基于遗传算法的PID控制器参数优化方法研究

PID Controller Parameters Optimization Method Based on Genetic Algorithm

提出了应用遗传算法对PID控制器的比例系数、积分系数寻找最优解的方法。基于遗传算法可直接对结构对象进行操作、对搜索空间中的多个解进行评估,减少了陷入局部最优解的风险,且算法本身易于实现。根据其适应度函数不仅不受连续可微的约束,而且其定义域可以任意设定等特点设计了基于工程PID遗传算法的编码以及基于该编码的交叉、变异算子。通过对电厂中过热汽温控制系统内回路一级过热器的辨识模型和300 MW循环流化床机组中给煤量与床温的关系中左裤衩腿密相区间的100%负荷工况点的传递函数模型进行仿真表明该种算法的有效性。

The proposed genetic algorithm was applied to the PID controller with the ratio coefficients Kp and the integral coefficients Ki looking for the optimal solution method. Based on genetic algorithm to structure the object can be directly to the search space of operation, many solution evaluation, reduce the getting into the local optimal solution of risk, and easy to realize . Its algorithm not only don＇t fitness of function by continuously differentiable constraints, but also can set arbitrary its domain. Considering these characteristics which based on the engineering designed of genetic algorithm, and the PID coding based on the code of crossover and mutation operator. This arti- cle useing the power plant superheated steam temperature control system in the superheater 1 loop identification model and the relationship between 300 MW circulating fluidized bed＇s coal quantity and the bed temperature about the left trousers leg＇s the interval of density operation points 100% load transfer function model simulation shows that the algorithm is effective.