欢迎来到沃文网! | 帮助中心 分享知识,传播智慧!
沃文网
换一换
首页 沃文网 > 资源分类 > DOC文档下载
 

pid控制外文加中文文献中文翻译.doc

  • 资源ID:1374       资源大小:405.12KB        全文页数:26页
  • 资源格式: DOC        下载权限:注册会员/VIP会员    下载费用:10积分 【人民币10元】
下载资源需要10积分 【人民币10元】
已注册用户请登录:
账号:
密码:
  忘记密码?
    
友情提示
2、PDF文件下载后,可能会被浏览器默认打开,此种情况可以点击浏览器菜单,保存网页到桌面,既可以正常下载了。
3、本站不支持迅雷下载,请使用电脑自带的IE浏览器,或者360浏览器、谷歌浏览器下载即可。
4、本站资源下载后的文档和图纸-无水印,预览文档经过压缩,下载后原文更清晰   

pid控制外文加中文文献中文翻译.doc

1PIDcontrollerFromWikipedia,thefreeencyclopediaAproportional–integral–derivativecontrollerPIDcontrollerisageneric.controlloopfeedbackmechanismwidelyusedinindustrialcontrolsystems.APIDcontrollerattemptstocorrecttheerrorbetweenameasuredprocessvariableandadesiredsetpointbycalculatingandthenoutputtingacorrectiveactionthatcanadjusttheprocessaccordingly.ThePIDcontrollercalculationalgorithminvolvesthreeseparateparameters;theProportional,theIntegralandDerivativues.TheProportionalvaluedeterminesthereactiontothecurrenterror,theIntegraldeterminesthereactionbasedonthesumofrecenterrorsandtheDerivativedeterminesthereactiontotherateatwhichtheerrorhasbeenchanging.Theweightedsumofthesethreeactionsisusedtoadjusttheprocessviaacontrolelementsuchasthepositionofacontrolvalveorthepowersupplyofaheatingelement.BytuningthethreeconstantsinthePIDcontrolleralgorithmthePIDcanprovidecontrolactiondesignedforspecificprocessrequirements.Theresponseofthecontrollercanbedescribedintermsoftheresponsivenessofthecontrollertoanerror,thedegreetowhichthecontrollerovershootsthesetpointandthedegreeofsystemoscillation.NotethattheuseofthePIDalgorithmforcontroldoesnotguaranteeoptimalcontrolofthesystemorsystemstability.Someapplicationsmayrequireusingonlyoneortwomodestoprovidetheappropriatesystemcontrol.Thisisachievedbysettingthegainofundesiredcontroloutputstozero.APIDcontrollerwillbecalledaPI,PD,PorIcontrollerintheabsenceoftherespectivecontrolactions.PIcontrollersareparticularlycommon,sincederivativeactionisverysensitivetomeasurementnoise,andtheabsenceofanintegralvaluemaypreventthesystemfromreachingitstargetvalueduetothecontrolaction.2AblockdiagramofaPIDcontrollerNoteDuetothediversityofthefieldofcontroltheoryandapplication,manynamingconventionsfortherelevantvariablesareincommonuse.1.ControlloopbasicsAfamiliarexampleofacontrolloopistheactiontakentokeeponesshowerwaterattheidealtemperature,whichtypicallyinvolvesthemixingoftwoprocessstreams,coldandhotwater.Thepersonfeelsthewatertoestimateitstemperature.Basedonthismeasurementtheyperacontrolactionusethecoldwatertaptoadjusttheprocess.Thepersonwouldrepeatthis-outputcontrolloop,adjustingthehotwaterflowuntiltheprocesstemperaturestabilizedatthedesiredvalue.FeelingthewatertemperatureistakingameasurementoftheprocessvalueorprocessvariablePV.ThedesiredtemperatureiscalledthesetpointSP.TheoutputfromthecontrollerandtotheprocessthetappositioniscalledthemanipulatedvariableMV.Thedifferencebetweenthemeasurementandthesetpointistheerrore,toohotortoocoldandbyhowmuch.Asacontroller,onedecidesroughlyhowmuchtochangethetappositionMVafteronedeterminesthetemperaturePV,andthereforetheerror.ThisfirstestimateistheequivalentoftheproportionalactionofaPIDcontroller.TheintegralactionofaPIDcontrollercanbethoughtofasgraduallyadjustingthetemperaturewhenitisalmostright.Derivativeactioncanbethoughtofasnoticingthewatertemperatureisgettinghotterorcolder,andhowfast,andtakingthatintoaccountwhendecidinghowtoadjustthetap.Makingachangethatistoolargewhentheerrorissmallisequivalenttoahighgaincontrollerandwillleadto3overshoot.Ifthecontrollerweretorepeatedlymakechangesthatweretoolargeandrepeatedlyovershootthetarget,thiscontrolloopwouldbetermedunstableandtheoutputwouldoscillatearoundthesetpointineitheraconstant,growing,ordecayingsinusoid.Ahumanwouldnotdothisbecauseweareadaptivecontrollers,learningfromtheprocesshistory,butPIDcontrollersdonothavetheabilitytolearnandmustbesetupcorrectly.Selectingthecorrectgainsforeffectivecontrolisknownastuningthecontroller.IfacontrollerstartsfromastablestateatzeroerrorPVSP,thenfurtherchangesbythecontrollerwillbeinresponsetochangesinothermeasuredorunmeasuredstotheprocessthatimpactontheprocess,andhenceonthePV.VariablesthatimpactontheprocessotherthantheMVareknownasdisturbancesandgenerallycontrollersareusedtorejectdisturbancesand/orimplementsetpointchanges.Changesinfeedwatertemperatureconstituteadisturbancetotheshowerprocess.Intheory,acontrollercanbeusedtocontrolanyprocesswhichhasameasurableoutputPV,aknownidealvalueforthatoutputSPandantotheprocessMVthatwillaffecttherelevantPV.Controllersareusedinindustrytoregulatetemperature,pressure,flowrate,chemicalcomposition,speedandpracticallyeveryothervariableforwhichameasurementexists.Automobilecruisecontrolisanexampleofaprocesswhichutilizesautomatedcontrol.Duetotheirlonghistory,simplicity,wellgroundedtheoryandsimplesetupandmaintenancerequirements,PIDcontrollersarethecontrollersofchoiceanyoftheseapplications.2.PIDcontrollertheoryNoteThissectiondescribestheidealparallelornon-interactingofthePIDcontroller.ForotherspleaseseetheSectionAlternativenotationandPIDs.ThePIDcontrolschemeisnamedafteritsthreecorrectingterms,whosesumconstitutesthemanipulatedvariableMV.Hence4wherePout,Iout,andDoutarethecontributionstotheoutputfromthePIDcontrollerfromeachofthethreeterms,asdefinedbelow.2.1.ProportionaltermTheproportionaltermmakesachangetotheoutputthatisproportionaltothecurrenterrorvalue.TheproportionalresponsecanbeadjustedbymultiplyingtheerrorbyaconstantKp,calledtheproportionalgain.TheproportionaltermisgivenbyWherePoutProportionaloutputKpProportionalGain,atuningparametereErrorSP−PVtTimeorinstantaneoustimethepresentChangeofresponseforvaryingKpAhighproportionalgainresultsinalargechangeintheoutputforagivenchangeintheerror.Iftheproportionalgainistoohigh,thesystemcanbecomeunstableSeethesectiononLoopTuning.Incontrast,asmallgainresultsinasmalloutputresponsetoalargeerror,andalessresponsiveorsensitivecontroller.Iftheproportionalgainistoolow,thecontrolactionmaybetoosmallwhenrespondingtosystemdisturbances.5Intheabsenceofdisturbances,pureproportionalcontrolwillnotsettleatitstargetvalue,butwillretainasteadystateerrorthatisafunctionoftheproportionalgainandtheprocessgain.Despitethesteady-stateoffset,bothtuningtheoryandindustrialpracticeindicatethatitistheproportionaltermthatshouldcontributethebulkoftheoutputchange.2.2.IntegraltermThecontributionfromtheintegraltermisproportionaltoboththemagnitudeoftheerrorandthedurationoftheerror.Summingtheinstantaneouserrorovertimeintegratingtheerrorgivestheaccumulatedoffsetthatshouldhavebeencorrectedpreviously.Theaccumulatederroristhenmultipliedbytheintegralgainandaddedtothecontrolleroutput.Themagnitudeofthecontributionoftheintegraltermtotheoverallcontrolactionisdeterminedbytheintegralgain,Ki.TheintegraltermisgivenbyChangeofresponseforvaryingKiWhereIoutIntegraloutputKiIntegralGain,atuningparametereErrorSP−PVτTimeinthepastcontributingtotheintegralresponseTheintegraltermwhenaddedtotheproportionaltermacceleratesthe6movementoftheprocesstowardssetpointandeliminatestheresidualsteady-stateerrorthatoccurswithaproportionalonlycontroller.However,sincetheintegraltermisrespondingtoaccumulatederrorsfromthepast,itcancausethepresentvaluetoovershootthesetpointvaluecrossoverthesetpointandthencreateadeviationintheotherdirection.Forfurthernotesregardingintegralgaintuningandcontrollerstability,seethesectiononlooptuning.2.3DerivativetermTherateofchangeoftheprocesserroriscalculatedbydeterminingtheslopeoftheerrorovertimei.e.itsfirstderivativewithrespecttotimeandmultiplyingthisrateofchangebythederivativegainKd.Themagnitudeofthecontributionofthederivativetermtotheoverallcontrolactionistermedthederivativegain,Kd.ThederivativetermisgivenbyChangeofresponseforvaryingKdWhereDoutDerivativeoutputKdDerivativeGain,atuningparametereErrorSP−PVtTimeorinstantaneoustimethepresentThederivativetermslowstherateofchangeofthecontrolleroutputandthiseffectismostnoticeableclosetothecontrollersetpoint.Hence,derivativecontrolis7usedtoreducethemagnitudeoftheovershootproducedbytheintegralcomponentandimprovethecombinedcontroller-processstability.However,differentiationofasignalamplifiesnoiseandthusthisterminthecontrollerishighlysensitivetonoiseintheerrorterm,andcancauseaprocesstobecomeunstableifthenoiseandthederivativegainaresufficientlylarge.2.4SummaryTheoutputfromthethreeterms,theproportional,theintegralandthederivativetermsaresummedtocalculatetheoutputofthePIDcontroller.Definingutasthecontrolleroutput,thefinalofthePIDalgorithmisandthetuningparametersareKpProportionalGain-LargerKptypicallymeansfasterresponsesincethelargertheerror,thelargertheProportionaltermcompensation.Anexcessivelylargeproportionalgainwillleadtoprocessinstabilityandoscillation.KiIntegralGain-LargerKiimpliessteadystateerrorsareeliminatedquicker.Thetrade-offislargerovershootanynegativeerrorintegratedduringtransientresponsemustbeintegratedawaybypositiveerrorbeforewereachsteadystate.KdDerivativeGain-LargerKddecreasesovershoot,butslowsdowntransientresponseandmayleadtoinstabilityduetosignalnoiseamplificationinthedifferentiationoftheerror.3.LooptuningIfthePIDcontrollerparametersthegainsoftheproportional,integralandderivativetermsarechosenincorrectly,thecontrolledprocesscanbeunstable,i.e.itsoutputdiverges,withorwithoutoscillation,andislimitedonlybysaturationormechanicalbreakage.Tuningacontrolloopistheadjustmentofitscontrolparametersgain/proportionalband,integralgain/reset,derivativegain/ratetotheoptimumvaluesforthedesiredcontrolresponse.Theoptimumbehavioronaprocesschangeorsetpointchangevariesdependingontheapplication.Someprocessesmustnotallowanovershootoftheprocess8variablebeyondthesetpointif,forexample,thiswouldbeunsafe.Otherprocessesmustminimizetheenergyexpendedinreachinganewsetpoint.Generally,stabilityofresponsethereverseofinstabilityisrequiredandtheprocessmustnotoscillateforanycombinationofprocessconditionsandsetpoints.Someprocesseshaveadegreeofnon-linearityandsoparametersthatworkwellatfull-loadconditionsdontworkwhentheprocessisstartingupfromno-load.Thissectiondescribessometraditionalmanualsforlooptuning.ThereareseveralsfortuningaPIDloop.Themosteffectivesgenerallyinvolvethedevelopmentofsomeofprocessmodel,thenchoosingP,I,andDbasedonthedynamicmodelparameters.Manualtuningscanberelativelyinefficient.Thechoiceofwilldependlargelyonwhetherornottheloopcanbetakenofflinefortuning,andtheresponsetimeofthesystem.Ifthesystemcanbetakenoffline,thebesttuningofteninvolvessubjectingthesystemtoastepchangein,measuringtheoutputasafunctionoftime,andusingthisresponsetodeterminethecontrolparameters.ChoosingaTuningAdvantagesDisadvantagesManualTuningNomathrequired.Online.Requiresexperiencedpersonnel.Ziegler–NicholsProven.Online.Processupset,sometrial-and-error,veryaggressivetuning.SoftwareToolsConsistenttuning.Onlineoroffline.Mayincludveandsensoranalysis.Allowsimulationbeforedownloading.Somecostandtraininginvolved.Cohen-CoonGoodprocessmodels.Somemath.Offline.Onlygoodforfirst-orderprocesses.3.1ManualtuningIfthesystemmustremainonline,onetuningistofirstsettheIandDvaluestozero.IncreasethePuntiltheoutputofthelooposcillates,thenthePshould9beleftsettobeapproximatelyhalfofthatvalueforaquarteramplitudedecaytyperesponse.ThenincreaseDuntilanyoffsetiscorrectinsufficienttimefortheprocess.However,toomuchDwillcauseinstability.Finally,increaseI,ifrequired,untiltheloopisacceptablyquicktoreachitsreferenceafteraloaddisturbance.However,toomuchIwillcauseexcessiveresponseandovershoot.AfastPIDlooptuningusuallyovershootsslightlytoreachthesetpointmorequickly;however,somesystemscannotacceptovershoot,inwhichcaseanover-dampedclosed-loopsystemisrequired,whichwillrequireaPsettingsignificantlylessthanhalfthatofthePsettingcausingoscillation.EffectsofincreasingparametersParameterRiseTimeshootSettlingTimeS.S.ErrorKpDecreaseIncreaseSmallChangeDecreaseKiDecreaseIncreaseIncreaseEliminateKdSmallDecreaseDecreaseDecreaseNone3.2Ziegler–NicholsAnothertuningisallyknownastheZiegler–Nichols,introducedbyJohnG.ZieglerandNathanielB.Nichols.Asintheabove,theIandDgainsarefirstsettozero.ThePgainisincreaseduntilitreachesthecriticalgainKcatwhichtheoutputoftheloopstartstooscillate.KcandtheoscillationperiodPcareusedtosetthegainsasshownZiegler–NicholsControlTypeKpKiKdP0.5Kc--PI0.45Kc1.2Kp/Pc-PID0.6Kc2Kp/PcKpPc/83.3PIDtuningsoftwareMostmodernindustrialfacilitiesnolongertuneloopsusingthemanual10calculationsshownabove.Instead,PIDtuningandloopoptimizationsoftwareareusedtoensureconsistentresults.Thesesoftwarepackageswillgatherthedata,developprocessmodels,andsuggestoptimaltuning.Somesoftwarepackagescanevendeveloptuningbygatheringdatafromreferencechanges.MathematicalPIDlooptuninginducesanimpulseinthesystem,andthenusesthecontrolledsystemsfrequencyresponsetodesignthePIDloopvalues.Inloopswithresponsetimesofseveralminutes,mathematicallooptuningisrecommended,becausetrialanderrorcanliterallytakedaysjusttofindastablesetofloopvalues.Optimalvaluesarehardertofind.Somedigitalloopcontrollersofferaself-tuningfeatureinwhichverysmallsetpointchangesaresenttotheprocess,allowingthecontrolleritselftocalculateoptimaltuningvalues.Otherulasareavailabletotunetheloopaccordingtodifferentperancecriteria.4ModificationstothePIDalgorithmThebasicPIDalgorithmpresentssomechallengesincontrolapplicationsthathavebeenaddressedbyminormodificationstothePID.OnecommonproblemresultingfromtheidealPIDimplementationsisintegralwindup.ThiscanbeaddressedbyInitializingthecontrollerintegraltoadesiredvalueDisablingtheintegralfunctionuntilthePVhasenteredthecontrollableregionLimitingthetimeperiodoverwhichtheintegralerroriscalculatedPreventingtheintegraltermfromaccumulatingaboveorbelowpre-determinedboundsManyPIDloopscontrolamechanicaldeviceforexample,avalve.Mechanicalmaintenancecanbeamajorcostandwearleadstocontroldegradationintheofeitherstictionoradeadbandinthemechanicalresponsetoansignal.Therateofmechanicalwearismainlyafunctionofhowoftenadeviceisactivatedtomakeachange.Wherewearisasignificantconcern,thePIDloopmayhaveanoutputdeadbandtoreducethefrequencyofactivationoftheoutputvalve.Thisisaccomplishedbymodifyingthecontrollertoholditsoutputsteadyifthechange11wouldbesmallwithinthedefineddeadbandrange.Thecalculatedoutputmustleavethedeadbandbeforetheactualoutputwillchange.Theproportionalandderivativetermscanproduceexcessivemovementintheoutputwhenasystemissubjectedtoaninstantaneousstepincreaseintheerror,suchasalargesetpointchange.Inthecaseofthederivativeterm,thisisduetotakingthederivativeoftheerror,whichisverylargeinthecaseofaninstantaneousstepchange.5.LimitationsofPIDcontrolWhilePIDcontrollersareapplicabletomanycontrolproblems,theycanperpoorlyinsomeapplications.PIDcontrollers,whenusedalone,cangivepoorperancewhenthePIDloopgainsmustbereducedsothatthecontrolsystemdoesnotovershoot,oscillateorhuntaboutthecontrolsetpointvalue.Thecontrolsystemperancecanbeimprovedbycombiningthefeedbackorclosed-loopcontrolofaPIDcontrollerwithfeed-forwardoropen-loopcontrol.KnowledgeaboutthesystemsuchasthedesiredaccelerationandinertiacanbefedforwardandcombinedwiththePIDoutputtoimprovetheoverallsystemperance.Thefeed-forwardvaluealonecanoftenprovidethemajorportionofthecontrolleroutput.ThePIDcontrollercanthenbeusedprimarilytorespondtowhateverdifferenceorerrorrema

注意事项

本文(pid控制外文加中文文献中文翻译.doc)为本站会员(a迷迷糊糊的)主动上传,沃文网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。 若此文所含内容侵犯了您的版权或隐私,请立即通知沃文网(发送邮件至2622162128@qq.com或直接QQ联系客服),我们立即给予删除!

温馨提示:如果因为网速或其他原因下载失败请重新下载,重复下载不扣分。




关于我们 - 网站声明 - 网站地图 - 资源地图 - 友情链接 - 网站客服点击这里,给沃文网发消息,QQ:2622162128 - 联系我们

网站客服QQ:2622162128   

沃文网 版权所有  | ©2017-2022

  |陕ICP备19002583号

  陕公网安备 61072602000132号

   违法和不良信息举报:0916-4228922