KPCA-LSSVM在磨煤机一次风量预测中的应用.pdfVIP

KPCA．LSSVM在磨煤机一次风量预测中的应用金秀章，等

KPCA．LSSVM在磨煤机一次风量预测中的应用

KPCA—LSSVMinPredictionoftheAirFlowofPulverizer

ofPrimary

Application

舍秀幸韩焘互

(华北电力大学控制与计算机工程学院，河北保定071003)

摘要：针对建立热电厂磨煤机一次风量软测量模型训练样本多、样本特征维数大等特点，考虑到现场测量所需的实时性和准确性，

KPCA．BP模型，这为现场磨煤机一次风量的准确、实时测量提供了一定的理论依据。

关键词：软测量样本优化核主成分分析最小二乘支持向量机BP神经网络

中图分类号：TPl81文献标志码：ADOI：10．16086／j．cnki。issnl000-0380。201503017

withthefeaturesofsoftmodelforofin

accordanceair(PA)flowcogenerationpower

Abstract：Inestablishingsensingprimarypulverizers

withtheofrealtime

andcharacteristics

trainingsamplessamplelargedimension，andconsideringrequirementsperformance

plant，e．g．，more

ontllecombinationofkernel

andinfieldmethodofmodelbasedI

precisionmeasurement．theestablishingsampleoptimizationprincipal

verificationofthemodelis

leastvectorsimulation

component8quremachine(LSSVM)ispmpesed．The

analysis(KPCA)-andsupport

conductedhistoricaldataofcertainresultsindicatethatthemethodbasedonICI)CA-LSSVM

usingoperatingpowerplant；thesampleoptimization

inand

softmodelisbetterthanorKPCA·BPmodel

LSSVM，BPcapabilityoperatingspeed．thiprovides

sensingaccuracy，tracking．8

basisforrealtimeandaccuratemeasurementofP