基于神經(jīng)網(wǎng)絡(luò)的電熔鎂爐智能控制系統(tǒng)研究

1、東北大學(xué)碩士學(xué)位論文基于神經(jīng)網(wǎng)絡(luò)的電熔鎂爐智能控制系統(tǒng)研究姓名：馬文來申請學(xué)位級別：碩士專業(yè)：電力電子與電力傳動指導(dǎo)教師：佟玉鵬20060201東北大學(xué)碩士學(xué)位論文ResearchonIntelligentControlSystemofElectricMeltingFurnaceforMagnesiabasedonNeuralNetworkAbstractMgOcrystalisakindofsuperfireproofmaterial

2、usedinmanyareas，anditsmeltingpointisabout2800“CThefireproofmaterialmadeofMgOcanresist2000。ChightemperatureThemainwaytoproduceMgOissmeltingprocessintheelectricfumace。ElectricMeltingFurnaceforMagnesia(EMFM)isacomplexplant，

3、whichexitsseriousnon—linearity，timechanging，strongcoupleThesingleelectrodecontrolmethod，whichbasedonclassicalcontroltheoryhasbeenusedwidelyInthisthesistheshortcomingsoftheclassicalsingleelectrodecontrolmethodhasbeenanaly

4、zed，inaccordancewithEMFMcontrolsystem’Scharacteristics，requestofdesignanddevelopingdirectionofEMFMcontrolsystem，anintelligentthreeelectrodescontrolmethodhasbeenproposedTheofflinecontrolmodeloftheEMFMisestablishedbasedonB

5、PneuralnetworkThemainworksofthethesisareasfollows：1Thebasicprinciple，technicalequipment，structureandthesmeltingprocessoftheElectricMeltingFurnaceforMagnesiaareintroduced2ThepresentstatusofElectricMeltingFurnaceforMagnesi

6、acontrolsystemisintroducedandseveralwidelyusedcontrolmethodsofEMFMarecompared3BasedonthereformedBPneuralnetworktheorytheofflineElectricMeltingFurnaceforMagnesiacontrolsystemmodelisbuildThethesisadoptsthesampledatatotrain

7、thenetwork，andthesimulationandresearchiSdoneonthecontrolsystemmodelAccordingtothesimulationresult，feasibilityandvalidityofthecontrolmethodwereproved4AVRsinglechipmicrocontrollerisadoptedasthecoreparttodesigntheEMFM’Sinte

8、lligentcontrollerandperipheralcircuitsofADconvertionserialcommunication，display，driveandcontr015Thesoftwarepartofsystemadoptsmodularstructuremethod，makeuseoftheresultoftheofflinenetworktrainingtocarryouttheneuralnetworkc