电力机车自动过分相永磁体逆问题研究

《电力机车自动过分相永磁体逆问题研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、湖南大学硕士学位论文电力机车自动过分相永磁体逆问题研究姓名：周立华申请学位级别：硕士专业：控制工程指导教师：罗隆福;邱煌明20041007Ꮉ⿟⸩຿ᄺԡ䆎᭛ᨬ㽕ᴀ᭛Џ㽕ᰃ᥶㋶⬉࡯ᴎ䔺㞾ࡼ䖛ߚⳌ∌⺕ԧ䗚䯂乬∖㾷ᮍ⊩ˈेᏆⶹ⺕എぎ䯈ߚᏗড䆒䅵⺕ԧ㒧ᵘḋᴀDŽೈ㒩ℸ䯂乬ˈߚϝ໻䚼ߚќҹ㾷އDŽ佪ܜˈҟ㒡њ᳝䰤ܗߚᵤᮍ⊩੠Ѡ㓈᳝䰤ܗߚᵤ䕃ӊFEMMˈᇍḋᴀ⺕ԧ䗤ϔ䖯㸠᳝䰤ܗᓎ῵੠᭄᥂ߚᵤˈ㦋প↣Ͼḋᴀᣛᅮぎ䯈ԡ㕂ⱘ⺕എᔎᑺ᭄᥂ˈ೼ℸ෎⸔Ϟˈ߽⫼MATLAB/SIMULINKᓎゟ䅵ㅫ఼ӓⳳ῵ൟˈߚ߿∖ߎ⺕ԧḋᴀ⺕എᔎᑺϢᅲ⌟⺕എᔎᑺⱘ䇃Ꮒᑇᮍ੠ˈ㦋ᕫ

2、⺕ԧḋᴀᑧDŽ✊ৢˈ䆒⊩ᓎゟҹ⺕ԧḋᴀሎᇌЎ䕧ܹǃҹḋᴀ⺕എᔎᑺϢᅲ⌟⺕എᔎᑺ䇃Ꮒᑇᮍ੠Ў䕧ߎⱘⳂᷛߑ᭄ˈ⬅Ѣ䕧ܹϢ䕧ߎП䯈໡ᴖⱘ䴲㒓ᗻ݇㋏ˈⳂࠡজ᮴⦄៤ⱘ㒳ϔⱘ῵ᓣ᪡԰ˈᴀ᭛䞛⫼݋᳝催ᑺ䴲㒓ᗻ䆚߿㛑࡯ⱘҎᎹ⼲㒣㔥㒰ˈे䗮䖛෎Ѣ䇃Ꮒড৥Ӵ᪁ㅫ⊩ⱘ໮ሖࠡ作㔥㒰˄BP⼲㒣㔥㒰˅䆁㒗ˈ݋ԧ䞛⫼њࡼ䞣ẃᑺϟ䰡ㅫ⊩੠L-MӬ࣪ㅫ⊩ᴹ䆁㒗BP⼲㒣㔥㒰ˈḍ᥂䇃Ꮒ㊒ᑺ㽕∖䗝⫼њL-MӬ࣪ㅫ⊩䆁㒗㒧ᵰˈᇚ䕧ܹⱘḋᴀሎᇌϢ䕧ߎⱘ䇃Ꮒᑇᮍ੠П䯈ᓎゟ䍋᯴ᇘ݇㋏ˈे᠔䳔ⱘⳂᷛߑ᭄DŽ᳔ৢˈ೼ҹ⺕ԧሎᇌЎ䆒䅵ব䞣ǃҹ䇃Ꮒᑇᮍ੠᳔ᇣЎⳂᷛߑ᭄ǃҹሎᇌ䰤ࠊЎ㑺ᴳᴵӊⱘ෎

3、⸔ϞˈᓎゟӬ᭄࣪ᄺ῵ൟˈ⫼䘫Ӵㅫ⊩Ӭ࣪ᮍ⊩ˈ∖ᕫ᳔᥹䖥ᅲ⌟ؐⱘ᳔Ӭ㾷DŽЎњ偠䆕䘫Ӵㅫ⊩ᇍ㾷އᴀ᭛䯂乬ⱘℷ⹂ᗻˈᴀ᭛জ䖤⫼њᢳ⠯乓⊩ᇍ݇㋏῵ൟⱘӬ࣪䅵ㅫˈϸ⾡ㅫ⊩ⱘ㒧ᵰ෎ᴀ䖒៤ϔ㟈DŽᴀ᭛ҟ㒡ⱘ䞛⫼ҎᎹ⼲㒣㔥㒰੠䘫Ӵㅫ⊩Ⳍ㒧ড়ⱘᮍᓣᴹ㾷އ∌⺕ԧ䗚䯂乬ˈ䕙ПҹMATLAB/SIMULINKǃMATLAB⼲㒣㔥㒰Ꮉ݋ㆅ(Nnbox)ǃMATLABӬ࣪Ꮉ݋ㆅ(OptimizationToolbox)ǃMATLAB䘫Ӵㅫ⊩Ꮉ݋ㆅ(GAToolbox)ˈ㛑໳䙓ܡ㓪ݭ໻䞣ⱘ䅵ㅫᴎ⿟ᑣˈҹ᳔ᖿⱘᮍᓣᡒࠄ݊ᇍᑨⱘ݇㋏῵ൟˈ᧰ᇏࠄܼሔ᳔Ӭ㾷DŽ݇䬂䆡˖䘫Ӵㅫ

4、⊩˗⬉࡯ᴎ䔺˗∌⺕ԧ˗Ӭ࣪˗FEMM˗BP⼲㒣㔥㒰˗ᢳ⠯乓⊩IᎹ⿟⸩຿ᄺԡ䆎᭛AbstractInthisthesis,theprincipalproblemthatneedssolutionistodiscoverythebestmethodoftheelectriclocomotive’sexcessivepermanentmagnet,thatistodiscoverythesampleoftheknownmagneticfield’sinverteddesign’sstructureofthemagnet.Basedonthi

5、sproblem,therearethreepartstosolveit.Firstly,themethodofthefiniteelementanalysisandthesoftwareofthetwoorientationfiniteelementanalysishavebeenintroduced.Throughcreatingthefiniteelementmodelandassayingitsdata,themagneticfieldintensitydataofeveryspecimen’sareobtained.Base

6、donthis,usingMATLAB/SIMULINK,thecalculatoremulationmodelabouttheimportofsample’sdimensionandtheoutputofthesquareofthemagneticfieldintensity’serroriscreated.Thespecimen’swarehouseaboutthemagnetismfieldintensityisobtainedbythecalculatoremulationmodel.Then,tryingtosetupata

7、rgetfunctionofinputtingthesizeofthemagnet’ssampleandoutputtingthesummationofthesquareoferrorofthesample’smagneticintensityandtherealmagneticintensity.Becauseofthecomplicatednonlinearrelationshipbetweentheinputandtheoutputandthenon-settledmodel’sprogress,thethesisadoptst

8、hemethodthatArtificialNeuralNetworks(ANN)withthecapabilityofhighnonlinearrecognition.Thatistosay,themodelthatr