论文部分内容阅读
针对非线性非高斯系统在实际工作环境中受强噪声干扰影响导致的故障诊断精度低的问题,提出了一种状态转移密度方差自适应更新的代价评估粒子滤波(CRPF)故障诊断方法。通过设计观测值与先验状态之间的相关性判别函数,根据噪声和误差的大小实时自适应调整状态转移密度方差,欲增强算法对强噪声干扰的适应能力;接着研究了残差自适应阈值的设计方法,通过引入滑动窗求区间均值代替基于参数置信区间自适应阈值的均值和方差,以期在保证故障诊断准确性的前提下减小计算时间。以160MW燃油机组为例,通过对不同强噪声环境下的汽包水位传感器故障诊断实例分析,结果表明该方法在复杂噪声环境下故障诊断的准确性得到了明显提高,同时减少了计算时间。
Aiming at the low fault diagnosis accuracy of nonlinear non-Gaussian system caused by the strong noise interference in practical working environment, a state-of-the-art density variance adaptive updating cost estimation particle filter (CRPF) fault diagnosis method is proposed. By designing the correlation discriminant function between observed and prior state, the variance of state transition density can be adaptively adjusted according to the size of noise and error in real time so as to enhance the adaptability of the algorithm to strong noise interference. Then, the residual adaptive threshold , The mean value and variance of the adaptive threshold based on parameter confidence interval are replaced by introducing the mean of sliding window interval to reduce the calculation time under the premise of ensuring the accuracy of fault diagnosis. Taking a 160 MW unit as an example, the fault diagnosis of the drum water level sensor under different strong noises is analyzed. The results show that the accuracy of fault diagnosis in complex noise environment is significantly improved and the calculation time is reduced.