CAJ | 학술논문

针对动车组运行环境多变、运行过程具有不确定和强非线性特性，提出了一种动车组运行过程自适应最小二乘支持向量机 LSSVM(Least Squares Support Vector Machine)建模及其速度控制方法。首先基于动车组牵引特性和运行过程数据建立其 LSSVM 模型，并采用粒子群优化算法确定模型参数；其次，依据模型校正策略，通过动车组运行的新数据采用自适应迭代算法来校正 LSSVM 模型参数，以改善模型的适应性；最后给出了基于LSSVM 模型的动车组速度预测控制方法。基于 CRH380AL 型动车组运行数据的对比仿真结果验证了本文方法的有效性。
침대동차조운행배경다변、운행과정구유불학정화강비선성특성，제출료일충동차조운행과정자괄응최소이승지지향량궤 LSSVM(Least Squares Support Vector Machine)건모급기속도공제방법。수선기우동차조견인특성화운행과정수거건립기 LSSVM 모형，병채용입자군우화산법학정모형삼수；기차，의거모형교정책략，통과동차조운행적신수거채용자괄응질대산법래교정 LSSVM 모형삼수，이개선모형적괄응성；최후급출료기우LSSVM 모형적동차조속도예측공제방법。기우 CRH380AL 형동차조운행수거적대비방진결과험증료본문방법적유효성。
In view of the dynamic operating environment,uncertainty and strong nonlinearity of the operation process of electrical multiple units (EMUs),the adaptive modelling and speed control methods of EMU opera-tion process,which are based on least squares support vector machine (LSSVM)algorithm,are proposed. Firstly,an LSSVM model of EMUs is established based on the real-world operation process data and traction characteristics of EMUs,whose parameters are determined by particle swarm optimization algorithm.Next, according to model correction strategy and new operation data of EMUs,these parameters are adjusted with an adaptive iteration algorithm,so as to improve the model’s adaptability.Finally,the predictive speed control method for the EMUs,which is based on the LSSVM model,is presented.Experimental results on the field data of CRH380AL (China’s high-speed EMUs type-380AL)verify the effectiveness of the proposed method.