基于最优熵滤波器的非线性非高斯带有数据缺失随机系统的故障隔离

Entropy Optimization Filtering for Fault Isolation of Nonlinear Non-Gaussian Stochastic Systems with Missing Data

文中提出了一种非线性非高斯带有数据缺失随机系统的故障隔离方法。利用EM算法对缺失数据进行修补,通过构建滤波器对系统的状态估计,将故障隔离问题简化为熵的最优化问题。滤波器的状态误差用非线性非高斯系统方程表示,并且获得状态误差的概率密度函数。通过在只存在目标故障时使状态误差的概率密度函数的熵最大化,而在只有非目标故障时使状态误差的概率密度函数的熵最小化,从而分离出目标故障,实现故障隔离。最后利用仿真示例,与完整数据下的故障隔离效果进行比较,验证了该方法的有效性。

This paper provides an entropy optimization filtering for fault isolation of nonlinear non-gaussian stochastic systems with missing data. First, using the EM algorithm to fix up the missing data; then, constructing a filter to estimate the state of the system, the fault isolation problem is simplified as the entropy optimization problem. State error can be expressed by the equations of non-linear non-Gaussian system, in order to obtain the probability density function of the state error. It needs to maximize the entropy of the probability density function of the state error in the presence of the target fault; While the non-target failure exits, the entropy of the probability density function of the state error is minimized, where the target fault is separated out for fault isolation. Finally, the simulation results are compared to one of the complete data, which demonstrate the effectiveness of the proposed control algorithm.