电梯能耗的混合预测控制方法  

Study of the Mixed Predictive Control Method to the Elevator Energy Consumption

作　　者：刘剑[1] 侯冉[1] 李晓刚[1] 刘健[2] 

机构地区：[1]沈阳建筑大学信息与控制工程学院,辽宁沈阳110168 [2]沈阳建筑大学理学院,辽宁沈阳110168

出　　处：《沈阳建筑大学学报（自然科学版）》2006年第2期319-322,共4页Journal of Shenyang Jianzhu University：Natural Science

基　　金：国家自然科学基金资助项目(69874026);建设部科技攻关项目(05-K6-20)

摘　　要：目的为实现电梯群控系统的最佳派梯调度及节约能源提供重要的决策依据.方法针对电梯能耗的预测问题,分别讨论了基于ARMA模型的预测算法和基于径向基函数(RBF)神经网络的预测方法.在此基础上,提出了一种将ARMA模型预测与RBF神经网络预测相结合的混合预测方法.新方法综合了两种算法的优点,能较好地满足电梯能耗的预测要求.探讨了新方法在电梯能耗预测中的应用情况,根据电梯实测数据进行了仿真试验,对实际能耗和预测能耗进行了比较和误差分析.结果达到了预测速度较快、预测精度较高的效果,验证了该方法的可行性.结论该混合预测方法应用于电梯能耗的多步预测时,具有较好的预测性能,取得了较好的预测结果.In order to implement the optimal dispatch of elevator group and reduce elevator energy consumption, a new prediction method is presented based on the ARMA model prediction and the RBF neural network model prediction. The method is a mixed prediction method that is composed of the ARMA model prediction and the RBF neural network prediction. It combines the advantages of the two methods so that it could well meet the needs of prediction for elevator energy consumption. And the application of this method to the elevator energy consumption prediction is studied in detail. The real observation data are used for simulation tests. The simulation results show that the mixed prediction method has good prediction performance while it is applied to the elevator energy consumption multi-step prediction. The feasibility of this method has been proved. Compared with the traditional methods, its prediction speed is quicker and the prediction precision is higher.

关 键 词：电梯能耗 ARMA模型 神经网络 混合预测 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]