目的解决火力发电机组的安全运行问题。方法将发电机组的部件故障、传感器故障和执行器故障定义为待辨识的参数,然后利用强跟踪滤波器对机组的状态和参数进行联合估计,最后利用修正的Bayes算法做出故障判决。结果仿真实验表明,该方法不仅能准确快速地识别故障,而且在故障发生时能够保持比较高的状态与参数估计精度。结论火电机组故障实时诊断的强跟踪滤波器方法是切实可行的。
Aim To solve a problem which makes fossil-fired nent's faults, sensor's faults and actuator's faults of fossil-fired power unit run safely. Methods Defining compopower unit as parameters to indentify, estimating the generators'state and parameters utilizing strong track filter, and then getting a judgment for fault according to modified Bayes algorithm, too. Results Simulations experiment indicates the method not only indentifies fault quickly, but also keeps better state and parameter's estimation accuracy when the fault appears. Conclusion The method of real time diagnosis by strong track filter for fossil-fired power unit is feasible.