SBR工艺中pH值变化时间序列的BP网络预测模型邵 青( 430072) 利用改进的BP算法结合MATLAB工具箱,对SBR工艺中pH值变化的时间序列建立了一种BP网络预测模型,并利用该模型对SBR工艺中pH值的变化规律及趋势进行了研究。结果表明:模型的计算值与实测值之间的误差很小,对未来时刻数据的预测精度也较高,模型较好地反映了SBR工艺中pH值变化的规律,这说明BP网络用于pH值变化规律的研究是一种行之有效的新途径。 SBR工艺 pH值 BP算法 预测精度Back_PropagationNetworkPredictingModelofTimeSeriesoftheVarietyofpHintheSequencingBatchReactor(SBR)ShaoQing(SchoolofCivilandArchitectureEngineering,WuhanUniversity430072)Abstract:Anback_propagationnetworkpredictingmodelwasestablishedbasedonthetimeseriesofthevarietyofpHintheSequencingBatchReac-tor(SBR),usingbothimprovedback_propagationalgorithmandMATLAB.ThemodelwasthenusedinthestudyonthechangingruleofpHandthedevelopingtrendintheSequencingBatchReactor.Theresultsshowedhighaccuracybothforthepresentdataandforthepredictingdata,whichshowedthattheestablishedANNmodelhadreflectedtheruleofthevarietyofpHintheSequencingBatchReactor.AlloftheaboveshowedthatitisaeffectiveandnewwaytodosomeresearchontheregulationofthevarietyofpHusingArtificialNeuralNetwork.Keywords:SequencingBatchReactor(SBR) pH Back_Propagation PredictingAccuracy SBR,:()(、)(,)()()。,pH,,,pH,OH-,。,OH-,,,。,pH,pH,,。(ArtificialNeuralNet-work,ANN),,,,。SBRpH,,,、、、。,,,,。ANNSBRpH。1 1.1 时间序列建模方法X(i),i=1,2,...,N,N。:X(t)=[X(t-1),...,X(t-p)]:(.);p。(.)。:,,3BP。,,(.)。X(i),i=1,2,...,N,40中国农村水利水电·2002年第8期N-p: X(1),…,X(p-1),X(p) X(p+1) N-p X(N-p),…;X(N-2),X(N-1) X(N)N-pBP,、,BP。1.2 所采用的BP网络算法,BP,,BP。1.3 用MATLAB实现的优越性pH,BP,,,,,,。,MATLAB。,MATLAB,:S、、。,。,,。2 2.1 样本选取及样本集的组织SBRpH,1。p=3,2。1 pH(h)00.51.01.52.02.53.0pH7.247.017.006.866.856.966.98(h)3.54.04.55.57.09.2pH6.936.966.936.927.007.142 pH12317.247.017.006.8627.017.006.866.8537.006.866.856.9646.866.856.966.9856.856.966.986.9366.966.986.936.9676.986.936.966.9386.936.966.936.929()6.966.936.927.0010()6.936.927.007.14 1~8,9~10。2.2 BP人工神经网络结构与参数3,3-8-1。3、81。,(1)。,Sigmoid,:f(x)=1/(1+e-x) (1)。Sigmoid-1~+1,,,-1~+1,。:V=A×lnV0+B(V0,V,A、B)。(2)。,,。,nS,。,(-1,1)。。lr=0.02,lr_inc=1.02;lr_dec=0.5;mom_const=0.98;err_ratio=1.03;err_goal=0.002;max_epoch=8000;3 3.1 样本训练结果150,,。,3,4。3 1231-0.24060.6419-1.2207-0.90532-0.7026-1.19680.18460.32693-0.87110.13041.08820.05154-0.60360.07941.2607-0.37945-1.05260.88860.37610.79626-0.4301-0.3741-0.97061.145570.37611.28750.9362-0.118180.79620.9192-0.50850.26874 12340.5636-0.93280.71400.365256780.60370.8354-0.2238-0.0560 BP,SBRpH。3.2 模型精度检验及预测,21~8,5。41SBR工艺中pH值变化时间序列的BP网络预测模型 邵 青石台寺泵站水锤计算及节制回流阀在水锤防护中的应用蒋 劲 李继珊( 430072) 李国栋( 441200) 对枣阳石台寺泵站4号水泵装置停泵水力过渡过程进行分析和计算,并提出防止水锤破坏的工程措施。针对该装置的具体条件,采用了专利产品节制回流阀消除水锤,并通过现场测试。测试结果与计算结果相符,证实采用节制回流水锤消除阀能解决该站4号水泵装置问题。 泵站 水锤 节制回流阀1 ,1995,、、。4,3。,,4,,,,,4,,,,,1999827。,4,、、、,,4,,。:24sh-19;QR=0.95m3/s;HR=27m;n=970r/min;Y-400-54-6;N=400kW;HD740-10。1070.4m,(0.6MPa)。1。2 ,,。2.1 水锤计算的特征线法5 (mg/L)(mg/L)(%)16.866.86200.0291026.856.8499-0.0014636.966.9543-0.0819046.986.98010.0014356.936.9296-0.0057766.966.96180.0258676.936.93100.0014486.926.92050.007239()7.006.9998-0.0028610()7.147.1390-0.01400 5,。SBRpH,SBRpH。SBRpH。◆1 ,.[J].,1996(11):82~902 ,.[J].,1997,21(6):522~5253 .MATLAB[M].:,19984 ,.[M].:,1992[] ,,33,(:2002-04-26),:42中国农村水利水电·2002年第8期