基于低复杂度最大空闲矩形的非线性传感器故障诊断方法

发布时间：2018-12-16 09:26

【摘要】：针对现存的很多传感器故障诊断方法假设前提多以及复杂度高的问题,提出一种分布式诊断方法来识别无线传感器网络(WSN)中的非线性故障。首先,对局部传感器的输出值进行分析,得到一系列特征值;然后,在交叉误差函数的基础上,将传感器非线性故障诊断等效为最大空闲矩形(LER)问题,并使用提出的低复杂度最大空闲矩形算法予以解决;最后,通过定义一个阈值来诊断有故障的传感器,且不需要使用参考传感器就可以检测一般非线性故障。仿真实验使用了双音谐波信号激励和白噪声信号激励,比较了双线性和指数非线性两种情况下的性能。相比集中式故障诊断方法,提出的算法节省了大量数据传输功率,且获得了非线性模型正常区域边界的准确值;相比最优LER算法,提出的低复杂度LER算法检测性能与之相似,但复杂度更低。
[Abstract]:A distributed fault diagnosis method is proposed to identify nonlinear faults in wireless sensor networks (WSN), aiming at the problems of many existing sensor fault diagnosis methods with multiple assumptions and high complexity. Firstly, the output value of the local sensor is analyzed and a series of eigenvalues are obtained. Then, based on the crossover error function, the sensor nonlinear fault diagnosis is equivalent to the maximum free rectangle (LER) problem, and the proposed algorithm is solved by using the low complexity maximum free rectangle algorithm. Finally, a threshold is defined to diagnose the fault sensor, and the general nonlinear fault can be detected without using the reference sensor. In the simulation experiment, the performance of bilinear and exponential nonlinearity is compared by using double tone harmonic signal excitation and white noise signal excitation. Compared with the centralized fault diagnosis method, the proposed algorithm saves a lot of data transmission power, and obtains the accurate value of the normal region boundary of the nonlinear model. Compared with the optimal LER algorithm, the detection performance of the proposed low complexity LER algorithm is similar to that of the optimal LER algorithm, but the complexity is lower.
【作者单位】：东莞理工学院城市学院计算机与信息科学系;广东工业大学实验教学部;
【基金】：国家自然科学基金资助项目(51305084) 广东省自然科学基金资助项目(9151170003000013) 东莞理工学院城市学院青年教师发展基金资助项目(2016QJZ0032)
【分类号】：TP212

【相似文献】