文献类型：期刊文献

中文题名：基于双滑模的飞机燃油油量传感器故障监测方法

英文题名：Fault Monitoring Method of Aircraft Fuel Sensor Based on Double Sliding Mode

作者：曲鸣飞[1];张鑫[1];于鑫[2]

第一作者：曲鸣飞

机构：[1]北京电子科技职业学院航空工程学院,北京100176;[2]北京联合大学城市轨道交通与物流学院,北京100011

第一机构：北京电子科技职业学院航空工程学院,北京100176

年份：2024

卷号：37

期号：11

起止页码：1952-1957

中文期刊名：传感技术学报

外文期刊名：Chinese Journal of Sensors and Actuators

收录：CSTPCD;;Scopus;北大核心:【北大核心2023】；CSCD:【CSCD2023_2024】；

基金：北京市教委课题项目(2022Z021-KWY)。

语种：中文

中文关键词：传感器;故障监测;滑膜观测器;李雅普诺夫矩阵;小波包分解法;核主成分分析法

外文关键词：aircraft fuel volume sensor;fault monitoring;synovial observer;Lyapunov matrix;wavelet packet decomposition method;Kernel principal component analysis

摘要：飞机燃油油量传感器故障监测由于输出信号的不稳定性,导致故障正确识别率低、残差监测值与标准残差间误差大,提出基于双滑模的飞机燃油油量传感器故障监测方法。以故障产生原因分析结果为基础,引入双滑膜,利用等效输出误差注入原理建立双滑膜观测器,结合李雅普诺夫矩阵关系优化双滑膜观测器测量矩阵,采集故障信息;通过小波包分解法分解采集的信息,提取特征;引入核主成分分析法,建立标准主成分信息模型,利用采集信息在主成分模型上的投影,对比传感器信息与核主成分信息的偏移,实现飞机燃油油量传感器故障监测。仿真结果表明,所提方法的故障正确识别率为100%,且残差监测值与标准残差间最大仅存在0.02的误差,该方法能够有效监测飞机燃油油量传感器故障。
Due to the instability of the output signal in the fault monitoring of aircraft fuel level sensors,the correct recognition rate of faults is low,and the error between the residual monitoring value and the standard residual is large.A dual sliding mode based fault monitoring method for aircraft fuel level sensors is proposed.Based on the analysis of the causes of faults,a dual sliding film observer is established according to the principle of equivalent output error injection.The measurement matrix of the dual sliding film observer is optimized by combining the Lyapunov matrix relationship to collect fault information.The collected information is decomposed and features are extracted through wavelet packet decomposition method.The kernel principal component analysis method is introduced,a standard principal component information model is established,the projection of collected information on the principal component model is utilized to compare the offset between sensor information and kernel principal component information,and fault monitoring of aircraft fuel level sensors is achieved.The simulation results show that the proposed method has a fault recognition accuracy of 100%,and there is only a maximum error of 0.02 between the residual monitoring value and the standard residual.The proposed method can effectively monitor the fault of the aircraft fuel level sensor.