基于人群搜索算法的自抗扰控制器参数优化

Parameter optimization of active disturbances rejection controller based on seeker optimization algorithm

火电机组过热蒸汽温度控制系统具有大迟延、大惯性的特性,在负荷变化或内外扰动的影响下,存在参数不易整定的问题。本文将自抗扰控制(ADRC)引入过热蒸汽温度控制系统,并基于人群搜索算法(SOA)对自抗扰控制器的主要参数进行寻优。以自抗扰控制器的6个主要参数为寻优目标,目标函数采用时间乘绝对误差积分(ITAE)准则,同时加入控制器输出量平方项的时间积分,以防止控制器输出量调节幅度过大,以及适当调整误差与控制量的权值。采用测试函数对本文SOA进行验证,并对某300 MW机组过热蒸汽温度控制系统分别采用SOA优化的ADRC-PI控制器和PI-PI控制器以及依据经验公式整定的PI-PI控制器进行时域性能、内外扰动及降负荷仿真对比。结果表明:本文SOA收敛精度和算法鲁棒性均较优;SOA优化的ADRC-PI控制器在抗内外扰动及降负荷过程控制效果均明显优于SOA优化的PI-PI及经验公式整定的PI-PI控制器;SOA优化的控制参数使系统的响应时间更快,超调更小,调节时间更短。

The superheated steam temperature control system of thermal power units has characteristics of large delay and large inertia.Affected by load changes and internal/external disturbances,the parameters are not easy to set.In this study,the active disturbances rejection control(ADRC)is introduced into the superheated steam temperature control system,and a seeker optimization algorithm(SOA)is used to optimize the main parameters of the ADRC.Using the six main parameters of the ADRC as the optimization aims,the objective function adopts the integrated time and absolute error criterion(ITAE).In order to prevent the output of the controller from over adjustment,the time integral of the square of controller output is added,and the weights of the error and the control are adjusted appropriately.The test function is used to verify the SOA,and the time-domain performance,internal and external disturbances and load reduction simulations of a superheated steam temperature control system of a 300 MW unit are compared using the SOA-optimized ADRC-PI controller and SOA-optimized PI-PI controllers and PI-PI controller based on empirical formulas.The results show that the convergence accuracy and the robustness of the SOA are excellent.Besides,in anti-internal and external disturbance and load reduction control effect,the SOA-optimized ADRC-PI controller is obviously better than the SOA-optimized PI-PI controller and PI-PI controller based on empirical formulas.The controlling parameters optimized by the SOA make the system’s response faster,overshoot small,and adjustment time shorter.