基于概率鲁棒的水轮机调速系统鲁棒性评估

针对常规鲁棒性评估方法不能定量给出评估结果的问题,提出了一种基于概率鲁棒方法的控制律鲁棒性定量评估方法。首先确定水轮机调速系统的数学模型和控制策略,并确定参数的摄动范围、置信水平等;然后对系统进行蒙特卡洛仿真,得到阶跃响应曲线簇并绘制性能指标的累计频率曲线;最后计算性能指标的可接受概率、样本均值等统计量,实现鲁棒性的定量评估。以紧水滩水力发电厂水轮机调速系统为研究对象,分别设计了基于PID和模糊内模的调速系统控制器,并用上述方法对两种控制器的鲁棒性进行评估。结果表明,模糊内模控制器的鲁棒性更优,与概率鲁棒的鲁棒性评估结果一致,验证了概率鲁棒对控制律鲁棒性定量评估的有效性。

Robustness Estimation of Turbine Governor System Based on Probabilistic Robustness

Aiming at the problem of not being able to give quantitative evaluation results for conventional robustness assessment methods, a robust quantitative evaluation method of control law based on probabilistic robust method was proposed. Firstly it determined the mathematical model and control strategy of the turbine governing system, and determined the perturbation range and confidence level of the parameters; then Monte Carlo simulation on the system to obtain the step response curve cluster was performed and the cumulative frequency curves of the performance index were plotted; Lastly it calculated the acceptable probability of the performance indicator, the sample mean and other statistics were to achieve a quantitative assessment of robustness. In this paper, the hydraulic turbine governing system of Jinshuitan Hydropower Plant was taken as the research object, and the PID and fuzzy internal model controller were designed respectively. The robustness of the two controllers was evaluated by the above method. The simulation results show that the robustness of the fuzzy internal model controller is relatively better, and it is consistent with the robustness evaluation based on the probabilistic robust method. The effectiveness of the probabilistic robust method for the robust evaluation of the control law robustness is verified.