陶立权1,2,刘程1,张正2,王倩1,王伟1,.基于观测器的传感器故障检测方法对比分析[J].航空发动机,2020,46(3):46-53
基于观测器的传感器故障检测方法对比分析
Comparison and Analysis of Sensor Fault Detection Methods Based on Observer
  
DOI:
中文关键词:  故障检测  性能指标  传感器  卡尔曼滤波器  滑模观测器  软硬故障  航空发动机
英文关键词:fault detection  performance index  sensor  Kalman filter  sliding-mode observer  soft and hard faults  aeroengine
基金项目:
作者单位E-mail
陶立权1,2,刘程1,张正2,王倩1,王伟1, 中国民航大学适航学院1民用航空器适航审定技术与管理研究中心2:天津300300 cauctao@sina.com 
摘要点击次数: 1686
全文下载次数: 1242
中文摘要:
      针对航空发动机传感器故障检测方法的适用范围缺乏量化数据问题,提出以故障检测性能指标为依据,对基于卡尔曼 滤波器和基于自适应滑模观测器的故障检测方法适用性进行对比分析。分别设计了基于卡尔曼滤波器的残差生成和自适应滑模观 测器的故障估计检测方法,根据故障检测策略实现故障检测;以蒙特卡罗法生成随机故障,对比分析在故障和干扰同时存在的情况 下2 种方法的故障检测效果和适用范围;基于DGEN 380 发动机进行仿真,仿真结果表明:基于卡尔曼滤波器的故障检测方法对高 斯噪声的处理能力较强,但故障检测率不足50%,尤其是针对软故障效果较差;而基于滑模观测器的故障检测方法在不考虑测量噪 声的情况下,针对软硬故障检测的效果均较好,但处理传感器测量噪声的能力较差。
英文摘要:
      In view of the lack of quantitative data on the application scope of aeroengine sensor fault detection methods,a comparison and analysis of the applicability of the fault detection methods based on Kalman filter and adaptive sliding-mode observer was proposed based on the performance index of fault detection. Kalman filter was designed to generate residual error and adaptive sliding-mode observer respectively to obtain fault estimation. Fault detection was realized based on fault detection strategy. Random faults were generated by Monte Carlo method,and the fault detection effect and application range of the two methods were compared and analyzed in the presence of both faults and disturbances. Simulation was carried out based on engine of DGEN 380. The simulation results show that the fault detection method based on Kalman filter has a strong ability to deal with Gaussian noise,but the fault detection rate is less than 50%,especially for soft faults. The fault detection method based on sliding-mode observer has better effect on soft and hard fault detection without considering measurement noise,but the ability to deal with sensor measurement noise is poor.
查看全文  查看/发表评论  下载PDF阅读器