论文部分内容阅读
针对一类非线性动态系统,在测量数据发生传输延迟或丢失的情况下,研究了动态系统缓变故障的预测问题.重新定义了粒子滤波器的似然函数,提出了滑动窗口粒子滤波(sliding-window particle filter)算法,并得到了故障幅值的初始估计.在通过滑动窗口在线小波去噪技术对故障幅值的初始估计降噪处理后,提出了基于ARIMA模型的时间序列预测算法.上述算法能够实时地对故障幅值进行迭代估计和预测.在给定故障阈值的条件下,算法能够提前预测系统发生失效的时间.三容水箱的仿真例子说明了算法的有效性.
In the case of a kind of nonlinear dynamic system, when the measured data is delayed or lost, the problem of predicting the gradual change of the dynamic system is studied. The likelihood function of the particle filter is redefined and the sliding window particle filter -window particle filter algorithm is proposed and an initial estimate of the fault amplitude is obtained.After the initial estimation of the fault amplitude is denoised by the on-line wavelet denoising technique of sliding window, a time series prediction algorithm based on ARIMA model is proposed The algorithm can iteratively estimate and predict the fault amplitude in real time.At the given fault threshold, the algorithm can predict the system failure time in advance.The simulation example of the three-tank shows the effectiveness of the algorithm.