论文部分内容阅读
针对连续铸钢结晶器液面的控制是一个非线性和动态的过程,提出了一种自适应稳态粒子群优化(PSO)算法,搜索最优的PID控制器的参数,PID控制器的参数被看成粒子的矢量,其平方误差的总和被定义为目标函数。为了促进种群的多样性和收敛速度,采用自适应变异操作和利用全局最优位置的稳态更新来及时获取进化过程的反馈信息。把适应性突变和全局最优位置的稳态更新添加到标准PSO中,以确保搜索平衡。基于PSO-PID自适应无模型钢的连续铸造中,自适应稳态的PSO和标准的PSO算法相比,有3个创新点。首先,利用全局最优位置的稳态更新来及时获取进化过程的反馈信息;其次,当突变条件具备时,粒子产生突变,种群中产生多样性粒子,在进化中对种群搜索有促进作用:最后,惯性权重系数w从0.9到0.4,自适应地减少,不是一个固定的常数,能保持探索和挖掘的平衡。仿真结果证明,新的PSO算法收敛速度比标准PSO算法的PID控制器速度快,可以准确地控制模具液面的水平:在自适应稳态粒子群优化算法的基础上,提出的增量式PID控制器能有效地调节结晶器液面,具有较好的鲁棒性。改进的PSO-PID的自适应稳态模型在连续铸钢的应用中取得了很好的效果。
The control of continuous casting mold liquid level is a non-linear and dynamic process. An adaptive steady-state particle swarm optimization (PSO) algorithm is proposed to search the optimal parameters of PID controller and PID controller The vectors considered as particles, whose sum of squared errors are defined as the objective function. In order to promote the diversity and convergence speed of population, the adaptive mutation operation and the steady-state update using the global optimal position are used to obtain the feedback information of the evolution process in time. Add steady-state updates of adaptive mutations and global optimal locations to the standard PSO to ensure search is balanced. In continuous casting based on PSO-PID adaptive modelless steel, there are three innovations in adaptive steady state PSO compared with the standard PSO algorithm. Firstly, the steady-state update of the global optimal position is used to obtain the feedback information of the evolutionary process in time. Secondly, when the mutation conditions are available, the particles are mutated and the diversity particles are generated in the population, which can promote the population search in evolution. Finally , Inertia weight coefficient w from 0.9 to 0.4, adaptively reduced, not a fixed constant, to maintain the balance of exploration and mining. The simulation results show that the new PSO algorithm converges faster than the PID controller of the standard PSO algorithm and can accurately control the level of the mold surface. Based on the adaptive steady-state particle swarm optimization algorithm, the incremental PID The controller can effectively adjust the mold liquid level and has good robustness. The improved PSO-PID adaptive steady-state model has achieved good results in the application of continuous casting.