最优PI控制器参数整定及评价

传统的PI(proportional-integral)控制器整定公式不能灵活地适应具体控制系统特定的性能要求。该文提出一种面向性能指标的最优PI控制器参数整定和鲁棒性评价的方法。首先基于广义Hermite-Biehler定理,计算保证闭环系统稳定的PI控制器参数范围;然后利用具有全局寻优能力的遗传算法,按照指定的性能指标在该范围内进行参数优化。根据Monte-Carlo随机试验原理,对模型参数在一定范围内摄动时的鲁棒稳定性和性能鲁棒性进行了定量的比较和评价。仿真研究表明,方法在指定的性能指标上取得了满意的控制效果,在模型发生摄动时与其他4种方法相比较,具有较好的性能鲁棒性。

Parameter tuning and evaluation for optimal PI controllers

Traditional PI controller tuning formulae can not be easily adapted to specific system requirements. This paper presents a parameter tuning and robustness evaluation method for performance-oriented optimal PI controllers. The generalized Hermite-Biehler theorem was used to define the PI controller parameter region which stabilizes a closed-loop system. The genetic algorithm with global searching was used to select optimum parameters from a region based on specified performance criteria. The Monte-Carlo method was used to compare the stability and performance for parameter uncertainties within a bounded range. Simulation results show that the proposed method demonstrates good stability for the specified indices and more robust performance than four other traditional methods for perturbations of the system parameters.