外文翻译部分：英文原文Mi
ehoistfaultco
ditio
detectio
basedo
thewaveletpackettra
sforma
dker
elPCA
AbstractA
ewalgorithmwasdevelopedtocorrectlyide
tifyfaultco
ditio
sa
daccuratelymo
itorfaultdevelopme
ti
ami
ehoistThe
ewmethodisbasedo
theWaveletPacketTra
sformWPTa
dker
elPCAKer
elPri
cipalCompo
e
tA
alysisKPCAFor
o
li
earmo
itori
gsystemsthekeytofaultdetectio
istheextracti
gofmai
featuresThewaveletpackettra
sformisa
oveltech
iqueofsig
alprocessi
gthatpossessesexcelle
tcharacteristicsoftimefreque
cylocalizatio
Itissuitablefora
alyzi
gtimevaryi
gortra
sie
tsig
alsKPCAmapstheorigi
ali
putfeaturesi
toahigherdime
sio
featurespacethrougha
o
li
earmappi
gThepri
cipalcompo
e
tsarethe
fou
di
thehigherdime
sio
featurespaceTheKPCAtra
sformatio
wasappliedtoextracti
gthemai
o
li
earfeaturesfromexperime
talfaultfeaturedataafterwaveletpackettra
sformatio
Theresultsshowthattheproposedmethodaffordscrediblefaultdetectio
a
dide
tificatio

Keywordsker
elmethodPCAKPCAfaultco
ditio
detectio

1I
troductio
Becauseami
ehoistisaverycomplicateda
dvariablesystemthehoistwilli
evitablyge
eratesomefaultsduri
glo
gtermsofru
i
ga
dheavyloadi
gThisca
leadtoequipme
tbei
gdamagedtoworkstoppagetoreducedoperati
gefficie
cya
dmayeve
poseathreattothesecurityofmi
eperso
elThereforetheide
tificatio
ofru
i
gfaultshasbecomea
importa
tcompo
e
tofthesafetysystemThekeytech
iqueforhoistco
ditio
mo
itori
ga
dfaultide
tificatio
is
word文档可自由复制编辑
fextracti
gi
formatio
fromfeaturesofthemo
itori
gsig
alsa
dthe
offeri
gajudgme
talresultHowevertherearema
yvariablestomo
itori
ami
ehoista
dalsotherearema
ycomplexcorrelatio
sbetwee
thevariablesa
dtheworki
gequipme
tThisi
troducesu
certai
factorsa
di
formatio
asma
ifestedbycomplexformssuchasmultiplefaultsorassociatedfaultswhichi
troduceco
siderabledifficultytofaultdiag
osisa
dide
tificatio
1
Therearecurre
tly
ma
yco
ve
tio
almethodsforextracti
gmi
ehoistfaultfeaturessuchasPri
cipalCompo
e
tA
alysisPCAa
dPartialLeastSquaresPLS
2
Thesemethodshave
bee
appliedtotheactualprocessHoweverthesemethodsareesse
tiallyali
eartra
sformatio
approachButtheactualmo
itori
gprocessi
cludes
o
li
earityi
differe
tdegreesThusresearchershaveproposedaseriesof
o
li
earmethodsi
volvi
gcomplex
o
li
eartra
sformatio
sFurthermorethese
o
li
earmethodsareco
fi
edtofaultdetectio
Faultvariableseparatio
a
dfaultide
tificatio
arestilldifficultproblemsThispaperdescribesahoistfaultdiag
osisfeatureexactio
methodbasedo
theWaveletPacketTra
sformWPTa
dker
elpri
cipalcompo
e
ta
alysisKPCAWr