节点文献

基于SVM的电梯制动器静态制动力矩估算方法研究

Estimation of static braking torque of elevator brake based on SVM

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 应征王学斌郭吉丰

【Author】 YING Zheng;WANG Xue-bin;GUO Ji-feng;National Elevator Product Quality Supervision and Inspection Center,Zhejiang Provincial Special Equipment Inspection and Research Institute;College of Electrical Engineering,Zhejiang University;

【机构】 浙江省特种设备检验研究院国家电梯产品质量监督检验中心浙江大学电气工程学院

【摘要】 针对电梯制动器静态制动力矩估算问题,对电梯制动器的动态制动性能与静态制动性能分别进行了研究,利用电梯制动器试验台测试制动器的制动性能,根据实验测试结果提出了一种基于支持向量机的电梯制动器静态制动力矩估算方法,将利用制动器动态制动性能实验得到的制动初始速度与制动过程的平均制动力矩作为估算算法训练数据的输入值,将相同制动器对应的静态制动力矩作为估算算法训练数据的输出值,经过支持向量机算法训练得到估算模型,采用网格搜索法进一步优化估算模型。最后,通过制动实验进一步采集制动性能数据作为验证数据。实验结果显示,电梯制动器的动态制动性能与制动初始速度有关。交叉验证的结果表明,基于SVM的电梯制动器静态制动力矩估算方法能够便捷准确地根据电梯制动器的动态数据估算出静态制动性能。

【Abstract】 Aiming at estimation of the static braking torques of the elevator brakes,the dynamic and static braking performance of the elevator brakes were studied respectively. Elevator brakes were tested by an elevator brake testing system for obtaining the dynamic and static braking torques of the elevator brakes. An estimation method of static braking torques of the elevator brakes based on support vector machines was proposed according to the experiment result. Mean braking torques and initial braking velocity in dynamic braking process were defined as training set input data; meanwhile static braking toques were defined as training set target data. The multivariate regression analysis based on SVM was employed to construct estimation model of static braking torque of elevator brake,the parameters of estimation model were optimized by using gird-search algorithm. At last,the results indicate that the initial braking velocity is one of factors affect the dynamic braking performance,and the presented method is achieved easily and high accurate.

【基金】 国家自然科学基金资助项目(51677171);浙江省质监科研计划资助项目(20150221)
  • 【文献出处】 机电工程 ,Journal of Mechanical & Electrical Engineering , 编辑部邮箱 ,2017年05期
  • 【分类号】TU857
  • 【被引频次】7
  • 【下载频次】139
节点文献中: 

本文链接的文献网络图示:

本文的引文网络