基于SOM-RBF算法的瓦斯涌出量动态预测模型研究.pdfVIP

第28卷 第8期 传 感 技 术 学 报 Vol_28 NO．8 2015年8月 CHINESEJOURNALOFSENSORSANDACTUATORS Aug．2015 GasEmissionQuantityDynamicPredictionModelofCoalMine BasedonSOM -RBFAlgorithm FUHua ，LIU Ting ，ZHANGShengqiang ，ZHAODonghong’，DINGGuanxi f1．FacultyofElectricalandControlEngineering，LiaoningTechnicalUniversity，HuludaoLiaoning125105，China； 2．LiaoyangPowerSupplyCompanyofLiaoyangElectricalPowerCompanyofStateGrid，LiaoyangLiaoning11100，China) Abstract：A newmodeldynamicpredictionmethodofcombinedartificialneuralnetworkcombiningself-organizing featuremapsandmulti—variableradialbasisfunctionispresented，whichadoptscollectingandprocessingdataby multi-sensorgasmonitoringsystem assamples，asasolutionofthemulti-factorpredictionproblem ofcoalminegas emission．Themodelingandpredictionmethodareutilizedasclusteringfirstly，andthenitisutilizedasclassification tosolvepredictionaccuracyloss，whichiscausedbythenumberlimitationoftrainingsamplesandtheirdispersion． Thepresentedmethodistestedonthehistoricaldatamonitoredinthemine，andsimulationresultsshowthat，thepre— sentedmodelhasahigherpredictionaccuracyandabetterperformanceofgeneralizationwithaveragepredictioner— ror2．16％ incomparisonwithotherpredictionmodels，andthenaveragerelationvarianceis0．0059androot-mean— squareerroris0．1311．Therefore，itcanbeapprovedthatthepresentedmodelrealizesthedynamicpredictionofab— soluteemissionquantityofcoalminegaseffectivelyandhasarelativelyhighpracticality． Keywords：multisensor；gasemission；self-organizingfeaturemaps；RadialBasisFunction；dynamicprediction EEACC： doi：10．3969／j．issn．1004—1699．2015．08．026 基于SOM．RBF算法的瓦斯涌出量动态预测模型研究 付 华 ，刘 汀，张胜强，赵东红 (1．辽宁工程技术大学电气与控制工程学院，辽宁葫芦岛125105； 2．国网辽宁省电力有限公司辽阳供电公司，辽宁辽阳 11100) 摘 要：针对煤矿瓦斯涌出量的多影响因素预测问题，以多传感器的瓦斯监测系统采集处理后的数据作为样本，提出了一种 自组织特征映射神经网络 (Self-organi