熱工故障的模糊神經(jīng)網(wǎng)絡(luò)診斷方法研究.pdf

1、華北電力大學(xué)（保定）碩士學(xué)位論文熱工故障的模糊神經(jīng)網(wǎng)絡(luò)診斷方法研究姓名：馬杏齋申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：控制理論與控制工程指導(dǎo)教師：馬良玉20071215華北電力人學(xué)碩士學(xué)位論文摘要摘要在了解熱工系統(tǒng)故障診斷研究現(xiàn)狀，分析故障診斷包含的主要環(huán)節(jié)的基礎(chǔ)上，深入研究了基于模糊RBF神經(jīng)網(wǎng)絡(luò)的熱工過(guò)程故障診斷方法。文中首先對(duì)故障診斷的參數(shù)預(yù)處理環(huán)節(jié)，征兆的模糊計(jì)算方法進(jìn)行了研究，討論了語(yǔ)義型和趨勢(shì)型兩種常見(jiàn)征兆的模糊計(jì)算和合成應(yīng)用方法。論述了R

2、BF神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu)及訓(xùn)練方法，重點(diǎn)對(duì)本文應(yīng)用的梯度訓(xùn)練法進(jìn)行了詳細(xì)分析。運(yùn)用模塊化方法，采用FORTRAN語(yǔ)言編制了征兆模糊計(jì)算及RBF神經(jīng)網(wǎng)絡(luò)診斷算法。以300Mw火電機(jī)組高加系統(tǒng)為例，對(duì)其故障規(guī)律進(jìn)行了分析，總結(jié)了故障模糊知識(shí)庫(kù)。應(yīng)用本文編制的診斷算法搭建了高加系統(tǒng)故障診斷模型，借助仿真系統(tǒng)進(jìn)行了詳細(xì)的仿真診斷試驗(yàn)，驗(yàn)證了故障診斷方法的有效性。關(guān)鍵詞：熱工系統(tǒng)，故障診斷，征兆模糊計(jì)算，RBF神經(jīng)網(wǎng)絡(luò)，高壓加熱器ABSTRACTInt

3、hispaper，F(xiàn)aultdiagnosisofthermalprocesswasstudieddeeplybasedOnfuzzyRBFneuralnetwork，withknowingthecufrentsituation0fthermalsystemfaultdiagnosis，andanalysingthekeysegmentoffaultdiagnoseFistly，Parameterpretreatmentoffaultd

4、iagnosisandSymptomfuzzycalculationwasstudied，twotypesofsymptom，trendsymptomandsemanticsymptomfuzzycalcuationgmethodsandintegratedcalculationwerediscussedThestructureandtrainingalgorithms0fRBFaftificialneuralnetworkwerere

5、alized，especiallythegradstrainingmethodThesymptomcalculcationandRBFdiagnosealgorithmwerebuiltinFORTRANlanguagebymodularizedmethodHaVetakenthehigh—pressureheaterof300MWthermalpOwerunitasexample，itsfaultcharacteristicrules

6、wereanalysed，anditsfauItfuzzyknowledgebasesweresummarizedThehighpressureheaterfaultdiagnosismodelwasbuiltwithdiagnosealgOrithms，thetestsofsimulatingfaultweredoneundersimulationsupportingsystem，theValidityofdiagnosemethod