F20110403_AACINV zhao_d_Page_21.jpg
22888659e3bd785c660cdee9ecfe05ad
0ccd0243c948ce78ddd81b7990589e835c918196
F20110403_AACIGA zhao_d_Page_45.tif
7ac8f8c174ceff61eef8dbdac7ad6434
661464e8d21e1e2d0ef2317a32aea1aafa4a32c8
33175 F20110403_AACIXR zhao_d_Page_65.QC.jpg
61ccae74a4129fc35c3ffb3ab61044db
18f7f519b1c30bce7034618c7de8b3c4d03f44c6
999816 F20110403_AACISU zhao_d_Page_73.jp2
b70a5fc0a5904618071630df1af1a82f
0eb929d35515a726e00765498d1a5940ec933de5
1586 F20110403_AACIIZ zhao_d_Page_45.txt
e7f56f7e69234789ca3340e1951c3e9d
810a54a85be45667b690e97f277b5988be686b3c
96601 F20110403_AACINW zhao_d_Page_22.jpg
98db8de305befa3359784f0b465583dc
50ca5145d01e32a399f2701b88089b67fa9b9f7b
F20110403_AACIGB zhao_d_Page_46.tif
ea00858f6e67eeafab476a845ca6425e
19d1fac02a593d5617b13299bb83160407e1d07f
32754 F20110403_AACIXS zhao_d_Page_70.QC.jpg
96494c0135632f9963c5e4e214d7f6bd
39b44d329b1c18dc9c2fbb18ee28908a6ac8596e
84073 F20110403_AACISV zhao_d_Page_74.jp2
aae7eef603253e958fbc6b173a0b5ae0
490df4550a8af8015b5d723b8dc8bf5b25089a10
83907 F20110403_AACINX zhao_d_Page_23.jpg
d179331d200da99a3dc4cc3640190416
1498a4942140db3ff4ccd5dcf3d5fbe403bc6745
F20110403_AACIGC zhao_d_Page_47.tif
1bb07760a49accd198d5f09780fd11d7
e1734045d89474a135039a719e9797eacf6386c5
35511 F20110403_AACIXT zhao_d_Page_76.QC.jpg
0022b34fbd6b051e0d7e5a0acdef58a0
01df5145a1ab5bd7bd13cf4e81b9aed004765974
F20110403_AACISW zhao_d_Page_75.jp2
132dba09b39fe0eda9271929be79e2b8
208300d034e9a2e4e160a99c0dfb8de07d68fa86
42916 F20110403_AACILA zhao_d_Page_23.pro
d090e5e8632724b141439a503bc762b3
4e5528abb497b730118f5fc80aeb564ab111000a
86328 F20110403_AACINY zhao_d_Page_24.jpg
aa4acf40899d5e0eee5055823392ac66
11a8f50b8c68bfa419ce69e9f6469ddaad8bbf83
F20110403_AACIGD zhao_d_Page_48.tif
22cefc30b98e45f80e3c9cca82456f5d
93870bc1e29f5760b1c49464c247217ad116502c
33658 F20110403_AACIXU zhao_d_Page_77.QC.jpg
427cf4f083291d382b7d77ee4fff5f82
e0c7af3ff4c4cf9eb55d18a7f99a183637f4555d
1051880 F20110403_AACISX zhao_d_Page_76.jp2
ed254b791b2d3564c28db8df5bd99dab
634e9690d36c2547599a9f3fae2ec806080a01a7
41250 F20110403_AACILB zhao_d_Page_24.pro
e8164e2114ba5e5063a5562d30c62914
9e00331d6b48338212d182f5e7c98767fc59b3e8
89447 F20110403_AACINZ zhao_d_Page_25.jpg
4506c49908305af74ad9aeb3d55db3e4
26dba010133d62e63ca362393ca784d430ffa13d
F20110403_AACIGE zhao_d_Page_49.tif
18a357c6928c6c51f336dc912aae8d32
4a778a117f109e53ce0e0568f3e64b8fbbafc737
562 F20110403_AACIXV zhao_d_Page_02thm.jpg
13f8ae43115e1d43d158b130631092e6
9132e1068cf5b117766796f40122fa911a2fcb14
43163 F20110403_AACILC zhao_d_Page_25.pro
e50c0a9ae7dbc0ed0cde8385c968db56
c26fb1b2afb3ff33b27074167b3044a109133891
F20110403_AACIGF zhao_d_Page_51.tif
2a3657de467c56feb6a1331019c94c72
8fb9b144713475644ea6ce7d1b967cf9eed1b9dc
6640 F20110403_AACIXW zhao_d_Page_11thm.jpg
74996053f388580deebc928915f30437
a6b1a8e64d8ce418da42d3d0726f8b60cfa382d7
1051953 F20110403_AACISY zhao_d_Page_77.jp2
39e0be9f95dd2330bfd495a8535f2f51
348f5cc6fee4d2dd64ce325f0f47143b8707b3f3
249699 F20110403_AACIQA zhao_d_Page_01.jp2
1a367b17d29e4357e3383d366745fd4b
cfcde13dc1b85f12a39c32d38e5fd33d26ba410a
41831 F20110403_AACILD zhao_d_Page_26.pro
be42861141290bc8853035e20a9ef848
e254b7e0b6b62709c61b4e989ac01860fee3e927
F20110403_AACIGG zhao_d_Page_52.tif
f9f2df9f6e79f98ce40dcaef968f7722
04f74f1b7b1d8ad1ec33dc8c86ee18b1e74a4107
4895 F20110403_AACIXX zhao_d_Page_12thm.jpg
b028ad316ebf5707d4359599d5d24286
afadd5d81e1f9c9e16052b4e798a9a6224a44f40
1051917 F20110403_AACISZ zhao_d_Page_78.jp2
7ce0eef0497e45ae939f4bf89d6fc2f8
0fcccced8fc20c52f8f02b17cd9ea5a952135e06
24237 F20110403_AACIQB zhao_d_Page_02.jp2
d70e399695dc0197be68d208313c0bd1
bbf4f3d719bd64ae130ccc1a4ce2ae9e0bca4a28
41162 F20110403_AACILE zhao_d_Page_27.pro
60f81f17864499a08fcc57963551abaa
54ddeb2e47f57880249f9ba702c7ba5f3ff5854d
F20110403_AACIGH zhao_d_Page_53.tif
a1278bd19e72b76cc287ed7e7fb4f47d
90ddcdf10bb321f5dfebdc4939781e9313bea4a1
6852 F20110403_AACIXY zhao_d_Page_13thm.jpg
ec8544562a6cbed683be6958f71b9146
7d0b9b9d19a5f1c45529539f3ce32407f94894d3
47560 F20110403_AACILF zhao_d_Page_28.pro
868d0cc83683ada92ee6293dd45430aa
10f615f72e3a17d376efe08aeaf8810a8e4ffa68
F20110403_AACIGI zhao_d_Page_54.tif
107923d99230d9a01d5ca04467a7cecc
a96c625b55d83bfa60e91a59204f168b0d4fa4b3
35262 F20110403_AACIQC zhao_d_Page_03.jp2
1307d7615cda8f8182ebaa1e3b3f8833
f24a01c508e421532798af4407af36412f12481a
7898 F20110403_AACIXZ zhao_d_Page_14thm.jpg
ce3309d26035d1513f4f5ea27922136f
25118ec8b862277b7073b51e1e05db9c66180cfe
25338 F20110403_AACIVA zhao_d_Page_32.QC.jpg
8fd571e98a3eebcbc04535626406250a
7f2a538907e6b974b79681bc839320a4dd320ff5
49917 F20110403_AACILG zhao_d_Page_29.pro
f456136ae865c28781827b319de2722d
73507cf4124b422d0682ee3ccb27ef6c8ab0844d
F20110403_AACIGJ zhao_d_Page_55.tif
cc3d031cea8ce05a484d99f7e1885a66
46124cff1c4878acfdd131e132c9ec0626897245
698080 F20110403_AACIQD zhao_d_Page_04.jp2
ced0ec5537b2858891fdbce2898acfab
fffcbbc80b0cb9f5283367a65ecbe2a69f915769
32139 F20110403_AACIVB zhao_d_Page_47.QC.jpg
db6cfd0e73df19394292d3cb23f40ee3
3bfb4815eb7c3d1a7e7e74f6fd62c9668316fd04
43018 F20110403_AACILH zhao_d_Page_30.pro
1eed8612c4689e1668c136a90210a611
3a9d4bd58f64037454fd059d42e8c532243e0068
F20110403_AACIGK zhao_d_Page_56.tif
8f154d8bd1eac45778c38832fa1aa4d3
9b0e038efa8d1cbb249692a1915324006fb3ec76
770692 F20110403_AACIQE zhao_d_Page_05.jp2
f96852ce524e9bf463b76641c66ceb2e
c246af07eac691813a5883aaab6576cb0883eb80
29969 F20110403_AACIVC zhao_d_Page_26.QC.jpg
44373604d3254f81f8e8bd9aa47c52de
50ec9a32ea7e1a26e803d558e7eecaa7dbaf4e01
43871 F20110403_AACILI zhao_d_Page_31.pro
82f1007b0fc437cae757bcdfecdc45d0
92cdc2f38aab15229f4b5b46d6cf523e3ed2abde
F20110403_AACIGL zhao_d_Page_57.tif
3863f6370b62452385fa2e18698e2d89
5dcc9d356d387c9a1beb5ba7c5ef5db4ce30888f
590109 F20110403_AACIQF zhao_d_Page_06.jp2
15874eeecd2fe2c09e58ce03cbce2b44
cc7a6ccca35e56450ce8634b3305f7056d8d903a
36301 F20110403_AACIVD zhao_d_Page_79.QC.jpg
ff418a55f9b793b41f9f26493d1ebbf8
0ff0dc934c494b41a12f56fb3d3bee80899a23f8
33036 F20110403_AACILJ zhao_d_Page_32.pro
fe7927884ebba3ba2cfd89aa8247d108
6c872558ffcf604576f1a5e2a62d0034d5d4cc23
F20110403_AACIQG zhao_d_Page_07.jp2
8cbe6455f77288924d7532ef82e20088
34d90bfe9d01e4ded4edff9d09f7c9004d0e7025
8623 F20110403_AACIVE zhao_d_Page_38thm.jpg
00d99cb62bf64968a6eb97623c83808f
583230360d1d6202abe417b33c55fb2fc1517033
39309 F20110403_AACILK zhao_d_Page_33.pro
ed1e1489d239509aa967b95383dc2baf
98da6a3a8e91017f9568e6ee988728db3b53bc2d
F20110403_AACIGM zhao_d_Page_58.tif
d311f2afd13bbd1e268f3d3d7f600289
53105edb9c4f333a35b6b7cd88869aa6a5b6c3cb
233336 F20110403_AACIQH zhao_d_Page_08.jp2
a98948ee4ef2a8e62c2ad4533aa073d5
2748ea9ecb756e334a543f577f1a6b41322521b8
7870 F20110403_AACIVF zhao_d_Page_50thm.jpg
56fae57085d45db13f47afa148d4f807
34bdf586d27cca676bcad5b54aff7cc045d7e1ea
38692 F20110403_AACILL zhao_d_Page_34.pro
bbc93773327e01d9151b355be7a2f48d
d4b75ccf92071ffa998ee3782713e1ebe44067c4
F20110403_AACIGN zhao_d_Page_59.tif
41ce4ef80889f8d815816131cdae9923
24c00405a752dfc7d365cb1367088cff879f3eb2
F20110403_AACIQI zhao_d_Page_09.jp2
e80b66cefcab82a3f7271e1b41df85b9
a41bf5108251bca7e5f1af9088f6915d7b4202e0
7594 F20110403_AACIVG zhao_d_Page_30thm.jpg
173ac82d0bdce85f03bb938109f276a9
2bef53aec89dbab22404809085b077b658e86cd5
50615 F20110403_AACILM zhao_d_Page_35.pro
b23bc9d86cbb67ffa842292c93990f6d
da64026a90c5d621f622f558cf0ec7dfacd61bf0
F20110403_AACIGO zhao_d_Page_60.tif
297b4001f0f5f7bedf34cd55fd92e8a7
17bf156378493bc92e3330be9f212c42096ea805
1051911 F20110403_AACIQJ zhao_d_Page_10.jp2
c32f75a3bb1149d57900cddd7ab01384
1f34d5cc8414cc5e77c34b4cc7f4a8f0d429c4f4
9914 F20110403_AACIVH zhao_d_Page_80.QC.jpg
fa27daab0f21811ffaac0679d3a0df47
47822b27c5d6c6eac6d297d4ffb53fc9cec5a727
35825 F20110403_AACILN zhao_d_Page_36.pro
66f4f0fc9a04c20140f4ce7165e9fa14
b09c9aa0f0ce014cfac63f2a76b5a8f59e220d67
F20110403_AACIGP zhao_d_Page_61.tif
4d2f4a459ab0a4cd5e36f0ca363bea39
740c9e804d6af561c2e0b10836a767d469249c88
910212 F20110403_AACIQK zhao_d_Page_11.jp2
8a495a91ef73ba35ef79653f98d89033
557c2a58300f65bd46f6d36a051b9c961082144e
8240 F20110403_AACIVI zhao_d_Page_70thm.jpg
688dbc8bbd9ee61ba949d6a773176ea7
bb3f5e7607d278017beaa5d37837a280e7faf5da
51317 F20110403_AACILO zhao_d_Page_37.pro
4296c844493d42f1902a8bdcd36c51e0
5a94bcac18d5c41aa74614d8df3d89ccd51151d7
F20110403_AACIGQ zhao_d_Page_62.tif
b12fa2887669423a602978d28f6c9334
dcfb5796b2e096f0949654282232a6363bb13a57
625732 F20110403_AACIQL zhao_d_Page_12.jp2
f904a3804485232fc22946e7e08aaf95
892143150ed6fed7136fc95842060238fd3046d3
4534 F20110403_AACIVJ zhao_d_Page_05thm.jpg
837f3ffafa54e1472af39f9bed1bee85
235ccef0b817b3333500b7c1e0829613160c0e17
57968 F20110403_AACILP zhao_d_Page_39.pro
02cb3d85f5c37cbba7b505f63cf1663b
72f40c22ab0b8d874dce87cde76870bbe603ffd9
F20110403_AACIGR zhao_d_Page_63.tif
1ae26cac7ac0a66cfd7f69fe03314477
37d4fbf069c05c39087b6dbc81c34e91e7e1cf40
940341 F20110403_AACIQM zhao_d_Page_13.jp2
44d81edcfa8de5022a45f8f6758f0702
ff3f3f3d7d93a3c5c5418378c462fb6d58c4ad18


Permanent Link: http://ufdc.ufl.edu/UFE0013600/00001

Material Information

Title: Analysis of in vitro and in vivo function of total knee replacements using dynamic contact models
Physical Description: Mixed Material
Language: English
Creator: Zhao, Dong ( Dissertant )
Fregly, Benjamin J. ( Thesis advisor )
Banks, Scott A. ( Reviewer )
George, Alan D. ( Reviewer )
Kim, Nam-Ho ( Reviewer )
Sawyer, W. Gregory ( Reviewer )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2006
Copyright Date: 2006

Subjects

Subjects / Keywords: Mechanical and Aerospace Engineering thesis, Ph. D.
Dissertations, Academic -- UF -- Mechanical and Aerospace Engineering
Genre: bibliography   ( marcgt )
non-fiction   ( marcgtt )
theses   ( marcgt )

Notes

Abstract: Despite the high incidence of osteoarthritis in human knee joint, its causes remain unknown. Total knee replacement (TKR) has been shown clinically to be effective in restoring the knee function. However, wear of ultra-high molecular weight polyethylene has limited the longevity of TKRs. To address these important issues, it is necessary to investigate the in vitro and in vivo function of total knee replacements using dynamic contact models. A multibody dynamic model of an AMTI knee simulator was developed. Incorporating a wear prediction model into the contact model based on elastic foundation theory enables the contact surface to take into account creep and wear during the dynamic simulation. Comparisons of the predicted damage depth, area, and volume lost with worn retrievals from a physical machine were made to validate the model. In vivo tibial force distributions during dynamic and high flexion activities were investigated using the dynamic contact model. In vivo medial and lateral contact forces experienced by a well-aligned instrumented knee implant, as well as upper and lower bounds on contact pressures for a variety of activities were studied. For all activities, the predicted medial and lateral contact forces were insensitive to the selected material model. For this patient, the load split during the mid-stance phase of gait and during stair is more equal than anticipated. The external knee adduction torque has been proposed as a surrogate measure for medial compartment load during gait. However, a direct link between these two quantities has not been demonstrated using in vivo measurement of medial compartment load. In vivo data collected from a subject with an instrumented knee implant were analyzed to evaluate this link. The subject performed five different overground gait motions (normal, fast, slow, wide, and toe out) while instrumented implant, video motion, and ground reaction data were simultaneously collected. The high correlation coefficient results support the hypothesis that the knee adduction torque is highly correlated with medial compartment contact force and medial to total force ratio during gait.
Subject: adduction, computational, contact, dynamic, gait, in, instrumented, knee, total
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 81 pages.
General Note: Includes vita.
Thesis: Thesis (Ph.D.)--University of Florida, 2006.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aleph - 003614801
System ID: UFE0013600:00001

Permanent Link: http://ufdc.ufl.edu/UFE0013600/00001

Material Information

Title: Analysis of in vitro and in vivo function of total knee replacements using dynamic contact models
Physical Description: Mixed Material
Language: English
Creator: Zhao, Dong ( Dissertant )
Fregly, Benjamin J. ( Thesis advisor )
Banks, Scott A. ( Reviewer )
George, Alan D. ( Reviewer )
Kim, Nam-Ho ( Reviewer )
Sawyer, W. Gregory ( Reviewer )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2006
Copyright Date: 2006

Subjects

Subjects / Keywords: Mechanical and Aerospace Engineering thesis, Ph. D.
Dissertations, Academic -- UF -- Mechanical and Aerospace Engineering
Genre: bibliography   ( marcgt )
non-fiction   ( marcgtt )
theses   ( marcgt )

Notes

Abstract: Despite the high incidence of osteoarthritis in human knee joint, its causes remain unknown. Total knee replacement (TKR) has been shown clinically to be effective in restoring the knee function. However, wear of ultra-high molecular weight polyethylene has limited the longevity of TKRs. To address these important issues, it is necessary to investigate the in vitro and in vivo function of total knee replacements using dynamic contact models. A multibody dynamic model of an AMTI knee simulator was developed. Incorporating a wear prediction model into the contact model based on elastic foundation theory enables the contact surface to take into account creep and wear during the dynamic simulation. Comparisons of the predicted damage depth, area, and volume lost with worn retrievals from a physical machine were made to validate the model. In vivo tibial force distributions during dynamic and high flexion activities were investigated using the dynamic contact model. In vivo medial and lateral contact forces experienced by a well-aligned instrumented knee implant, as well as upper and lower bounds on contact pressures for a variety of activities were studied. For all activities, the predicted medial and lateral contact forces were insensitive to the selected material model. For this patient, the load split during the mid-stance phase of gait and during stair is more equal than anticipated. The external knee adduction torque has been proposed as a surrogate measure for medial compartment load during gait. However, a direct link between these two quantities has not been demonstrated using in vivo measurement of medial compartment load. In vivo data collected from a subject with an instrumented knee implant were analyzed to evaluate this link. The subject performed five different overground gait motions (normal, fast, slow, wide, and toe out) while instrumented implant, video motion, and ground reaction data were simultaneously collected. The high correlation coefficient results support the hypothesis that the knee adduction torque is highly correlated with medial compartment contact force and medial to total force ratio during gait.
Subject: adduction, computational, contact, dynamic, gait, in, instrumented, knee, total
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 81 pages.
General Note: Includes vita.
Thesis: Thesis (Ph.D.)--University of Florida, 2006.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aleph - 003614801
System ID: UFE0013600:00001


This item has the following downloads:


Full Text












ANALYSIS OF IN VITRO AND IN VIVO FUNCTION OF
TOTAL KNEE REPLACEMENTS USING DYNAMIC CONTACT MODELS
















By

DONG ZHAO


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2006

































Copyright 2006

by

Dong Zhao

































To my parents, Xinsheng Zhao and Yaqing Zhang















ACKNOWLEDGMENTS

I would like to express my gratitude to my advisor, Dr. Benjamin J. Fregly, for his

patient guidance and endless support throughout my doctoral studies.

My thanks also go to Dr. Scott A. Banks, Dr. Alan D. George, Dr. Nam Ho Kim,

and Dr. W. Gregory Sawyer, for serving on my committee. I thank them for providing

valuable comments and help in the progress of my research. I thank Dr. Victoria Good,

from Smith & Nephew, Inc. Memphis, TN; Dr. Darryl D. D'Lima, from the Shiley

Center for Orthopaedic Research & Education at Scripps Clinic, La Jolla, CA; Dr.

Hideyuki Sakoda, Nakashima Medical Division, Nakashima Propeller Co., Ltd., Japan;

and Kim H. Mitchell, from the Biomotion Foundation of West Palm Beach, FL, for

providing me data used in this dissertation.

I would like to thank all the students from the Computational Biomechanics

Laboratory. I appreciate Carlos Marquez, Jeffery A. Reinbolt, and Kelly Rooney

proofreading my manuscripts. I thank Yi-Chung Lin and Jeffery A. Reinbolt for their

input during the Matlab code development. I especially appreciate the discussions with

Carlos Marquez on all kinds of issues.
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TA BLE S ......... ..................... ....... ........... ....... ............ .. vii

LIST OF FIGURES ......... ......................... ...... ........ ............ ix

ABSTRACT .............. ......................................... xi

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

B a c k g ro u n d ................................................................. .................................... 1
Specific A im s........................................................ 2
D issertation O organization .................................................................. .....................2

2 COMPUTATIONAL PREDICTION OF KNEE ARTHROPLASTY DAMAGE
USING A PIN-ON-PLATE WEAR FACTOR AND EVOLVING SURFACE
G E O M E T R Y ...................................... ................................ ................ 5

Introduction..................................... ........................... ..... ..... ........ 6
M eth o d s ........................................................................................... 8
Computational Framework for Damage Prediction ...........................................8
Analytical Validation of Computational Framework ................ ...................14
Experimental Evaluation of Computational Framework............... .................. 16
R e su lts ...................................... .......................................................1 8
D iscu ssio n ...................................... ................................................. 2 3

3 IN VIVO MEDIAL AND LATERAL TIBIAL LOADS DURING DYNAMIC
AND HIGH FLEXION ACTIVITIES.................................... ........................ 29

Introduction..................................... .................................. ......... 30
M eth o d s ............................................................................. 3 1
D ata C o llectio n ............................................................................... 3 1
Model Development .................. .....................................................33
R e su lts ...........................................................................................3 5
M edial to Total Force R atio ........................................ .......................... 35
Contact Pressure and CoP ............................................................................35









D isc u ssio n ............................................................................................................. 3 8
In V ivo K nee Load D istribution...................................... ........................ 38
Choice of M material M odel ................................................................ ...............40
C oP E rror A naly sis........... ........................................................ .. .... ..... .. 4 1
D ata Synchronization A analysis ............................................. ............... ... 41
Difference between Overground and Treadmill Gait ............. ................43
Sensitivity Study of Contact Force Prediction to Kinematic Errors....................44
C onclu sion ......... ... ............... .................................... ...........................46

4 THE RELATIONSHIP BETWEEN THE KNEE ADDUCTION TORQUE AND
MEDIAL CONTACT FORCE DURING A VARIETY OF GAIT PATTERNS ......47

Introduction ........................... ...................................... ...... ........ .. 48
M materials and M methods ....................................................................... ..................49
D ate C collection ........... ............................. ... ...... ............... ... ............ 4 9
External Knee Adduction Torque......................... ........................... 50
M edial C contact Force ................... .. ...... .................. ....... .......... .....51
Statistics A naly sis........... ...... .................................................... .. .... .. .. 52
R esu lts ......... .. ...... .......................................................................... ......... 53
D isc u ssio n ............. ........... ................. ......................... ................................ 5 5

5 CONCLUSION..................... ..................61

LIST OF REFEREN CES ........................ ............................ ..............................63

B IO G R A PH IC A L SK ETCH ...................................................................................... ..69















LIST OF TABLES


Table p

2-1 Percent RMS error over all cycles used in the damage prediction process. For
update interval, mc means millions of cycles, % refers to percent of plate
thickness. .............................................................................19

2-2 Comparison of experimental and predicted damage depths and areas (variable
step updating with nonlinear material model) for the one insert whose surface
geometry was measured using a coordinate measuring machine. Results for
other updating and material model combinations were similar. ...........................21

2-3 The surface updating methodology, material model and creep model should be
used to achieve desired goals. Goal 1: Predict wear volume at any time point.
Goal 2: Predict wear volume, damage depth, and damage area at any time point.
Goal 3: Predict wear volume, damage depth, and damage area only at the final
tim e p o in t ...................................................................... 2 7

3-1 Medial-to-total peak pressure (Pmax) and average pressure (Pave) for gait, stair,
lunge, and kneel activities using linear and nonlinear material models. For gait
and stair, results are reported at maximum load. For kneel and lunge, results are
reported for mean load with additional results for minimum and maximum load
over the interval indicated in parentheses. .................................... .................38

3-2 Medial-to-total force percentage, and anterior-posterior center of pressure error
for gait, stair, lunge, and kneel activities using linear and nonlinear material
models. For gait and stair, results are reported at maximum load. For kneel and
lunge, results are reported for mean load with additional results for minimum
and maximum load over the interval indicated in parentheses. ............................39

3-3 Sensitivity study of the data synchronization......................................................42

3-4 Test cases of sensitivity study of contact force prediction to kinematic errors........45

3-5 CoP location as a function of locked medial-lateral translation or varus-valgus
ro tatio n v alu e ...................................................................... 4 5

4-1 The average and standard deviation of the peaks of measured total force,
predicted medial force, and calculated adduction torque.........................................53









4-2 The average and standard deviation of the correlation coefficients between the
internal knee abduction torque (i.e., negative of the external knee adduction
torque) and medial contact force and or medial to total force ratio for a variety
of gait patterns. (p < 0.001 for all the cases). ................ ............................... 54

4-3 Student's t-test results of the subject's normal walking to the other walking
pattern s. .............................................................................56















LIST OF FIGURES


Figure page

1-1 Thesis concepts. ................................................................3

2-1 Experimental measurements and the computational model to predict total knee
anthroplasty damage. A) Experimental wear factor measurement. B)
Computational damage prediction. C) Experimental damage measurement. ............7

2-2 Computational framework of the damage prediction with surface geometry
updating ..................................................... ........................... 9

2-3 Cylinder-on-plate contact model with surface updating. ........................................15

2-4 Wear depth and area predictions of the cylinder-on-plate model. A) Wear depth
with fixed step updating. B) Wear depth with variable step updating. C) Wear
area with fixed step updating. D) Wear area with variable step updating. ..............19

2-5 Cross section of the worn plate of the cylinder-on-plate tests with different
updating methodology. A) No update (Interval 5e6, 1 simulation). B) Fixed step
(Interval 2.5e6, 2 simulations). C) Fixed step (Interval 2.5e5, 20 simulations). D)
Variable step (Threshold 5%, 7 simulations). E) Variable step (Threshold 0.5%,
19 sim ulations, 8 updates) ............................................... ............................. 20

2-6 Simulated and gravimetrically measured volume loss of the tibial inserts due to
mild wear. Simulation error bars indicate + 1 SD in wear factor measurement. .....21

2-7 Damage depth and area predictions of the TKR. A) Damage depth with fixed
step updating. B) Damage depth with variable step updating. C) Damage area
with fixed step updating. B) Damage area with variable step updating...................22

2-8 Measured and predicted TKR damage. A) CMM measurement of one tibial
insert tested on an AMTI simulator machine after 5e6 cycles of simulated gait.
B ) Predicted tibial insert dam age. ................................................. .....................22

2-9 Damage contour maps predicted by the computer simulation. A) No surface
updating with the linear material model. B) No surface updating with the
nonlinear material model. C) Fixed step updating with the linear material
model(Interval 2.5e5, 20 simulations). D) Fixed step updating with the nonlinear
material model (Interval 2.5e5, 20 simulations). E) Variable step updating with









the linear material model (Threshold 0.5%, 13 simulations). F) Variable step
updating with the nonlinear material model (Threshold 0.5%, 11 simulations). .....24

3-1 The experimental and computational methods used to develop and evaluate the
dynamic contact model of the instrumented knee implant................................ 33

3-2 Medial and lateral contact forces calculated. A) During gait activity. B) During
sta ir a ctiv ity .............................................................................................................3 5

3-3 Maximum and average contact pressures calculated. A) Linear material model
during gait. B) Nonlinear material model during gait. C) Linear material model
during gait. D) Nonlinear material model during stair.........................................36

3-4 Experimental and simulated center of pressure locations. A) In the anterior-
posterior direction for gait. B) In the medial-lateral direction for gait. C) In the
anterior-posterior direction for stair. D) In the medial-lateral direction for stair.
Results are plotted only for the linear material model since results for the
nonlinear material model were indistinguishable. The small plot in the lower
right plot shows the dimension and origin of the tibial tray.................................37

3-5 Error in predicted anterior-posterior center of pressure location as a function of
the inverse of the applied axial load during gait. .........................................42

3-6 Total tibial load and medial to total load ratio. A) During treadmill gait. B)
D during overground gait. ........................................ .............................................44

4-1 The experimental and computational methods used to calculate the knee
adduction torque and medial contact force during gait. ........................................52

4-2 Measured total and predicted medial contact force and internal abduction toque
at the knee joint. A) Medial contact force during one normal gait. B) Medial
contact force during one toe out gait trial. C) Internal abduction torque during
one normal gait. D) Internal abduction torque during one toe out gait trial. E)
Medial force ratio during one normal gait. F) Medial force ratio during one toe
ou t g ait trial. ....................................................... ................ 5 5

4-3 The correlation between internal knee abduction torque and medial contact force
or medial to total force ratio. A) For one normal gait trial. B) For one toe out
gait trail. These two trials are the same as those shown in 4-2. ............................56















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

ANALYSIS OF IN VITRO AND IN VIVO FUNCTION OF
TOTAL KNEE REPLACEMENTS USING DYNAMIC CONTACT MODELS

By

Dong Zhao

August 2006

Chair: Benjamin. J. Fregly
Major Department: Mechanical and Aerospace Engineering

Despite the high incidence of osteoarthritis in human knee joint, its causes remain

unknown. Total knee replacement (TKR) has been shown clinically to be effective in

restoring the knee function. However, wear of ultra-high molecular weight polyethylene

has limited the longevity of TKRs. To address these important issues, it is necessary to

investigate the in vitro and in vivo function of total knee replacements using dynamic

contact models. A multibody dynamic model of an AMTI knee simulator was developed.

Incorporating a wear prediction model into the contact model based on elastic foundation

theory enables the contact surface to take into account creep and wear during the dynamic

simulation. Comparisons of the predicted damage depth, area, and volume lost with worn

retrievals from a physical machine were made to validate the model.

In vivo tibial force distributions during dynamic and high flexion activities were

investigated using the dynamic contact model. In vivo medial and lateral contact forces

experienced by a well-aligned instrumented knee implant, as well as upper and lower









bounds on contact pressures for a variety of activities were studied. For all activities, the

predicted medial and lateral contact forces were insensitive to the selected material

model. For this patient, the load split during the mid-stance phase of gait and during stair

is more equal than anticipated.

The external knee adduction torque has been proposed as a surrogate measure for

medial compartment load during gait. However, a direct link between these two

quantities has not been demonstrated using in vivo measurement of medial compartment

load. In vivo data collected from a subject with an instrumented knee implant were

analyzed to evaluate this link. The subject performed five different overground gait

motions (normal, fast, slow, wide, and toe out) while instrumented implant, video motion,

and ground reaction data were simultaneously collected. The high correlation coefficient

results support the hypothesis that the knee adduction torque is highly correlated with

medial compartment contact force and medial to total force ratio during gait.














CHAPTER 1
INTRODUCTION

Background

The human knee joint is one of the most important joints in human body and

frequently affected by osteoarthritis (OA, a degenerative joint disease causing pain and

loss of function). Knee osteoarthritis symptoms of pain and dysfunction are the primary

reasons for total knee replacement (TKR). Despite the high incidence of OA in knee

joint, its causes remain unknown. Mechanical loading at the knee joint is believed to be

one cause of OA. Unfortunately, noninvasive in vivo direct measurement of knee joint

load distribution is not available yet. Researchers have investigated the relationship

between dynamic gait measurements and clinical outcome in order to find a marker that

can predict knee OA. The peak external knee adduction moment is identified as an

important clinical marker (Prodromos et al., 1985; Andriacchi 1994; Wang et al., 1990;

Hurwitz et al., 1998; Sharma et al., 1998). However, a direct link between these two

quantities has not been demonstrated using in vivo measurement of medial compartment

load.

TKR has been shown clinically to be effective in restoring the knee function.

However, wear of ultra-high molecular weight polyethylene (UHMWPE) has limited the

longevity of TKRs (Sharkey et al., 2002). Consequently, knee simulator machines are

commonly used to evaluate wear performance of new knee implant designs and materials

(Walker et al., 1997). Wear testing on a simulator machine is time consuming and

expensive due to the large number of low-frequency cycles that must be run (Muratoglu









et al., 2003). Moreover, different stations on the same test machine sometimes produce

different wear results. A computationally effective computer simulation could be an

alternative way to evaluate TKRs.

Specific Aims

A contact model capable of calculating contact pressure distribution was applied to

the virtual displacement controlled AMTI knee simulator machine and in vivo artificial

simulation. A multibody dynamic model of an AMTI knee simulator was developed.

Incorporating a wear prediction model into the contact model based on elastic foundation

theory enabled the contact surface to take into account creep and wear during the

dynamic simulation. Comparisons of the predicted damage depth, area, and volume lost

with worn retrievals from a physical machine were made to validate the model. In vivo

tibial force distributions during different activates was studied. The reliability of the

calculated medial and lateral contact forces was evaluated based on the contact model's

ability to reproduce the experimentally measured medial-lateral (ML) and anterior-

posterior (AP) CoP locations. The relationship between the knee adduction torque and in

vivo medial contact force during normal, fast, slow, wide, and toe out gait was

investigated. The hypothesis tested was that the knee adduction torque is highly

correlated with medial compartment force and force ratio (i.e., ratio of medial to total

contact force) during a variety of gait activities.

Dissertation Organization

In Chapter 2, conceptual and computational details are presented for a methodology

to incorporate a wear prediction model into the contact model for specific application to

TKR wear analysis. A multibody dynamic contact model with surface updating of an

AMTI knee simulator machine was constructed to predict the wear volume, damage








depth and areas of a tibial insert. An updating rule for automatic contact surface updating

during wear simulation was determined using a cylinder-on-plate model. This rule then

was applied to the TKR simulation. The wear areas and depth with fixed step updating

and auto-updating were compared with retrieved worn tibial inserts to validate the model.

Furthermore, this chapter evaluated whether a wear factor determined from a pin-on-plate

test could be incorporated into the differential elements of a joint-level computational

simulation to predict wear in a TKR.

Wear Performance

t
In Vitro Function


Dynamic Contact Model


In Vivo Function


Internal Loads External Predictors

Figure 1-1. Thesis concepts.

In Chapter 3, in vivo medial and lateral tibial loads during gait, stair rise/descent,

kneel, and lunge activities for a single knee implant patient were determined. Contact

forces and pressures during each activity were found using a novel combination of

experimental and computational techniques. The results provided insight into the in vivo

loading conditions experienced by a well-aligned knee implant for a variety of activities.

In Chapter 4, the relationship between the knee adduction torque and in vivo medial

contact force during normal, fast, slow, wide, and toe out gait was studied. A single

patient with an instrumented knee implant provided a unique opportunity to perform the









investigation. The adduction torque curve for each gait pattern was obtained using

standard external gait measurements. The corresponding internal axial contact force was

measured by the instrumented knee implant. A linear regression equation was used to

determine medial contact force from the implant load cell measurements, where the

regression coefficients were found by using a dynamic contact model to compute medial

and lateral contact force from fluoroscopic and axial load data collected from the same

patient performing treadmill gait. The hypothesis tested was that the knee adduction

torque is highly correlated with medial compartment force and force ratio (i.e., ratio of

medial to total contact force) during a variety of gait activities.

Chapter 5 summarizes the research, with suggestion for future improvements to the

contact and damage model.














CHAPTER 2
COMPUTATIONAL PREDICTION OF KNEE ARTHROPLASTY DAMAGE USING
A PIN-ON-PLATE WEAR FACTOR AND EVOLVING SURFACE GEOMETRY

Wear of ultra-high molecular weight polyethylene (UHMWPE) remains a primary

factor limiting the longevity of total knee replacements (TKRs). However, wear testing

on a simulator machine is time consuming and expensive, making it impractical for

iterative design purposes. The objectives of this paper were first, to evaluate whether a

constant wear factor from pin-on-plate tests can be used in a computational model to

predict accurate joint-level damage in a TKR, and second, to evaluate how choice of

surface updating methodology (fixed or variable step) and material model (linear or

nonlinear) affect the predictions. A computational damage model was constructed of a

commercial knee implant in an AMTI simulator machine. The damage model combined a

dynamic contact model with a surface update model to predict how wear plus creep alter

the tibial insert geometry over multiple simulations. The computational framework was

validated by predicting progressive wear in a cylinder-on-plate system for which an

analytical solution was derived. The implant damage model was then evaluated for 5

million cycles of simulated gait by comparing damage predictions with damage

measurements made from an AMTI machine. Using a wear factor measured for the same

material pair, the model predicted tibial insert wear volume to within 2% error and

damage depths and areas to within 18% and 10% error, respectively. Choice of material

model had little influence, while inclusion of surface updating affected damage depth and

area but not wear volume predictions. Specific updating method was important only









during the first half million cycles, where variable step was needed to capture rapid

geometry changes due to creep. Overall, our results indicate that accurate TKR damage

predictions can be made with a computational model using a constant wear factor

obtained from pin-on-plate tests performed with the same material pair, and furthermore,

that surface updating approach matters only during the initial "break in" period of the

simulation.

Introduction

Wear of ultra-high molecular weight polyethylene remains a primary factor limiting

the longevity of total knee replacements (Sharkey et al. 2004). Consequently, knee

simulator machines are commonly used to evaluate wear performance of new knee

implant designs and materials (Barnett et al., 2002; Burgess et al., 1997; DesJardins et al.,

2000; Muratoglu et al., 2003, Walker et al., 1997). Wear testing on a simulator machine

is time consuming and expensive due to the large number of low-frequency cycles that

must be run. Moreover, different stations on the same machine sometimes produce

different wear results. Consequently, it is impractical to use a knee simulator machine to

test the sensitivity of a TKR design to the material used, implant geometry, implant

alignment, and loading conditions.

Because of these issues, researchers have sought alternative methods to speed up

the implant design and evaluation process. Pin-on-plate wear tests (Fisher et al., 2004)

using the same material pairs as in a TKR are faster and cheaper to perform than are tests

on a knee simulator machine. With a simplified motion path and well controlled sliding

speeds, accurate wear factors can be obtained under test conditions similar to those of

knee simulator machines. However, pin-on-plate wear tests do not provide the joint-level









damage evaluation desired for design purposes since they do not account for TKR surface

geometry.








A B C
Figure 2-1. Experimental measurements and the computational model to predict total
knee anthroplasty damage. A) Experimental wear factor measurement. B)
Computational damage prediction. C) Experimental damage measurement.

In contrast, computer simulation can be an efficient and reproducible method for

predicting TKR wear performance, with simulator machines providing a well-controlled

test bed for evaluation purposes. Motion and load inputs are well defined, the number of

loading cycles is known precisely, and wear volume can be measured gravimetrically at

known intervals for comparison. Though computational damage predictions of TKRs

have been developed for the AMTI (Zhao et al., 2006) and Stanmore (Knight et al., 2005,

Rawlinson et al., 2006) simulator machines, these predictions possess several important

limitations. Wear factors were taken from the literature or fine tuned to match

experimental wear volume measurements. Predicted damage depths and areas were not

assessed quantitatively. Different types of material models were used in different studies.

Apart from Knight et al. (2005), results from a single simulation were extrapolated out to

the full number of loading cycles, which does not account for the gradual evolution of the

worn surface geometry that occurs in real life. It remains unknown whether a wear factor

from pin-on-plate tests performed on the same material pair as the implant can be used in

a computational damage simulation to predict accurate joint-level damage (depth, area,

volume). Furthermore, how features of the damage model (linear wear plus nonlinear









creep) interact with choice of surface updating method (fixed or variable time intervals)

and material model (linear or nonlinear) to affect the damage predictions is also an open

issue.

The objectives of this paper were first, to evaluate whether a wear factor from pin-

on-plate tests can be used in a computational model to predict accurate joint-level

damage in a TKR, and second, to evaluate how choice of surface updating methodology

and material model affect the predictions. Two methods for performing periodic updating

of the worn surface geometry are investigated updating based on fixed time intervals

(i.e., fixed step), and updating based on a fixed amount of surface change (i.e., variable

step). The computational framework is validated analytically using a cylinder-on-plate

wear problem with known analytical solution and evaluated experimentally by

performing a TKR damage prediction for a commercial knee implant tested in an AMTI

simulator machine. The wear factor for the TKR damage prediction is determined

experimentally from pin-on-plate tests performed on the same material pair as the

implant. Predicted damage depths, damage areas, and wear volumes after 5 million cycles

of simulated gait are compared to measurements obtained from physical simulator

machine.

Methods

Computational Framework for Damage Prediction

A computational methodology was developed to simulate progressive surface

damage (= wear + creep) over multiple loading cycles. The methodology combines a

multibody dynamic contact model with a surface update model, where the two models

iterate to predict progressive surface damage (Figure 2-1). The dynamic model is

constructed within the Pro/MECHANICA MOTION simulation environment (PTC,









Waltham, MA) and incorporates a custom elastic foundation contact model developed

specifically for multibody simulations (Bei and Fregly, 2004). The dynamic contact

model is encapsulated as a dynamic link library using the Equations of Motion

functionality within Pro/MECHANICA MOTION, allowing the model to be incorporated

into user-written C++ code. A computational damage model written in C++ is then used

to combine the dynamic contact model with a custom surface update model, where

analyses performed with both models can be iterated automatically.


Figure 2-2. Computational framework of the damage prediction with surface geometry
updating.

A five-step process is followed to perform computational damage predictions with

surface updating (Figure 2-2). The first step is a static analysis to find the initial pose of









the contacting bodies prior to performing a forward dynamic simulation. The second step

is a forward dynamic simulation (i.e., forward analysis) of the entire system over one

cycle to predict the relative motion of the contacting bodies at each time instant. The third

step is an inverse dynamic analysis to predict contact pressures and sliding conditions on

the articulating surfaces at desired time points given the relative motion of the contacting

bodies. The fourth step is a damage analysis to calculate the change in damage depth

3Damage at each point across the contact surfaces for the specified number of cycles. If the

final number of cycles has not been reached, the next step is an update analysis to modify

the surface geometry to reflect the total damage depth sustained at each point thus far.

Finally, the entire process is iterated, starting with the static analysis, until the specified

number of motion cycles has been simulated.

Surface updating is modeled by using a modified version of an elastic foundation

contact model. The traditional elastic foundation model scatters a "bed of springs" over

the three-dimensional surfaces to push them apart. The springs represent an elastic layer

of known thickness covering one or both bodies, where each spring is independent from

its neighbors. For a rigid femur contacting a deformable tibial insert of finite thickness, a

uniform grid of spring "elements" is placed on the tibial insert contact surfaces (Bei and

Fregly, 2004), and the contact pressure p for each spring is calculated from (Johnson,

1985; An et al., 1990; Blankevoort et al., 1991)

(1- v)E(p) d
p= (2-1)
(1+ v)(1- 2v) h

where E(p) is Young's modulus of the elastic layer (a constant for a linear

material model and a function of pressure p for a nonlinear material model; Nufio and









Ahmed, 2001), v is Poisson's ratio of the layer, h is the layer thickness at the spring

location, and d is the spring deflection, defined as the interpenetration of the undeformed

(and unworn) surfaces in the direction of the local surface normal. Both h and d are

calculated on an element-by-element basis across the contact surfaces of the tibial insert.

The modification to this traditional formulation is to offset the interpenetration d

and thickness h of each element by the amount of surface damage 3Damage sustained by

the element up to the current number of cycles used in the damage prediction process.

The modified elastic foundation formula is thus


(1-v)E(p) (d- Dmge) )
p = (2-2)
(1+ v)(1- 2v) (h- Damage)

At the end of each damage analysis, 8Daage is calculated on an element-by-element

basis and written to a file. At the start of the next iteration, the update analysis reads in

and stores the value of 3Dage for each element so that it can be used in the subsequent

element pressure calculations. Thus, the actual surface geometry is never altered, and the

interpenetration d between the undeformed contact surfaces continues to be calculated as

if no surface damage had occurred.

The calculation of 3Daage in the damage analysis accounts for the combined effects

of material lost due to mild wear 8wer and surface deformation due to compressive creep


Creep


8Damage = Wear Creep (2-3)
The total depth of material removed from an element 8Wear is predicted using an iterative

version of Archard's classic law for mild wear (Archard and Hirst, 1956):






12



9wear = NVJ k'I PIAt (2-4)
J= 1-=1 )
where i represents time frames within a one-cycle inverse analysis, n is the total number

of time frames in the analysis,j represents an individual inverse anlaysis, m is the total

number of inverse analyses performed thus far, and k is a constant wear factor. At any

time frame i within a one-cycle inverse analysis, p, is the element contact pressure, IvI

the magnitude of the element's relative sliding velocity, and At the time increment used

in the analysis, so that Iv]|At represents the sliding distance experienced by the element.


For any inverse analysis, (k :nl p, vAt)j is the one-cycle mild wear depth from

Archard's wear law, and N, is the incremental number of cycles for which the one-cycle

wear depth is to be extrapolated. In practice, only the incremental increase in wear depth

Awear is calculated for the current iteration, and this value is added to the current value

of Damage during the damage analysis.

The total depth of surface deformation on each element due to compressive creep

,reep is calculated based on curve-fitted UHMWPE creep data reported by Lee and

Pienkowski (1997 and 1998 a, b):




cep = C1 +C2 Log At .-NJ -4 4 J j (2-5)
I L YN\
J1
where C, = 3.491 x 10-3 and C, = 7.966 x 10-4 are constants and all other quantities are as

defined in Eq. (2-4). The unit for pressure is MPa, the unit for time minutes, and the unit

n
for thickness mm. The quantity Ipn is the average pressure on the element over the
S1=









course of dynamic simulation, while the quantity m -j(l lp, /n)j is the average

pressure over all cycles N simulated thus far. Unlike the creep model reported in Fregly et

al., (2005), creep recovery is built into this model since the within-cycle average pressure

uses both loaded and unloaded time frames and the between-cycle average pressure

accounts for all cycles simulated thus far. Unlike for wear, the incremental change

(increase or decrease) A5re, in creep cannot be calculated directly. Instead, the total

amount of creep deformation is calculated from Eq. (2-5) and the incremental change

determined by subtracting the previous value stored in memory. This incremental change

is also added Damage during the damage analysis.

The calculation of Cre is complicated by the fact that it relies on past history. To

address this problem, we store the most recent values of m 1Nj (Y" ,p, /n)J and


- N, in memory for use during the subsequent damage analysis. Combining these

quantities with their incremental changes from the current iteration provides all of the

information needed to solve Eq. (2-5).

The incremental number of cycles Nm at which the contact surface geometry is

updated can be either fixed or variable. Fixed step updating requires a user-defined

update interval so that Nm is known in advance. Variable step updating alters the surface

geometry when the calculated surface change (= Aea,, + Ad8,ep ) ofany element exceeds

a specified threshold value. In this study, we define this threshold as a specified

percentage of the current element thickness h- 3amge With this method, the value of









Nm required to trigger an update is not known in advance and is found iteratively by

solving a nonlinear rootfinding problem.

Analytical Validation of Computational Framework

The computational framework was validated by predicting progressive wear in a

simple cylinder-on-plate system for which an analytical wear solution was derived. In

this system, a rigid cylinder of radius R, is pressed onto a fixed plate of thickness h,

width w, and length / by a constant vertical load Fn. The cylinder was considered to be

rigid and the plate linearly elastic with Young's modulus E and Poison's ratio v. The

cylinder rotated about its long axis at a constant angular speed S/R, (i.e., a constant

linear sliding speed of S at each point on the plate surface). Mild wear between the

cylinder and plate was described by a constant wear factor k. Contact between the

cylinder and plate was described by an elastic foundation model, the change in plate

thickness due to wear was assumed to be negligible, and the shape of the worn plate

surface was assumed to be cylindrical. With these assumptions, we derived analytical

solutions (see Appendix for details) for wear volume Vwea, and maximum wear depth


3wear and wear area Aa,, for any number of cycles N:

Vwa = kNF S
8wea = Rw cosO'-R cosO (2-6)
2R, sin 0'
sin0
In these equations, R, is the radius of the worn surface and 0 and 0' are the

angles between R, and R,, respectively, and a vertical axis (Figure 2-3). The unknowns

on the right hand side of Eqs. (2-6) are R,, 0, and 0'. The derivation produces three










coupled nonlinear equations in these three unknowns, which are solved using nonlinear

rootfinding methods. Creep was not included in the analytical model since no closed-

form solution can be derived in this case.


Rw -


7Vw7ar, *w*'
Vctac
Worn surface

VGeometry VContact + Vwear




Figure 2-3. Cylinder-on-plate contact model with surface updating.

A computational damage model of the cylinder-plate system was constructed to

ensure that the entire framework was working properly as well as to develop empirical

rules for fixed and variable step updating. The model was constructed such that the

dimensions and material properties of the cylinder and plate were similar to those of a

total knee replacement. The cylinder radius was 40 mm and the plate dimensions were 40

mm long by 20 mm wide by 10 mm thick. The cylinder length was 30 mm and the

cylinder completely covered the width of the plate. The plate was assumed to be linear

elastic with a Young's modulus of 463 MPa (Kurtz et al., 2002) and Poisson's ratio of

0.46 (Bartel et al., 1995) to emulate polyethylene. The normal force was 1000 N to

approximate the maximum load experienced on one condyle. The angular speed of the

cylinder was a constant 2;r rad/sec (60 RPM) corresponding to a 1 Hz frequency. Wear

results were generated over 5 million cycles using a constant wear factor of 1 x 10 7

mm3 / Nm The contact element grid density on the plate surface was set to 400 x 1 so

that the length of each element was 0.1 mm and the width spanned that of the plate.









The cylinder-on-plate computational model was used to generate wear predictions

using fixed and variable step surface updating with different update intervals. The fixed

step predictions used update intervals of 5, 2.5, 1, 0.5, 0.25, and 0.1 million cycles (mc),

while the variable step predictions used update thresholds of 5, 2.5, 1, 0.5, 0.25 and 0.1%

of the plate thickness. For both methods, the update value that produced wear depth, area,

and volume errors on the order of 5% at the final time point was selected as the value to

use in the subsequent TKR damage predictions.

Experimental Evaluation of Computational Framework

The ultimate goal of the computational framework is to predict surface damage in

commercial TKR designs. Thus, to evaluate the framework for this situation, we

compared computational damage predictions with experimental damage measurements

for a commercial TKR design. Wear testing on an AMTI knee simulator machine was

performed on three implants of the same design (Hi-tech Knee II cruciate-retaining,

Nakashima Medical Division, Nakashima Propeller Co., Ltd., Japan). The machine

performed 5 million cycles of simulated gait at a frequency of 1.0 Hz. The lubricant used

in the tests was distilled water (74.7%) + bovine calf serum (25%) + sodium azide

(0.3%). The temperature was maintained at 37.0 + 0.0 C. The tibial inserts were made

from GUR1050 powder by direct compression molding. Before testing, the contact

surfaces of one unworn femoral component and tibial insert were digitized using a

coordinate measuring machine (CMM) with an accuracy of 0.01 mm. Wear volumes of

the three tibial inserts were measured gravimetrically every 1 million cycles, while

damage depths and areas for one insert were measured at 5 million cycles using the

CMM.









The wear factor for the computational model was obtained by pin-on-plate tests

performed with the same material pair as used in the implant. In this way, the wear factor

used in the computational damage predictions was consistent with the wear tests

performed in the simulator machine. The polyethylene pin followed a rectangular wear

path of 25 mm by 10 mm. The sliding velocity was 35 mm/s and the total sliding distance

was 28 km. The lubricant used in the test was the same as that used in the AMTI

machine, and the temperature was maintained at 36.5 + 1.0 C. The wear factor was

calculated based on gravimetrically measured volume loss using a polyethylene density

of 0.943 gm/cm3 (Nakashima Propeller Co., 2005). The resulting wear factor from three

pin-on-plate tests was 2.59 + 0.63 x 10 mm3 /Nm, with the mean value being used in

the TKR computational damage prediction.

A multibody dynamic contact model of one station of an AMTI knee simulator

machine was constructed within Pro/MECHANICA MOTION to predict surface damage

in the same TKR design. The degrees of freedom in the dynamic model were constructed

to match those of the simulator machine, and the implant components were positioned in

the model to match their positioning in the physical machine. An additional 6 degree-of-

freedom (DOF) joint between the femoral component and tibial insert was used to

measure relative (i.e., joint) kinematics for contact calculations. Each DOF in the model

was either motion or load controlled to mimic the function of the AMTI machine. Motion

and load inputs to the model were taken as the feedback (i.e., achieved) waveforms

measured by the machine during the actual wear tests. Contact surfaces for the femoral

component and tibial insert were reverse-engineered from the pre-test CMM data using

Geomagic Studio (Raindrop Geomagic, Research Triangle Park, NC).









Linear and nonlinear material models with a Poisson's ratio of 0.46 were used for

all damage predictions. The linear model used a Young's modulus of 463 MPa. The

nonlinear model used a published relationship for the tangent modulus E as a function of

contact pressure (Bei and Fregly, 2004):


E =/{1 I 1i+nK ] (2-7)
2 po Po

where sE = 0.0257, p, = 15.9, and n = 3 are material parameters (Fregly et al., 2003).

Based on computational results for the cylinder-on-plate system, damage predictions with

both material models were performed over 5 million cycles of simulated gait using fixed

step updating with an interval of 0.25 mc and variable step updating with a threshold of

0.05%, as well as with no surface updating.

Results

The wear predictions for the cylinder-on-plate system reproduced the analytical

wear results as long as an appropriate surface updating criterion was used. Wear volume

predictions were insensitive to the exclusion or inclusion surface updating or to the

choice of updating method, in all cases matching the analytical result to within 0.001%

root-mean-square (RMS) error after 5 million cycles. In contrast, wear depth and area

predictions were highly sensitive to update criteria but less sensitive to update method

(Figure 2-4). As the update criterion was reduced from 5 mc or 5%, wear depth and area

predictions converged to the analytical solution. When the update criterion was too loose,

wear depth predictions were too high and wear area predictions too low with the

corresponding wear contours exhibiting unrealistic sharp edges (Figure 2-5). An update






19


criteria of approximately 0.25 mc (fixed step) and 0.05% (variable step) were required to

match the analytical damage depth and area to within about 5% RMS error (Table 2-1).

Table 2-1. Percent RMS error over all cycles used in the damage prediction process. For
update interval, mc means millions of cycles, % refers to percent of plate
thickness.
Updating Update Number of Wear Wear Wear
Method Interval Simulations Depth (%) Area (%) Volume (%)
Fixed 5 mc 1 202.2 67.1 0.0
2.5 mc 2 152.9 51.0 0.0
1 mc 5 45.9 24.0 0.0
0.5 mc 10 8.9 10.3 0.0
0.25 mc 20 2.7 4.9 0.0
0.1 mc 50 2.4 2.1 0.0
Variable 5% 6 42.6 22.1 0.0
2.5% 8 19.5 13.5 0.0
1% 13 3.5 7.4 0.0
0.5% 18 2.6 5.1 0.0
0.25% 28 2.6 3.1 0.0
0.1% 66 2.6 1.5 0.0


1.5


-*-5 mc
- 2.5 mc
1 me
-- 0.5 mc
-- 0.25 mc
- 0.1 mc
- Analytical


" -I


30C

2 20C

100lo

c


Figure 2-4


1.5


A 0


0


1 2 3


4 5


2 3


Million cycles C Million cycles D
SWear depth and area predictions of the cylinder-on-plate model. A) Wear
depth with fixed step updating. B) Wear depth with variable step updating. C)
Wear area with fixed step updating. D) Wear area with variable step updating.


For the TKR system, the computational damage predictions closely matched the

experimental damage measurements. Regardless of surface updating method (none, fixed,


-- 5%
-2.5%
1%
-- 0.5%
- 0.25%
-0.1%
- Analytical


2


e.7











or variable) or choice of material model (linear or nonlinear), the predicted wear volume


at 5 million cycles was 40.5 mm3, an error of only 1.5% compared to 39.9 + 3.4 mm3

from the three gravimetric wear measurements performed at the end point of each AMTI

test (Figure 2-6).


E -Plate
E Cylinder
0)

-5 A


5






0)
-5B





-5

0)





0)
"-5 D





-5
-20 Length (mm) 20 E
Figure 2-5. Cross section of the worn plate of the cylinder-on-plate tests with different
updating methodology. A) No update (Interval 5e6, 1 simulation). B) Fixed
step (Interval 2.5e6, 2 simulations). C) Fixed step (Interval 2.5e5, 20
simulations). D) Variable step (Threshold 5%, 7 simulations). E) Variable step
(Threshold 0.5%, 19 simulations, 8 updates).

In contrast, the predicted damage depths and areas were less accurate, with medial-

lateral errors of 16% and 18% in damage depth and 10% and less than 1% in damage area

for the best run (Table 2-2; nonlinear material model with variable step updating). Depth

and area predictions were insensitive to choice of material model but sensitive to selected

update method (Figure 2-7). No surface updating resulted in overprediction of damage









depth and underprediction of damage area. Only variable step updating captured a rapid

change in surface geometry during the first half-million cycles, with depth predictions for

fixed step updating overshooting the variable step results. Nonetheless, after only 1

million cycles, depth and area predictions from both updating methods converged to

approximately the same trajectory for the remaining 4 million cycles. Fixed step

predictions with an update interval of 5 mc required 20 simulations, while variable step

predictions with a threshold of 0.5% required 13 simulations for the linear material model

and 11 for the nonlinear model.

Table 2-2. Comparison of experimental and predicted damage depths and areas (variable
step updating with nonlinear material model) for the one insert whose surface
geometry was measured using a coordinate measuring machine. Results for
other updating and material model combinations were similar.
Experiment Simulation
Damage Lateral Medial Total Lateral Medial Total
Depth (mm) 0.43 0.44 ---- 0.35 0.37 ----
Area (mm2) 212 229 441 233 228 461


60

S-*- Insert 1
E Insert 2
E 40 Insert 3
E -*- Simulation

20
20 T




0 1 2 3 4 5

Million cycles

Figure 2-6. Simulated and gravimetrically measured volume loss of the tibial inserts due
to mild wear. Simulation error bars indicate + 1 SD in wear factor
measurement.











+ Linear 5 me
E Nonlinear 5 mc E
S0.75 Linear 2.5 mc 0.75
Nonlinear 2.5 me
a) Linear 0.5%
S0.5 Nonlinear 0.5% 0 0.5
E E
S0.25 J 0.25

0 A B

300 300

2 2
200 m 200





0 1 2 3 4 5 0 1 2 3 4 5
E 100 E 100


012345 012345
Million cycles C Million cycles D


Figure 2-7. Damage depth and area predictions of the TKR. A) Damage depth with fixed
step updating. B) Damage depth with variable step updating. C) Damage area
with fixed step updating. B) Damage area with variable step updating.









0.01
Lateral Medial A Lateral Medial B


Figure 2-8. Measured and predicted TKR damage. A) CMM measurement of one tibial
insert tested on an AMTI simulator machine after 5e6 cycles of simulated gait.
B) Predicted tibial insert damage.

The predicted damage regions for the TKR system were in good qualitative

agreement with the worn tibial insert obtained from the AMTI machine (Figure 2-8). The

predicted and measured damage scars were similar in shape and location on the insert,

with the predicted locations of maximum damage on the medial and lateral sides being

similar to those on the worn insert (x's in Figure 2-8). At the anterior-lateral corer of the

medial damage scar, a small outcropping region observed experimentally was reproduced









by the computational model. Similarly, at the anterior-medial corner of the lateral damage

scar, a small incropping region observed experimentally was also reproduced by the

model. Switching from the linear to the nonlinear material model or from fixed to

variable step updating had little influence on the predicted damage scars, while switching

from no updating to either updating method resulted in a noticeable increase in the size of

the damage scars.

Discussion

This study used a computational model to predict damage in a commercial knee

implant design after 5 million cycles of simulated gait on an AMTI knee simulator

machine. A wear factor from pin-on-plate tests performed using the same material pair as

in the implant was used to generate the computational predictions. Compared to damage

measurements made on the same implant following physical testing on an AMTI

machine, wear volumes were predicted to within 2% error, damage depths to within 18%

error, and damage areas to within 10% error. Choice of material model (linear versus

nonlinear) had little affect on the predictions, while choice of updating method (fixed

versus variable) only had an influence at the start of simulation (Figure 2-7 and Figure 2-

9). In addition to this experimental evaluation, the computational framework was

validated using an analytical wear solution derived for a cylinder-on-plate system.

Computational damage models may prove valuable in the future for screening new knee

implant designs rapidly or performing sensitivity and optimization studies that would be

too time consuming to complete with physical simulator machines.

The analytical wear solution for the cylinder-on-plate system permitted an objective

evaluation of the proposed computational framework with surface updating. The virtually

identical wear volume results for the analytical and numerical models indicated that the








computations in the numerical model were being performed correctly. The small error

between analytical and numerical wear depth for tight update criterion is expected since

the worn surface is not a perfect cylinder as assumed in the analytical derivation. The

jagged worn surfaces predicted by loose update criteria are caused by high contact

pressures on the edges of the worn surface from the previous update (Figure 2-5). This

analytical solution may prove useful as a benchmark test case for other studies involving

computational wear prediction.
0.90




A 0.01B





C0.01 D





E 0.01 F
Figure 2-9. Damage contour maps predicted by the computer simulation. A) No surface
updating with the linear material model. B) No surface updating with the
nonlinear material model. C) Fixed step updating with the linear material
model(Interval 2.5e5, 20 simulations). D) Fixed step updating with the
nonlinear material model (Interval 2.5e5, 20 simulations). E) Variable step
updating with the linear material model (Threshold 0.5%, 13 simulations). F)
Variable step updating with the nonlinear material model (Threshold 0.5%, 11
simulations).

While choice of material model did not have a significant affect on our results, the

nonlinear material model was still preferable. It exhibited slightly faster convergence









characteristics (i.e., smaller oscillations of damage depth, less underestimation of damage

area) compared to the linear material model for a given update method and criterion. For

example, when we performed additional damage predictions with fixed-step updating

using an interval of 1 me, damage depth for the nonlinear material model stopped

oscillating by 2 me, while for the linear material it still exhibited small oscillations up to

4 mc. Furthermore, for a given variable step update threshold, the nonlinear model

required fewer simulations than did the linear material model. Thus, given a choice

between the two material models, the nonlinear model appears to be preferable from a

computational speed and stability perspective.

Though both the fixed step and variable step updating matched the experimental

damage results well, variable step updating is advantageous for two reasons. First, it

gives a smoother worn surface than does fixed step updating for the same number of

simulations. Second, it can capture rapid changes in surface geometry that cannot be

captured by fixed step updating. In our damage prediction, the initial rapid change in

damage depths and areas was cause by our creep model. These rapid changes are likely

reflective of the situation in vivo. Nonetheless, if accurate damage depth and area

predictions are desired only at the 5 million cycle end point, either fixed or variable step

updating will work equally well. In fact, fixed step updating may possess a slight

advantage, since the number of simulations that will be performed is known in advance.

Four modeling assumptions were involved in our TKR damage prediction process.

The first was that the elastic foundation contact model is a reasonable approximation of

the full three-dimensional elasticity problem. Though this model tends to slightly

overestimate contact area and underestimate contact depth, these inaccuracies become









even smaller as the worn surface geometry becomes more conformal. As suggested by

the quality of our predictions, the elastic foundation model appears to be adequate for

predicting knee replacement contact mechanics as long as subsurface stress information

is not required. The second assumption was that the surface updating method only

modifies the surfaces in the direction of their original unworn surface normals. As the

damage process progresses, the normal direction of each contact element is not adjusted.

However, since the damage depths remained "small" after 5 million cycles, and since the

damage predictions were in good agreement with experimental observations, the

influence of this assumption was likely negligible. The third assumption was that the

wear factor used in Archard's wear law was a constant. The literature reports that wear

factor measurements can be affected by a number of conditions such as surface roughness

(Lancaster et al., 1997), contact pressure (Barbour et al., 1997), lubricant (Saikko and

Ahlroos, 2000), wear path (Endo et al., 1999), and time-varying loading (Barbour et al.,

1997). Despite this fact, the constant wear factor used in our study, which was not fine-

tuned to match the AMTI damage measurements, worked exceptionally well for

predicting joint-level wear volume with the computational model. The fourth assumption

was that creep deformation after 5 million cycles was not recoverable. By accounting for

the time history of loading and unloading, our creep model can reproduce loading and

unloading curves reported in the literature (Lee and Pienkowski, 1997 and 1998 a, b). If

we allowed our creep model to relax for a long enough time, all of the creep deformation

would be recovered, since our model calculates creep based on the average pressure over

all cycles. Thus, the creep predicted by our model may represent a combination of

viscoelastic and plastic effects. If we leave out our creep model, wear volume and









damage area are still predicted accurately as long as surface updating is used, while

damage depth is underestimated by approximately 50%. Additional experimental creep

data would be needed to refine the current creep model further.

The most significant limitation of our study was that only a single implant design

was available for analysis. It is difficult to find pin-on-plate wear factor data for the same

material pair used in an implant tested on a simulator machine. Furthermore, it is difficult

to find experimental data for pin-on-plate and corresponding simulator tests that are

performed well. Whether or not the accuracy of our predictions is generalizable to other

implant designs and types of simulator machines will require further investigation.

Table 2-3. The surface updating methodology, material model and creep model should be
used to achieve desired goals. Goal 1: Predict wear volume at any time point.
Goal 2: Predict wear volume, damage depth, and damage area at any time
point. Goal 3: Predict wear volume, damage depth, and damage area only at
the final time point
Goal 1 Goal 2 Goal 3
Surface Updating None Variable Fixed or Variable
Material Model Linear or Nonlinear Nonlinear Linear or Nonlinear
Creep Model None Included Included*
*If damage depth prediction is not needed, then creep model can be excluded.

Our results may have important implications for the use of computational tools in

the knee implant design process (Table 2-3). If wear volume after a larger number of

cycles (e.g., 5 million) is the only desired goal, then surface updating is not necessary,

nor is a creep model. Wear volume results extrapolated from a single simulation without

creep will be just as accurate. If damage depth and area are to be predicted as well, but

only at the endpoint, then fixed or variable step updating will work well, and the creep

model should be included. For example, using fixed step updating every 1 mc with either

material model and the creep model, we can match damage depth, damage area, and wear

volume predictions at 5 mc calculated using a much tighter fixed step update criterion.






28


Finally, if damage depth and area are desired at any point during the damage prediction

process, variable step updating becomes preferable to fixed step updating and the

nonlinear material model preferable to the linear one.














CHAPTER 3
IN VIVO MEDIAL AND LATERAL TIBIAL LOADS DURING DYNAMIC AND
HIGH FLEXION ACTIVITIES

Though asymmetric loading between the medial and lateral compartments of total

knee replacements may contribute to implant loosening and failure, the in vivo contact

load distribution during dynamic daily activities remains unknown. This study reports in

vivo medial and lateral contact forces experienced by a well-aligned knee implant as well

as upper and lower bounds on contact pressures for a variety of activities. In vivo implant

motion and load data were collected from a single knee replacement patient performing

gait, stair rise/descent, lunge, and kneel activities. In vivo motion was measured using

video fluoroscopy, while in vivo axial loads were collected simultaneously using an

instrumented tibial component. A dynamic contact model employing linear and nonlinear

polyethylene material properties was constructed to calculate medial and lateral contact

forces and pressures based on the measured kinematics, axial loads, and centers of

pressure. For all activities, the predicted medial and lateral contact forces were insensitive

to the selected material model. The percentage of medial to total contact force ranged

from 18 to 60 for gait, 47 to 65 for stair, and 55 to 60 for kneel and lunge. In contrast, the

predicted contact pressures were sensitive to the selected material model, with the

nonlinear material giving lower, more uniform pressure results. Contact pressures were

higher for stair than for gait, with kneel producing the lowest pressures and lunge the

highest. For this patient, the load split during the mid-stance phase of gait and during stair

is more equal than anticipated.









Introduction

Asymmetric loading between the medial and lateral compartments of the knee has

been hypothesized to contribute to the development of knee osteoarthritis (OA) (Jackson

et al., 2004). In artificial knees, asymmetric forces exerted on the tibial insert may

contribute to mechanical failure as well as loosening of the implant (Andriacchi et al.,

1986). Ever since Morrison's (Morrison 1970) seminal work, modeling studies have

predicted larger contact forces on the medial than on the lateral side of the knee during

gait (Hsu et al., 1990; Hurwitz et al., 1998; Johnson et al., 1980; Noyes et al., 1992;

Shelburne et al., 2005; Schipplein and Andriacchi, 1991). These studies have reported

between 50 and 100% of the total contact force passing through the medial compartment,

with the predicted medial-lateral load split varying considerably over the gait cycle

(Hurwitz et al. 1998). While recent experimental studies have measured in vivo axial

loads in the femur and tibia using telemetry (D'Lima et al., 2005; Taylor et al., 1997 and

1998), none have been able to partition the total axial load into medial and lateral contact

forces. Thus, the actual in vivo distribution of tibial contact forces during dynamic,

weight-bearing activities remains unknown.

The lack of accurate in vivo medial and lateral contact force data is also

problematic for predicting abrasive/adhesive wear in total knee replacements. Medial-

lateral load split affects medial and lateral contact pressures, which in turn affect mild

wear of the polyethylene tibial insert (Blunn et al., 1991). For lack of better data,

researchers often offset the axial load 5 mm to the medial side when evaluating the wear

performance of a new implant design using a knee simulator machine (DesJardins et al.,

2000). Though in vivo contact pressures have been predicted by computational models









(Fregly et al., 2005; Perie and Hobatho, 1998), they too suffer from assumptions related

to the unknown medial-lateral load split.

This chapter seeks to determine in vivo medial and lateral tibial loads during gait,

stair rise/descent, kneel, and lunge activities for a single knee implant patient. Contact

forces and pressures during each activity are found using a novel combination of

experimental and computational techniques. The results provide insight into the in vivo

loading conditions experienced by a well-aligned knee implant for a variety of activities.

Methods

Data Collection

Data were collected from one total knee arthroplasty patient (male, right knee, age

80, mass 68 kg) eight months after surgery. Institutional review board approval and

patient informed consent were obtained. The patient received a well-aligned custom tibial

prosthesis design instrumented with four uniaxial force transducers, a microtransmitter,

and an antenna. On the frontal radiograph, the line across the undersurface tibial tray is

90.10.5 degrees to the tibial shaft line (line through the mid-point of the tibial cut and

the center of the talus. The femoral component is aligned at 6 degrees valgus to the

anatomic axis of the femur. We recorded in vivo tibial force data simultaneously with

fluoroscopic motion analysis data during treadmill gait, stair rise/descent, kneel, and

lunge activities (Banks and Hodge, 1996). One channel of the analog output of the force

plate was directly connected to the instrumented knee output recording instrument during

the data collection. However, there was not a marker to synchronize the measurements

when regular ground reaction force measurement was not available (e.g. treadmill gait).

The fluoro kinematic and the instrumented knee data were collected with different sample

frequencies. Also, there were delays between the tibial force and analog ground reaction









force on the order of a few milliseconds. These limitations required a clear definition of

the motion cycle and a synchronization analysis. For the gait activity, the motion cycle

was defined to begin at right heel strike, which was determined using synchronized tibial

force and ground reaction force data collected during over ground gait trials. The feature

in the instrumented knee load data curve corresponding to the heel strike event was

identified by determining where the vertical ground reaction force measured during

overground gait become nonzero. Then the initial (0 shift) heel strikes in the instrumented

knee load data of treadmill gait were identified using this feature. For the stair activity,

the stair motion cycle was defined as a rise followed by a descent with the patient's right

leg supporting the body. For the kneel and lunge activities, they were only partial weight

bearing, since the patient was standing on the opposite leg during these activities. These

activities were dynamic to reach the final static pose, only this final pose (which was held

by the patient for several seconds) was analyzed.

For kneel and lunge, joint pose at the most highly flexed position was determined

by averaging five and nine, respectively, time frames of fluoroscopic data since those

were the numbers of fluoroscopic time frames collected for those activities.. The recorded

tibial force data were denser sampled and covered more time than the fluoro data, Since

no synchronization marker was available, the tibial force data corresponding to the

selected time frames were visually identified to make sure all the fluoro data being

covered. The mean, minimum, and maximum force and corresponding center of pressure

measurements were used in the subsequent analyses. The duration of the resulting motion

cycles was 1.11 s for gait and 3.20 s for stair.









Model Development

Dynamic contact models (gait and stair) and static contact models (kneel and lunge)

of the patient's knee implant were constructed to predict in vivo contact forces, pressures,

and areas on the medial and lateral contact surfaces of the tibial insert. The models were

implemented within the Pro/MECHANICA MOTION simulation environment (PTC,

Waltham, MA) (Figure 3-1). A six degree-of-freedom (DOF) joint between the fixed

femoral component and moving tibial insert was used to measure relative (i.e., joint)

kinematics for contact calculations. Femoral AP translation, internal-external rotation,

and flexion-extension were prescribed to match the fluoroscopically measured kinematics

while the other three DOFs were predicted via forward dynamic simulation (Fregly et al.,

2005). The location at which the axial force was applied to the back side of the tibial tray

was prescribed to match the center of pressure measured experimentally.

In Vivo Dynamic Computational
Data Contact Model Predictions




V Medial & lateral force
Axial force & CoP -



SPressure distribution



Kinematics
Center of pressure
Figure 3-1. The experimental and computational methods used to develop and evaluate
the dynamic contact model of the instrumented knee implant.









A custom deformable contact model based on elastic foundation theory (An et al.,

1990; Blankevoort et al., 1991; Fregly et al., 2003; Fregly et al., 2005; Komisteck et al.,

2005) was incorporated into the multibody dynamic model to predict contact forces and

pressures between the implant surfaces. The contact model utilized springs distributed

uniformly over the articulating surfaces of the tibial insert to prevent excessive

interpenetration, where each spring was associated with a single tibial surface element of

known area. The contact pressure p for each element was calculated from Eq. 2-1. The

three-dimensional contact force vector acting on each side of the tibial insert was

computed by multiplying the element pressures by their respective areas and performing

a vector summation over all elements. Medial and lateral contact forces were calculated

by taking the axial components of the two contact force vectors. Evaluation of the

calculated contact forces was performed by comparing measured and predicted CoP

locations for the entire tibial insert in the medial-lateral (ML) and anterior-posterior

directions.

Element contact pressures were calculated using linear and nonlinear material

models with a Poisson's ratio of 0.46 (Bartel et al., 1995; Kurtz et al., 2002). Both linear

and nonlinear models were used to study the sensitivity of predicted medial-lateral load

calculations to the choice of material model. For the linear material model, Young's

modulus was set to 463 MPa for all elements (Bartel et al., 1995; Kurtz et al., 2002). For

the nonlinear material model, Young's modulus was set to a different value for each

element depending on its current level of contact pressure. The relationship between

Young's modulus and contact pressure was derived from a modified nonlinear power law

material model (Eq. 2-2). Values for these parameters were set to sE =0.0257, p, =15.9










MPa, and n =3 based on fitting of experimental stress-strain data for polyethylene

(Cripton et al., 1993; Kurtz et al.; 2002). During the dynamic simulation, Eq. 2-3 was

solved for each element using a nonlinear root finding algorithm.

Results

Medial to Total Force Ratio

The ratio of medial to total contact force as calculated by the contact model varied

between as well as within activities. Predicted contact forces were insensitive to choice of

material model. Medial contact force followed the trend of total contact force better than

did lateral contact force for both gait and stair (Figure 3-2). Stair exhibited the largest

medial and lateral contact forces, followed by gait, lunge, and kneel. The ratio of medial

to total force (reported as a percentage) ranged from 18.0% to 60.4% for the entire gait

cycle, averaging 54.7% during mid-stance phase and 33% during swing phase, and from

46.9% to 64.6% for the entire stair cycle. The ratio was much more constant for stair than

for gait over a single motion cycle. At maximum load, the ratio was between 53% and

60% for all four activities (Table 3-1).


Total
-- Medial
S3 ---- Lateral


FU-


S\ -0 Stair Up- ----- Stair Down-
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Gait Cycle (%) A Stair Cycle (%) B
Figure 3-2. Medial and lateral contact forces calculated. A) During gait activity. B)
During stair activity.

Contact Pressure and CoP

Medial and lateral contact pressures (maximum and average) followed similar

trends to medial and lateral contact forces (Figure 3-3). In contrast to the contact force











results, maximum contact pressure results were highly sensitive to choice of material

model while average pressure results were moderately sensitive, especially for stair. For

all activities, the linear material model produced larger maximum and average pressures

than did the nonlinear material model except during periods of low axial loads (e.g.,

kneel and the swing phase of gait). For both material models, medial pressures were

larger than lateral ones during the mid-stance phase of gait and during the entire stair

cycle and smaller during lunge and kneel. Also in contrast to contact force results, lunge

exhibited the largest medial and lateral contact pressures, followed by stair, gait, and

kneel (Table 3-1, Figure 3- 4).

50 50
-- Medial linear
0-40 40 ---- Lateral linear
-- Medial nonlinear
---- Lateral nonlinear
30 30

S20 20
E
U 10 10

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Gait Cycle (%) A Gait Cycle (%) B
50 50

|40 (40
t \- t \ \
30 30

S20 ^^^ ^.rr.^- 20

U10 10
0-Stair Up -- Stair Down < +-Stair Up Stair Down1
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Stair Cycle (%) C Stair Cycle (%) D
Figure 3-3. Maximum and average contact pressures calculated. A) Linear material
model during gait. B) Nonlinear material model during gait. C) Linear
material model during gait. D) Nonlinear material model during stair.

Predicted CoP locations matched the experimental measurements over most of the

gait cycle and the entire stair cycle. Similar to contact forces, results from the two

material models were nearly identical. Without changing any fluoroscopically measured











motion inputs, the contact simulations reproduced the ML location of the CoP extremely

closely for all activities with root-mean-square (RMS) errors of 0.6 mm for gait and 0.0

mm for stair, kneel and lunge. In contrast, the simulations matched the AP location of the

CoP well only for the stance phase of gait and for stair, with RMS errors of 4.8 mm over

the entire gait cycle, 1.9 mm during stance phase, and 1.3 mm for stair. To reduce the AP

CoP error due to the error of the fluoroscopic measurement, the input AP translations to

the simulation were adjusted based on the AP CoP error prediction with the original AP

translation (Eq. 3-1).


Aod +1
AP= A ErrorA PCo
41d -1


20 40 60
Gait Cycle (%)


80 100


if ErrorAPP < -
if ErrorAPcoP < 1
if Error,P, >I


Gait Cycle (%)



O o





4-Stair Up ------Stair Down--


(3-1)


0 20 40 60 80 100 0 20 40 60 80 100
Stair Cycle (%) C Stair Cycle (%) D
Figure 3-4. Experimental and simulated center of pressure locations. A) In the anterior-
posterior direction for gait. B) In the medial-lateral direction for gait. C) In the
anterior-posterior direction for stair. D) In the medial-lateral direction for
stair. Results are plotted only for the linear material model since results for the
nonlinear material model were indistinguishable. The small plot in the lower
right plot shows the dimension and origin of the tibial tray.


10

-5



3.
0
0-10

-15
0


Experiment
-- Simulation
I"i
/$

^-'I


4- Stair Up -. 4-- Stair Down -









When the fluoroscopically measured AP translations were adjusted within their

range of experimental error ( 1 mm), RMS errors in the AP CoP location were reduced

to 3.5 mm over the entire gait cycle, 0.5 mm during stance phase, and 0.5 mm for stair

(Figure 5-4). At maximum load for each activity, CoP error in the AP direction was

smallest for stair and largest for kneel, ranging from 0.2 to 11.5 mm after AP translation

adjustment (Table 3-2).

Table 3-1. Medial-to-total peak pressure (Pmax) and average pressure (Pave) for gait,
stair, lunge, and kneel activities using linear and nonlinear material models.
For gait and stair, results are reported at maximum load. For kneel and lunge,
results are reported for mean load with additional results for minimum and
maximum load over the interval indicated in parentheses.
Pmax (MPa) Pave (MPa)
Material Model Activity Medial Lateral Medial Lateral
Gait 27.2 25.2 13.4 12.4
Stair 39.4 31.2 19.2 15.5
S15.6 20.5 8.7 10.7
Linear Kneel
(12.4, 17.2) (19.0, 26.1) (7.0,9.3) (10.6, 11.1)
37.0 67.0 20.2 28.6
Lunge (31.4,38.3) (64.5,80.9) (16.0,24.2) (28.1,32.2)
Gait 17.5 16.8 12.0 11.5
Stair 21.4 18.9 14.9 13.2
Non r K l 13.0 15.0 9.1 10.8
Nonlinear Kneel
(11.3, 13.7) (14.4, 16.1) (8.3, 9.6) (10.6, 11.2)
21.6 27.2 15.9 18.6
Lunge (19.4,23.8) (26.1,29.9) (14.4, 17.3) (17.4,20.1)


Discussion

In Vivo Knee Load Distribution

This study used a novel combination of an instrumented knee implant, fluoroscopic

motion analysis, and a dynamic contact model to determine in vivo medial and lateral

contact forces and pressures on the tibia during gait, stair rise/descent, kneel, and lunge

activities. Medial and lateral contact force calculations were insensitive to choice of

material model (linear or nonlinear), with the ratio of medial to total contact force being









close to 50% for high-loading phases of all activities. In contrast, contact pressure results

were highly sensitive to choice of material model and thus provided only an upper and

lower bound on the pressures one might expect to observe in vivo. Center of pressure

calculations were also insensitive to choice of material model, and the reliability of the

calculated medial and lateral contact forces was supported by the model's ability to

reproduce the experimentally measured AP and ML CoP locations. Though a constant

load split of 55% medial-45% lateral would be a reasonable approximation for the current

patient performing gait, stair rise/descent, kneel, and lunge activities, the reasonableness

of this approximation for the general patient population is not known.

Table 3-2. Medial-to-total force percentage, and anterior-posterior center of pressure
error for gait, stair, lunge, and kneel activities using linear and nonlinear
material models. For gait and stair, results are reported at maximum load. For
kneel and lunge, results are reported for mean load with additional results for
minimum and maximum load over the interval indicated in parentheses.
Material Load Medial/Total Force AP CoP Error (mm)
Model Activity (BW) (%) Original Adjusted
Gait 2.2 53.4 2.9 1.0
Stair 3.5 56.0 0.3 0.2
S0.3 57.0 11.6 10.4
Linear Kneel
(0.2, 0.4) (55.2, 59.6) (12.6,11.2) (11.5,10.1)
1.6 57.7 2.9 1.7
ge (1.2, 2.2) (56.6, 57.5) (3.1, 1.2) (2.0, 0.1)
Gait 2.2 53.4 2.8 1.1
Stair 3.5 56.0 0.3 0.2
Non r K l 0.3 57.0 11.6 10.5
Nonlinear Kneel
(0.2, 0.4) (55.2, 59.6) (12.6,11.3) (11.5,10.2)
1.6 57.7 3.2 2.1
Luge (1.2,2.2) (56.6, 57.5) (3.4, 1.6) (2.3, 0.5)


While previous studies reported that the majority of the load passes through the

medial compartment during gait, our study found a more equal medial-lateral load split.

To our knowledge, all previous estimates of medial-lateral load split during gait have

been based on biomechanical models without internal load measurements (Hsu et al.,









1990; Hurwitz et al., 1998; Johnson et al.,1980; Noyes et al., 1992; Shelbume et al.,

1991; Schipplein et al., 1991). Most estimates were derived from models that required

prediction of individual muscle forces, and these predictions may have been sensitive to

errors in muscle origin and insertion locations, muscle moment arms, and peak isometric

strength values. Furthermore, previous studies did not have a method for evaluating the

reasonableness of the predicted medial-lateral load split. In our study, prediction of

individual muscle forces was not needed due to the availability of internal force

measurements, and the load split calculations could be evaluated using the experimentally

measured CoP locations.

Choice of Material Model

In general, the nonlinear material model predicted lower, more evenly distributed

contact pressures than did the linear material model. Without in vivo contact pressure

measurements to evaluate the predictions, the pressure results can only be used to provide

approximate upper and lower bounds on in vivo contact pressures. Nonetheless, the trends

in contact pressures observed between different activities were the same for both material

models. Consequently, the pressure results are still valuable for assessing which types of

activities are likely to contribute the most significantly to mild wear. High contact

pressures occurred during knee and lunge because only the posterior tips of the femoral

component were in contact with the tibial insert during deep flexion. While such high

pressures would cause the polyethylene to yield, it is possible that these pressures occur

in real life and cause initial plastic deformation following implantation. Alternatively,

these high maximum pressures may indicate that the nonlinear material model provides a

more reasonable estimate of in vivo contact pressures than does the linear material model

at high loads.









CoP Error Analysis

Errors in the predicted CoP locations were small apart from the AP direction during

the swing phase of gait and during kneel. The CoP locations measured by the

instrumented implant have been shown to be accurate to within + 2 mm for loads ranging

from 800 to 2000 N (D'Lima et al., 2005). Maximum errors in predicted CoP location

were on the order of 10 mm in the AP direction for loads in the range of 150 N (swing

phase of gait) to 200 N (kneel). Since the experimental CoP calculation involves dividing

by the total axial load, the calculation will be most sensitive to axial load measurement

errors when the load is "small." Consequently, when we plotted AP CoP error as a

function of the inverse of the axial load, we found a high correlation (R2 = 0.80, Figure 3-

5). Thus, sensitivity of the experimental CoP calculations to errors in the measured axial

load may help explain the large AP CoP errors during periods of light loading. Data

averaging may have also contributed to AP CoP errors, consistent with the observation

that the unaveraged stair data set had the lowest errors. Since researchers are most

interested in phases of high rather than low loading, CoP errors during light loading

periods are not a critical limitation of the study.

Data Synchronization Analysis

While the computational procedure minimized errors in the predicted AP CoP

location for gait, this procedure made the data synchronization process objective without

having a significant influence on the results. The goal of performing this procedure was

to ensure that the internal force data and fluoroscopic kinematic data were synchronized

as accurately as possible. If there were no experimental or modeling errors present, we

would expect the model to reproduce perfectly the experimentally measured CoP

locations. Thus, we sought to determine whether a slightly different synchronization of









the data would alter the AP CoP errors significantly. As shown in Table 3-3, the

predicted AP CoP location changed by only 0.2 mm when switching from visual

synchronization (i.e., 0% phase shift) to synchronization based on minimum AP CoP

error (i.e., -2%). We believe that our objective approach for synchronizing the data is

more robust than an approach based solely on visual inspection of the data.

10


E 8
0



1) 4 -
<

0 2

O0

0 1 2 3 4 5 6 7

Load 103 N

Figure 3-5. Error in predicted anterior-posterior center of pressure location as a function
of the inverse of the applied axial load during gait.

Table 3-3. Sensitivity study of the data synchronization.
Phase shift +5% 0% -1% -2% -3% -5%
Flexion angle (Deg) 24.6 13.5 12.7 12.3 12.2 12.4
AP CoP error (mm) 3.4 2.1 2.0 1.9 1.9 2.4
ML CoP error (mm) 0.2 0.1 0.2 0.2 0.3 0.2

Video-based motion analysis does not provide kinematics inputs that are accurate

enough to perform contact analyses with the implant components. Comparison of surface

marker and fluoroscopic knee kinematics has revealed differences as large as 10 mm in

joint translations and 5 deg in joint rotations (Cereatti et al., 2006, Halloran et al., 2005).









Difference between Overground and Treadmill Gait

To evaluate how similar loading is between overground and treadmill gait, we

followed a two-step approach. First, we evaluated the accuracy with which medial

contact force can be estimated directly from the two medial load cell measurements.

There is no guarantee that the sum of the two medial load cell measurements will equal

the medial contact force calculated by the contact model. However, when we performed

the comparison for the treadmill gait data, we found that the two medial force curves

were nearly identical, with maximum and root-mean-square differences of 0.03 and 0.01

BW, respectively (Figure 3-6). When we performed the same comparison for the stair,

kneel, and lunge data, the maximum error was 0.14 BW with an RMS difference of 0.04

BW for stair. Thus, while we could not prove ahead of time that the sum of the two

medial load cell measurements essentially equals the medial contact force, our contact

model calculations strongly support this conclusion. Given this finding, we then used the

load cell measurements directly to compare the total axial load and medial-lateral load

split between treadmill and overground gait. The total axial load profiles were very

similar with an RMS difference 0.15 BW. In contrast, the medial-lateral load split during

overground gait was larger during mid-stance but comparable at the start and end of

stance phase. However, the average value of medial force ratio increased by only 3% for

the entire stance phase and by only 6% during mid stance. Thus, the medial-lateral load

split during overground gait was similar, though not identical, to that measured during

treadmill gait.

The primary limitation of this study is that only a single patient with an implanted

knee could be analyzed. This limitation is due to the difficulty of obtaining the in vivo

experimental measurements of implant motions and loads required as inputs to the










contact model. Although the patient's walking speed (1.24 + 0.03 m/s, Andriacchi et al.,

1982) and ground reaction forces were similar to those of normal subjects (Chao et al.,

1983), we are limited by having only a single subject with an artificial knee for analysis.

The extent to which these results apply to knee implant patients in general or to non-

implanted knees is not known. At a minimum, we can safely conclude that in well-

aligned implanted knees, it is possible to achieve an approximately 50-50 medial-lateral

load split during dynamic activities such as gait and stair and during static activities such

as kneel and lunge.

2.5 70

2 / \ / --Treadmill 60 -
Sl --Tdmlreadmill
/ Overground 50 /
1.5 --
0 -
40
0 -
I.- 30-
0.5- 20

C 10
0 20 40 60 80 100 0 20 40 60 80 100
Gait Cycle (%) A Gait Cycle (%) B
Figure 3-6. Total tibial load and medial to total load ratio. A) During treadmill gait. B)
During overground gait.

Sensitivity Study of Contact Force Prediction to Kinematic Errors

The calculated CoP location is dependent on the varus-valgus rotation and medial-

lateral translation predicted by the contact model, as well as on the predicted medial and

lateral contact forces. We did not use fluoroscopically measured motions for these

degrees of freedom nor for superior-inferior translation (Fregly et al., 2005). If the

predicted varus-valgus rotation or medial-lateral translation were significantly in error,

then the calculated CoP location in the medial-lateral direction in particular would have

been significantly in error as well. To mimic a rigid body analysis, we performed static

analyses for the initial time frame of gait but with medial-lateral translation or varus-









valgus rotation locked to specific values rather than left free. Table 3-4 summarizes the

tests that we performed:

* Test 1 corresponds to the way the degrees of freedom were handled in the dynamic
simulations. Fixed values were taken from the first time frame of the gait
simulation.

* Test 2 investigates whether a wrong contact model prediction of medial-lateral
translation would have produced a bad medial-lateral CoP prediction. The medial-
lateral translation was fixed at the nominal value found in Test 1 +/- 2 mm.

* Test 3 investigates whether a wrong contact model prediction of varus-valgus
rotation would have produced a bad medial-lateral CoP prediction. The varus-
valgus rotation was fixed at the nominal value found in Test 1 +/- 0.1 deg.

Table 3-4. Test cases of sensitivity study of contact force prediction to kinematic errors.
Degree of Freedom Test 1 Test 2 Test 3
Anterior-posterior Fixed Fixed Fixed
translation
Superior-inferior Free Free Free
translation
Medial-lateral Free Fixed at various Free
translation values
Flexion-extension Fixed Fixed Fixed
Internal-external Fixed Fixed Fixed
rotation
Varus-valgus Free Free Fixed at various
rotation values
The changes in the predicted medial-lateral CoP location as a function of locked

medial-lateral translation or varus-valgus rotation value are provided in table 3-5.

Table 3-5. CoP location as a function of locked medial-lateral translation or varus-valgus
rotation value.
Locked ML Translation Locked VV Rotation
Offset (mm) ML CoP Offset (deg) ML CoP
(mm) (mm)
2 26.9163 0.1 22.0988
1 16.0899 0.01 -1.15671
0.5 6.03472 0.001 -3.83896
0 -4.13701 0 -4.13701
-0.5 -13.5634 -0.001 -4.42925
-1 -25.1905 -0.01 -7.0102
-2 -34.6436 -0.1 -22.0979









As shown by this table, small errors in the fluoroscopically measured ML

translation and VV rotation would produce large errors in the predicted ML CoP location

(and hence in the predicted medial and lateral contact forces). For example, a 0.1 deg

change in varus-valgus rotation would cause 100% of the contact force to pass through

one condyle or the other. Since single plane fluoroscopy can measure ML translation to

within about 6.6 mm and varus-valgus rotation to within about 1.1 deg (Banks and

Hodge, 1996), using the fluoroscopically measured kinematics directly for these two

degrees of freedom would not produce ML CoP predictions consistent with the

experimental data. In contrast, leaving these degrees of freedom free to equilibrate

produces medial-lateral CoP (and hence medial-lateral contact force) predictions that are

consistent with the experimental data. The model's predictions for these two degrees of

freedom were within the error bars of their fluoroscopically measured motions,

supporting the reasonableness of this approach.

Conclusion

In summary, this chapter has presented a novel approach of combining in vivo load

and motion measurements with a computational model to predict the medial-lateral force

distribution in an implanted knee during a variety of activities. The in vivo medial-to-total

force ratio and contact pressure results may be valuable for improving our understanding

of knee joint mechanics and developing computational and experimental testing protocols

to assess wear performance of new knee implant designs.














CHAPTER 4
THE RELATIONSHIP BETWEEN THE KNEE ADDUCTION TORQUE AND
MEDIAL CONTACT FORCE DURING A VARIETY OF GAIT PATTERNS

The external knee adduction torque has been proposed as a surrogate measure for

medial compartment load during gait. However, a direct link between these two

quantities has not been demonstrated using in vivo measurement of medial compartment

load. This study uses in vivo data collected from a single subject with an instrumented

knee implant to evaluate this link. The subject performed five different overground gait

motions (normal, fast, slow, wide, and toe out) with simultaneous collection of

instrumented implant, video motion, and ground reaction data. For each trial, the knee

adduction torque was measured externally while the total axial force applied to the tibial

insert was measured internally. Based on data collected from the same subject performing

treadmill gait under fluoroscopic motion analysis, a regression equation was developed to

predict medial contact force from the implant load cell measurements. Pearson

correlation coefficients were calculated for the stance phase and entire gait cycle to

quantify the relationship between the knee adduction torque and both the medial contact

force and the medial to total contact force ratio. For the complete gait cycle, correlation

coefficients were between 0.83 and 0.96 (p < 0.001) for medial contact force and between

0.73 and 0.95 (p < 0.001) for medial to total force ratio. Correlation coefficients were

slightly lower for the stance phase alone. These results support the hypothesis that the

knee adduction torque is highly correlated with medial compartment contact force and

medial to total force ratio during gait.









Introduction

The human knee joint is critical for locomotion and is commonly affected by

osteoarthritis (OA). Adverse mechanical loading, and in particular high medial contact

force, is believed to contribute to the development of knee OA (Jackson et al., 2004). The

ability to measure or predict high medial contact force in individual patients would be

valuable for identifying those at highest risk for developing knee OA as well as for

devising new treatment approaches. Unfortunately, noninvasive in vivo measurement of

medial compartment contact force is not yet available.

For this reason, researchers have investigated the use of external measures available

from gait analysis as surrogates for internal medial force. To date, the peak knee

adduction torque has been identified as the best candidate, in part because of its ability to

predict OA disease progression (Sharma et al., 1998) and long-term outcome following

high tibial osteotomy surgery (Prodromos et al., 1985). Schipplein and Andriacchi (1991)

were the first to propose that the knee adduction torque is the primary determinant of

medial compartment load during gait. Their conclusions were based on medial

compartment load predictions made by a statically determinate muscle model. Using the

same model, Noyes et al. (1992) found a statistically significant correlation between the

peak knee adduction torque and the predicted peak medial compartment load in a group

of ACL-deficient patients. More recently, Hurwitz et al. (2002) showed that the peak

knee adduction torque was the best single predictor of the medial to lateral ratio of

proximal tibial bone density. While these results support the hypothesis that the knee

adduction torque during gait serves as a surrogate for medial compartment load, no study

has been able to correlate these quantities based on internal medial load measurements.









This paper investigates the relationship between the knee adduction torque and in

vivo medial contact force during normal, fast, slow, wide, and toe out gait. A single

patient with an instrumented knee implant provided a unique opportunity to perform the

investigation. The adduction torque curve for each gait pattern was obtained using

standard external gait measurements. The corresponding internal axial contact force was

measured by the instrumented knee implant (D'Lima et al., 2005). A linear regression

equation was used to determine medial contact force from the implant load cell

measurements, where the regression coefficients were found by using a dynamic contact

model to compute medial and lateral contact force from fluoroscopic and axial load data

collected from the same patient performing treadmill gait. The hypothesis tested was that

the knee adduction torque is highly correlated with medial compartment force and force

ratio (i.e., ratio of medial to total contact force) during a variety of gait activities.

Materials and Methods

Date Collection

Data were collected from a single patient with an instrumented knee implant (male,

right knee, age 80, mass 68 kg, height 1.705 m) eight months after surgery. Institutional

review board approval and patient informed consent were obtained. In vivo tibial force

data were recorded simultaneously with video motion (Motion Analysis Corporation,

Santa Rosa, CA) and ground reaction (AMTI Corporation, Watertown, MA) data for five

patterns of overground gait: normal (1.24 0.03 m/s), fast (1.52 0.04 m/s), slow (0.80

+ 0.05 m/s), wide stance (1.03 0.05 m/s), and toe out (1.10 + 0.05 m/s). These gait

patterns were chosen since walking speed and foot path have been shown to influence the

knee adduction torque (van den Bogert et al., 1994; Reinschmidt et al., 1997; Zeller et al.,

2003). Four uniaxis loads cells in the implant provided a measure of total axial load but









not the load distribution between the medial and lateral compartments. The Cleveland

Clinic marker set with additional markers placed on the foot segment was used to create

segment coordinate systems and provide three-dimensional movement data. Raw marker

and ground reaction force data were filtered using a fourth-order, zero phase-shift, low

pass Butterworth Filter with a cutoff frequency of 6 Hz (Reinschmidt et al., 1997; van

den Bogert et al., 1994; Zeller et al.; 2003). The subject performed three trials of each

gait pattern using a self-selected walking speed and foot path. For each trial, one

complete motion cycle, starting and ending with right heel strike, was chosen for

analysis.

External Knee Adduction Torque

The knee adduction torque was calculated using traditional bottom-up inverse

dynamics. The dynamical equations were derived for the foot-shank system with Autolev

(Levinson and Kane, 1990) using Kane's method. The foot possessed six degrees-of-

freedom relative to the laboratory-fixed coordinate system. The ankle joint was modeled

as two non-intersecting pin joints whose axes were found via optimization of additional

motion data collected from the patient performing an ankle circumduction movement

(Reinbolt et al., 2005). Foot and shank masses, mass centers, and moments of inertia

were estimated using regression relationships (De Leva et al., 1996). Optimal alignment

of the shank segment with the shank markers was performed using an approach based on

the singular value decomposition (Soderkvist et al., 1993). Subsequent alignment of the

foot segment with the foot markers was performed using optimization (Reinbolt et al.,

2005). The inverse dynamics reaction torque in the knee was calculated about the knee

joint center, defined as the midpoint between the medial and lateral femoral epicondyles.

The external knee adduction torque, which is due primarily the moment of the ground









reaction force vector about the knee center, was taken as the negative of the internal knee

abduction torque calculated from inverse dynamics.

Medial Contact Force

To determine medial contact force, we followed a two-step process to convert the

four implant load cell measurements into a medial force prediction. In the first step, we

used a dynamic contact model of the patient's knee implant to predict medial contact

force given fluoroscopic and instrumented implant data collected simultaneously from the

same patient during treadmill gait. The model was constructed within the

Pro/MECHANICA MOTION simulation environment (PTC, Waltham, MA) (Figure. 4-

1) and incorporated a custom elastic foundation contact model with nonlinear material

properties. A six degree-of-freedom (DOF) joint between the fixed femoral component

and moving tibial insert was used to measure relative (i.e., joint) kinematics for contact

calculations. Anterior-posterior translation, internal-external rotation, and flexion-

extension were prescribed to match the fluoroscopically measured kinematics while the

other three DOFs were predicted via forward dynamic simulation. The location at which

the measured axial force was applied to the tibial tray was prescribed to match the center

of pressure measured experimentally. The medial and lateral contact forces acting on the

tibial insert were calculated from the contact pressures acting across the surfaces.

In the second step, we used linear regression to relate medial contact force to the

four implant load cell measurements. The medial contact force (FM) predictions

generated for treadmill gait were fitted as a function of the four implant load cell

measurements (FAM, FPM, FAL, and FPL) using linear least squares (Eq. 4-1, R2 =

0.99, root-mean-square (RMS) error = 0.01 body weight (BW)).









F, = C1F4 + C2FPM + C3FAL + C4FPL (4-1)

The resulting coefficients Cl, C2, C3, and C4 in Eqn. 1 were then used to compute

medial contact force during overground gait from the four load cell measurements. The

dynamic contact model was not be used to predict medial contact force directly since

fluoroscopic measurement of internal implant motion was not available during

overground gait.




V Gait Dynamic
Analysis Contact Model





SRegression
Model

Load measurements




| Adduction torque Medial contact force



Figure 4-1. The experimental and computational methods used to calculate the knee
adduction torque and medial contact force during gait.

Statistics Analysis

To analyze the relationship between the knee adduction torque and the medial

contact force or medial to total force ratio within the gait cycle, we performed a linear

regression analysis with Pearson correlation coefficients on each of the 15 individual gait









trials. Also, two-tailed Student's t-test (a = 0.05) were used to determine the statistical

significance of the reductions in peak knee adduction torques and medial contact force

between difference walking patterns.

Results

Peak total and medial axial load as well as peak adduction torque varied between

gait patterns (4-1). Among the fifteen trials, the largest peak axial force was 2.74 BW,

which occurred during one fast gait trial, with a corresponding peak medial force of 1.73

BW. The smallest peak axial force was 2.06 BW, which occurred during one slow gait

trial, with a corresponding peak medial force of 1.28 BW. For all trials, the shape of the

medial contact force curve closely followed that of the total contact force curve (Figure

4-2), and for 10 of the 15 trials, both force curves exhibited two distinct peaks. The

corresponding adduction torque curves exhibited two distinct peaks for 9 of the 15 trials.

However, the shape of each adduction torque curve did not necessarily follow that of the

corresponding medial force curve, nor did the adduction torque peaks necessarily match

those of the medial force (Figure 4-2). Among the 15 trials, the largest adduction torque

peak was 3.37 %BW*H, which occurred during one toe out gait trial, while the smallest

peak was 2.13 %BW*H during one fast gait trial.

Table 4-1. The average and standard deviation of the peaks of measured total force,
predicted medial force, and calculated adduction torque.
Normal Fast Slow Wide Toe Out
Total Force (BW) 2.490.14 2.590.13 2.350.27 2.470.31 2.440.25
Medial Force (BW) 1.580.10 1.630.09 1.51+0.21 1.660.13 1.570.14
Torque (%BW*H) 2.560.18 2.530.40 2.71+0.29 2.440.20 3.130.41

The right knee adduction torque curves exhibited similar trends to medial contact

force for all the gait patterns. The slow, wide, and toe out gait adduction torque curves









did not have the common two peaks as the normal and fast gait did. The toe out gait had

the biggest peak measured total axial load while gait had the smallest (Figure 4-2). The

two peaks in the adduction torque curve during stand phase occurring 4 to 6% early in the

cycle than did the corresponding peaks in the medial contact force curve for fast gait.

Table 4-2. The average and standard deviation of the correlation coefficients between the
internal knee abduction torque (i.e., negative of the external knee adduction
torque) and medial contact force and or medial to total force ratio for a variety
of gait patterns. (p < 0.001 for all the cases).
Medial Force Medial Force Ratio
Gait
Stance Phase Full Cycle Stance Phase Full Cycle
Normal 0.870.05 0.940.02 0.900.06 0.930.03
Fast 0.720.09 0.890.03 0.820.10 0.880.04
Slow 0.840.05 0.920.02 0.880.02 0.930.01
Wide 0.81+0.04 0.870.02 0.770.03 0.780.05
Toe Out 0.770.06 0.860.04 0.870.03 0.900.02

Statistically significant correlations (p < 0.001 for all trials) were found between

the knee adduction torque and both the medial contact force and the medial to total

contact force ratio (Table 4-2). When the entire gait cycle was analyzed for each trial, the

Pearson correlation coefficient between adduction torque and medial contact force ranged

from 0.83 to 0.96. For the medial to total force ratio, the range was from 0.73 to 0.95.

When analysis of each trial was repeated using only the stance phase, the Pearson

correlation coefficients were slightly lower, ranging from 0.62 to 0.92 for medial force

and from 0.71 to 0.95 for medial to total force ratio (Figure 4-3). When the hypothesis

that fast, slow, wide, and toe out walk patterns can change the first and second peaks of

knee adduction torque and medial contact force during the normal gait was test using the

two-tailed Student's t-test, the results indicated that the null hypothesis could not be

rejected at the 5% significance level (Table 4-3).






















0 20 40 60
Gait Cycle (%)


0 L ----------
0 20 40 60
Gait Cycle (%)
75

60

O 45
01)
L? 30

g 15
75


80 100


0 20 40 60
Gait Cycle (%)


80 100
B


80 100


Gait Cycle (%)


75

60
o
Y 45

L 30

g 15
2


0 20 40 60 80 100 0 20 40 60 80 100
Gait Cycle (%) E Gait Cycle (%) F
Figure 4-2. Measured total and predicted medial contact force and internal abduction
toque at the knee joint. A) Medial contact force during one normal gait. B)
Medial contact force during one toe out gait trial. C) Internal abduction torque
during one normal gait. D) Internal abduction torque during one toe out gait
trial. E) Medial force ratio during one normal gait. F) Medial force ratio
during one toe out gait trial.

Discussion

This study used a combination of an instrumented knee implant, video-based


motion analysis, fluoroscopic motion analysis, and a dynamic contact model to evaluate

the hypothesis that the knee adduction torque measured externally can be used as an


indicator of internal medial contact force and medial to total contact force ratio during










gait. The large correlation coefficients found in our study strongly support this

hypothesis. No assumptions about muscle origin and insertion sites, moment arms,

strength, or indeterminacy were required to predict medial contact force (Noyes et al.,

1992, Schipplein and Andriacchi, 1991), which instead was calculated directly from the

load cell measurements provided by the instrumented implant. Our findings may be

useful for studies that seek to monitor or alter medial compartment load in response to

various treatment paradigms for medial compartment osteoarthritis.

75 75

60 60
o o
rO 45 rO 45

S30 \ 30

) 15 o ) 15

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Gait Cycle (%) A Gait Cycle (%) B
75 75
R= 0.95 R= 0.88 "". *"
S60 p< 0. 60 p< 0.001



S30. 30
0 O*
S15 i- 15-
0 0
-1 0 1 2 3 4 -1 0 1 2 3 4
Adduction Torque (%BW*H) C Adduction Torque (%BW*H)
Figure 4-3. The correlation between internal knee abduction torque and medial contact
force or medial to total force ratio. A) For one normal gait trial. B) For one toe
out gait trail. These two trials are the same as those shown in Figure 4-2.

Table 4-3. Student's t-test results of the subject's normal walking to the other walking
patterns.
t value Adduction Torque Peaks Medial Force Peaks
First Second First Second
Fast -1.85 0.55 -1.97 1.70
Slow -2.98 -0.75 1.35 0.22
Wide 0.98 0.77 2.69 -0.96
Toe Out -6.20 0.56 -0.35 0.42









In contrast to our study, previous studies investigated the correlation between peak

knee adduction torque and peak medial compartment load during only the stance phase of

gait (Noyes et al., 1992, Schipplein and Andriacchi 1991). These studies used larger

numbers of subjects to facilitate statistical analysis of the data. When we repeated our

linear regression analyses using only the peak values of knee adduction torque and

medial compartment load, we did not find any statistically significant correlations.

However, distinctive peaks in adduction torque and medial load did not exist for all trials,

and only a single patient with a knee replacement was analyzed due to the need for

internal load data. By analyzing within-cycle data for each trial, we were still able to

investigate the correlation between the knee adduction torque and medial compartment

load in a specific subject, even though large changes in peak values were not produced by

the different gait patterns.

A previous study also reported a strong correlation between peak external knee

extension torque (i.e., inverse dynamics knee flexion torque) and peak internal medial

compartment load during gait (Noyes et al., 1992). To assess whether a similar strong

correlation exists for our within-cycle data, we calculated the knee flexion/extension

torque from inverse dynamics for the 15 trials. For the entire cycle, Pearson's correlation

coefficients between the flexion/extension torque and medial contact force ranged from -

0.34 (p < 0.001) to 0.54 (p < 0.001). When only the stance phase was analyzed, the

correlations ranged from -0.37 (p = 0.002) to 0.40 (p < 0.001). Analysis of either the

magnitude of the flexion-extension curve or the medial force ratio produced weaker

correlations. Thus, given the wide variability in these correlation results, we did not find









knee flexion/extension torque to be a consistent indicator of medial compartment load for

our subject.

We also assessed the extent to which the external knee adduction torque is balanced

by internal medial and lateral contact forces. To perform the assessment, we calculated

the moment of the four axial forces, as measured the implant load cells, about the

geometric center of the insert and took the component in the anterior-posterior direction.

For the complete gait cycle, the correlation coefficients between the external knee

adduction torque and the internal moment due to axial contact forces were between 0.70

(p < 0.001) and 0.92 (p < 0.001). However, the magnitude of the internal moment was

roughly a factor of 10 smaller than that of the external adduction torque. This finding

suggests that muscles and ligaments may play a critical role in balancing knee loads in

the coronal plane. Future data collected with an implant capable of measuring six

components of internal knee load will be used to evaluate this initial hypothesis.

While most studies use inverse dynamics to calculate the knee adduction torque, it

is also possible to use the moment of the ground reaction force vector about the knee

center as an estimate. We therefore investigated whether use of this alternate approach

would significantly alter the results of our study. We treated the ground reaction force

vector as a bound vector applied at the center of pressure under the foot. After calculating

the moment of this vector about the knee center, we used the component in the anterior-

posterior direction of the shank to approximate the adduction torque. During stance phase

where the alternate method produces non-zero results, the adduction torque trends for the

two methods were similar, with an RMS difference of 0.28 + 0.09 %BW*H for the 15

trials. When we repeated the linear regression analyses, the Pearson's correlation









coefficients over the 15 trials changed by -0.01 0.02 for medial contact force and by

0.01 0.02 for medial to total contact force ratio. Thus, omission of inertia forces acting

on the shank and foot segments would not have altered our adduction torque or

correlation coefficient results significantly.

The five different gait patterns analyzed in our study were selected because of their

potential to alter the peak knee adduction torque. Increased toe out angle (Hurwitz et al.,

2002; Prodromos et al., 1985) and reduced walking speed (Wang et al., 1990) have

received the most attention as possible ways to reduce the adduction torque. Walking

with an increased stance width may also have a minor positive effect. Compared to the

subject's normal gait pattern, these walking modifications did not produce clear

adduction torque reductions (Table 4-3). However, we only had access to a single

subject, the patient had an artificial rather than natural knee, and the subject performed

only three trials of each gait pattern. More extensive data would be required to draw any

conclusions about the effectiveness of these gait modifications for reducing the knee

adduction torque and medial contact force in the general population.

Three important modeling assumptions were involved in the generation of the

medial contact force predictions. The first was that the dynamic contact model could

predict accurate medial and lateral contact forces during treadmill gait. We evaluated this

assumption by comparing the total contact force and center of pressure measured by the

instrumented implant with the same quantities predicted by the contact model. Over the

entire cycle, the RMS error in total contact force was 0.002 BW, while the RMS error in

medial-lateral and anterior-posterior center of pressure was within 0.6 mm. A second but

related assumption was that the contact force predictions were insensitive to choice of









polyethylene material model. When we compared contact force predictions made using

linear and nonlinear polyethylene material models, RMS differences were within 0.003

BW. The third important modeling assumption was that a linear regression model could

be used to calculate medial contact force from the four implant load cell measurements.

To evaluate this assumption, we used additional fluoroscopic data collected from the

same subject performing a stair rise activity. When medial contact force calculated by the

dynamic contact model was compared to medial contact force calculated by the

regression equation derived from treadmill gait, the RMS difference between the two

approaches was 0.04 BW. In addition, the RMS difference between the total contact force

measured during overground gait and the sum of the medial and lateral contact forces

predicted by separate regression equations was 0.002 BW. Given the small magnitude of

these differences, we do not believe that our modeling assumptions had a significant

influence on our results.

Conclusion

In summary, this study demonstrated that the knee adduction torque measured

externally is highly correlated with internal medial compartment force as well as medial

to total force ratio. The extent to which these results apply to the general population, to

knee OA patients, or to activities other than gait is not known. Despite these limitations,

our results strongly support the hypothesis that the knee adduction torque can be used as

a surrogate measure for medial compartment load during gait.














CHAPTER 5
CONCLUSION

This dissertation presents analyses of in vitro and in vivo function of total knee

replacements using dynamic contact model. A novel approach of combining in vivo load

and motion measurements with a computational model to predict the medial-lateral force

distribution in an implanted knee during a variety of activities is presented. The in vivo

medial-to-total force ratio and contact pressure results may be valuable for improving our

understanding of knee joint mechanics and developing computational and experimental

testing protocols to assess wear performance of new knee implant designs.

The fact that knee adduction torque measured externally is highly correlated with

internal medial compartment force as well as the medial to total force ratio was

demonstrated. The extent to which these results apply to the general population, to knee

OA patients, or to activities other than gait is not known. Despite these limitations, our

results strongly support the hypothesis that the knee adduction torque can be used as a

surrogate measure for medial compartment load during gait.

Though the elastic foundation model is simple, it is able to predict contact pressure

efficiently with satisfactory accuracy. Therefore it is a good candidate for the joint

mechanics study of the human musculoskeletal model. It is also effective for wear

analysis of TKR both in vivo and in vitro combining with the Archard's wear law and

creep model. The wear model has been proven to be very accurate. However,

experimental data are not sufficient for the UHMWPE creep modeling. To reduce the

simulation time of surface-updating dynamic analysis, a custom developed numerical






62


integrator or other simulation methods (e.g., surrogate model) may be used to replace the

current commercial numerical integrator.















LIST OF REFERENCES


An, K.N., Himenso, S., Tsumura, H., Kawai, T., and Chao, E.Y.S., 1990. Pressure
distribution on articular surfaces: application to joint stability analysis. Journal of
Biomechanics 23, 1013-1020.

Andriacchi, T.P., Stanwyck T.S., Galante J.O., 1986. Knee biomechanics and total knee
replacement. Journal of Arthroplasty 1, 211-9.

Andriacchi, T.P., 1994. Dynamics of knee malalignment. Orthopedic Clinics of North
America 25, 395-403.

Archard, J.F. and Hirst, W., 1956. The wear of metals under unlubricated conditions.
Proceedings of the Royal Society A236, 397-410.

Banks, S.A. and Hodge., W.A., 1996. Accurate measurement of three-dimensional knee
replacement kinematics using single-plane fluoroscopy. IEEE Transactions on
Biomedical Engineering 43, 638-649.

Banks, S.A., Markovich G.D., and Hodge W.A., 1997. The mechanics of knee
replacements during gait: in vivo fluoroscopic analysis of two designs. American
Journal of Knee Surgery 10, 261-267.

Banks, S.A., Markovich G.D., and Hodge W.A., 1997. In vivo kinematics of cruciate-
retaining and substituting knee arthroplasties. Journal of Arthroplasty 12, 297-304.

Barbour, P.S., Barton, D.C. Fisher, J., 1997.The influence of stress conditions on the
wear of UHMWPE for total joint replacements. Journal of materials science.
Materials in medicine. 8, 603-611.

Barnett, P. I., McEwen, H.M., Auger, D.D., Stone, M.H., Ingham, E., Fisher, J., 2002.
Investigation of wear of knee prostheses in a new displacement/force-controlled
simulator. Proceedings of the Institution of Mechanical Engineers. Part H, Journal
of engineering in medicine. 216, 51-61.

Bartel, D.L., Burstein, A.H. and Edwards, D.L., 1985. The effect of conformity and
plastic thickness on contact stress in metal-backed plastic implants. Journal of
biomechanical Engineering 107, 193-199.

Bartel, D., Rawlinson J.J., Burstein A.H., Ranawat C.S., and Flynn W.F., 1995. Stresses
in polyethylene components of contemporary total knee replacements. Clinical
orthopaedics and related research 317, 76-82.









Bei, Y. and Fregly, B.J., 2004. Multibody dynamic simulation of knee contact mechanics.
Medical Engineering and Physics 26, 777-789.

Blankevoort, L., Kuiper, J.H., Huiskes, R., and Grrotenbeor, H.J., 1991. Articular contact
in a three-dimensional model of the knee. Journal of Biomechanic 24, 1019-1031.

Blunn, G.W., Walker, P.S., Joshi, A., Hardinge, K., 1991. The dominance of cyclic
sliding in producing wear in total knee replacements. Clinical and Orthopaedic
Related Research 273, 253-260.

Burgess I.C., Kolar M., Cunningham J.L., Unsworth A.,1997. Development of a six
station knee wear simulator and preliminary wear results. Proceedings of the
Institution of Mechanical Engineers. Part H, Journal of engineering in medicine.
211, 37-47.

Cripton, P.A., 1993. Compressive characterization of ultra-high molecular weight
polyethylene with applications to contact stress analysis of total knee replacements,
Queen's University, Kingston, Ontario.

D'Lima, D.D., Patil S, Steklov N, Slamin JE, Colwell CW Jr., 2006. Tibial forces
measured in vivo after total knee arthroplasty. Journal of Arthroplasty 21, 255-62.

D'Lima, D.D., Townsend, C.P., Arms, S.W., Morris, B.A., Colwell, C.W. Jr., 2005. An
implantelemetry device to measure intra-articular tibial forces. Journal of
Biomechanics 38, 299-304.

De Leva, P., 1996. Adjustments to Zatsiorsky-Seluyanov's segment inertia parameters.
Journal of Biomechanic 29, 1223-30.

Dennis, D., Komistek R.D., Mahfouz M.R., 2003. In vivo fluoroscopic analysis of fixed-
bearing total knee replacements. Clinical and Orthopaedic Related Research 410,
114-130.

DesJardins, J.D., Walker, P.S., Haider, H., Perry, J., 2000. The use of a force-controlled
dynamic knee simulator to quantify the mechanical performance of total knee
replacement designs during functional activity. Journal of Biomechanics 28, 1231-
1242.

Endo, M. M., Barbour, P. S., Barton, D. C., Wroblewski, B. M., Fisher, J., Tipper, J. L.,
Ingham, E., Stone, M. H., 1999. A comparison of the wear and debris generation of
GUR 1120 (compression moulded) and GUR 4150HP (ram extruded) ultra high
molecular weight polyethylene. Bio-medical Materials and Engineering 9, 113-124.

Fregly, B.J., Bei, Y., and Sylvester, M.E, 2003. Experimental evaluation of a multibody
dynamic model to predict contact pressures in knee replacements. Journal of
Biomechanics 36, 1659-1668.









Fregly, B.J., Sawyer, W.G., Harman, M.K., Banks, S.A., 2005. Computational wear
prediction of a total knee replacement from in vivo kinematics. Journal of
Biomechanics, 305-314.

Fregly, B.J., Reinbolt J.A., Koh B.I., 2006. Chmielewski TL Evaluation of a patient-
specific cost function to predict the influence of foot path on the knee adduction
torque during gait. In Proceedings of the 7th International Symposium on
Computer Methods in Biomechanics and Biomedical Engineering. Juan-les-Pins,
France, 137

Fisher, J., McEwen, H. M., Tipper, J. L., Galvin, A. L., Ingram, J., Kamali, A., Stone, M.
H. and Ingham, E., 2004, Wear, debris, and biologic activity of cross-linked
polyethylene in the knee: benefits and potential concerns. Clinical Orthopaedics
and Related Research 428, 114-119.

Hsu, R. W., Himeno, S., Coventry, M.B., Chao, E.Y., 1990. Normal axial alignment of
the lower extremity and load-bearing distribution at the knee. Clinical
Orthopaedics and Related Research 255, 215-227.

Hurwitz, D.E., Sumer, D.R., Andriacchi, T.P., and Sugar, D.A., 1998. Dynamic knee
loads during gait predict proximal tibial bone distribution. Journal of Biomechanics
31,423-430.

Hurwitz, D.E., Ryals A. B., Case J. P., Block J. A., and Andriacchi T. P., 2002. The knee
adduction moment during gait in subjects with knee osteoarthritis is more closely
correlated with static alignment than radiographic disease severity, toe out angle
and pain. Journal of Orthopeadic Research 20, 101-107.

Jackson, B. D., Wluka, A.E., Teichtahl, A.J., Morris, M.E., Cicuttini, F.M., 2004.
Reviewing knee osteoarthritis-a biomechanical perspective. Journal of Sports
Science and Medicine 7, 347-357.

Johnson, F., Leitl., S., Waugh, W, 1980. The distribution of load across the knee: A
comparison of static and dynamic measurements. Journal of Bone and Joint
Surgery 62B, 346-349.

Johnson, K. L., 1985. Contact Mechanics. Cambridge University Press. Cambridge.

Komistek, R.D., Kane, T.R., Mahfouz, M., Ochoa, J.A., Dennis, D.A., 2005. Knee
mechanics: a review of past and present techniques to determine in vivo loads.
Journal of Biomechics 38, 215-228.

Kurtz, S. M., Jewett, C.W., Bergstrom, J.S., Foulds, J.R., Edidin, A.A., 2002. Miniature
specimen shear punch test for UHMWPE used in total joint replacements.
Biomaterials 23, 1907-1919.









Lancaster, J.G., Dowson, D., Isaac, G.H. and Fisher, J., 1997. The wear of ultra-high
molecular weight polyethylene sliding on metallic and ceramic counterfaces
representative of current femoral surfaces in joint replacement. Proceedings of the
Institution of Mechanical Engineers. Part H, Journal of engineering in medicine
211, 17-24.

Lee, K.Y. and Pienkowski, D., 1997. Reduction in the initial wear of ultrahigh molecular
weight polyethylene after compressive creep deformation. Wear, 203-204, 375-
379.

Lee, K.Y. and Pienkowski, D., 1998. iscoelastic recovery of creep-deformed ultra-high
molecular weight polyethylene (UHMWPE). ASTM Special Technical.
Publication, 1307, pp. 30-36.

Lee, K.Y. and Pienkowski., D., 1998. Compressive creep characteristics of extruded
ultrahigh molecular-weight polyethylene. Journal of Biomedical Material Research
39, 261-265.

Levinson, D. and. Kane, T., 1990. AUTOLEV a new approach to multibody dynamics,
pp. 81-102. In W. Schiehlen, Ed.. Multibody Systems Handbook, Springer, Berlin.

Morrison, J. B., 1970. The mechanics in knee joint in relation to normal walking. Journal
of Biomechanics 3, 51-60.

Mundermann, A., Dyrby CO, Hurwitz DE, Sharma L, Andriacchi TP., 2004. Potential
strategies to reduce medial compartment loading in patients with knee osteoarthritis
of varying severity: reduced walking speed. Arthritis Rheum 50, 1172-1178.

Muratoglu O.K., Perinchief R.S., Bragdon C.R., O'Connor D.O., Konrad R., Harris W.H.,
2003. Metrology to quantify wear and creep of polyethylene tibial knee inserts.
Clinical Orthopaedics and Related Research 410, 155-164.

Noyes, F.R., Schipplein, O.D., Andriacchi, T.P., Saddemi, S.R., and Weise, M., 1992.
The anterior cruciate ligament-deficient knee with varus alignment. An analysis of
gait adaptations and dynamic joint loadings. American Journal of Sports Medince
20,707-716.

Nuio, N. and Ahmed, A.M., 2001. Sagittal profile of the femoral condyles and its
application to femorotibial contact analysis. Journal of Biomechanical Engineering
123, 18-26.

Perie, D. and H., M. C., 1998. In vivo determination of contact areas and pressure of the
femorotibial joint using non-linear finite element analysis. Clinical Biomechanics
13,394-402.

Prodromos, C.C., Andriacchi, T.P., and Galante, J.O., 1985. A relationship between gait
and clinical changes following high tibial osteotomy. Journal of Bone and Joint
Surgery 67A, 1188-1194.









Rawlinson, J.J. and Bartel, D.L., 2002. Flat medial-lateral conformity in total knee
replacements does not minimize contact stress. Journal of Biomechanics 35, 27-34.

Reinbolt, J.A., Schutte, J. F., Fregly, B. J., Koh, B. I., Haftka, R. T., George, A. D., and
Mitchell, K. H., 2005. Determination of patient-specific multi-joint kinematic
models through two-level optimization. Journal of Biomechanics 38, 621-662.

Reinschmidt C, van den Bogert, A., Nigg B.M., Lundberg A., Murphy N., 1997. Effect of
skin movement on the analysis of skeletal knee joint motion during running.
Journal of Biomechanic 30, 729-32.

Saikko, V. and Ahlroos, T., 2000. Wear simulation of UHMWPE for total hip
replacement with a multidirectional motion pin-on-disk device: effects of
counterface material, contact area, and lubricant. Journal of biomedical materials
research 49, 147-154.

Schipplein, O.D. and Andriacchi, T.P., 1991. Interaction between active and passive knee
stabilizers during level walking. Journal of Orthopaedic Research 9, 113-119.

Sharkey P.F., Hozack W.J., Rothman R.H., Shastri S., Jacoby S.M., 2002. Why are total
knee arthroplasties failing today? Clinical Orthopaedics and Related Research 404,
7-13.

Sharma, L., Hurwitz, D.E., Thonar, E. J-M. A., Sum, J.E., Lenz, M.E., Dunlop, D.D.,
Schnitzer, T.J., Kirwan-Mellis, G. and Andriacchi, T.P., 1998. Knee adduction
moment, serum hyaluronan level, and disease severity in medial tibiofemoral
osteoarthritis. Arthritis and Rheumatism 41, 1233-1240.

Shelburne, K.B., Torry, M.R., Pandy, M.G., 2005. Muscle, ligament, and joint-contact
forces at the knee during walking. Medicine and Science in Sports and Exercise 37,
1948-1956.

Soderkvist, I. and Wedin PA., 1993. Determining the movements of the skeleton using
well-configured markers. Journal ofBiomechanic 26, 1473-1477.

Taylor, S.J., Perry, J.S., Meswania, J.M., Donaldson, N., Walker, P.S., Cannon, S.R.,
1997. Telemetry of forces from proximal femoral replacements and relevance to
fixation. Journal ofBiomechanics 30, 225-234.

Taylor, S.J., Walker, P.S., Perry, J.S., Cannon, S.R., Woledge, R., 1998. The forces in the
distal femur and the knee during walking and other activities measured by
telemetry. Journal of Arthroplasty 13, 428-437.

van den Bogert, A., Smith GD, Nigg BM., 1994. In vivo determination of the anatomical
axes of the ankle joint complex: an optimization approach. Journal of Biomechanic
27, 1477-1488.






68


Walker, P. S., 1997. Knee simulating machine for performance evaluation of total knee
replacements. Journal of Biomechanics 30, 83-89.

Wang, J.W., Kuo, K.N., Andriacchi, T.P., and Galante, J.O., 1990. The influence of
walking mechanics and time on the results of proximal tibial osteotomy. Journal of
Bone and Joint Surgery 72A, 905-913.

Zeller, B., McCrory JL, Kibler WB, Uhl TL., 2003. Differences in kinematics and
electromyographic activity between men and women during the single-legged
squat. American Journal of Sports Medicine 31, 449-456.















BIOGRAPHICAL SKETCH

Dong Zhao was born on Nov. 7, 1977, in Anshan, a city of industry, in northeast

China. He has been fascinated by mechanical systems since he was a kid. His father's

tools were his favorite toys, and he enjoyed disassembling and reassembling bicycles,

watches, and motorcycles in his teens. These days, he disassembles and reassembles

things such as computer hard drives and car engines for fun. His interests led him to

choose mechanical engineering as his undergraduate major at Beijing University of

Aeronautics and Astronautics, one of the best engineering universities in China. He

received his B.S. in July 1999. Realizing the importance of electronic systems in

mechanical engineering, he studied robotics as a graduate. He got his M.S. in April 2002.

His exposure to bioengineering-oriented robot projects made him decide on a long-term

goal of designing medical and orthopaedic devices for patients. He believes computer

simulations are the next step for improving the design process of mechanical and

electronic systems. In August 2002, he went overseas and enrolled in the Ph.D. program

in the mechanical and aerospace engineering department at the University of Florida. He

joined Dr. B.J. Fregly's Computational Biomechanics Laboratory and worked on

dynamic contact model and damage model development, in vivo human knee joint load

distribution analyses, and computer simulations of knee simulator machines.