基于改进PSO算法的目标温度模糊神经网络控制器

Target temperature fuzzy neural network controller based on improved PSO algorithm

以某钢厂引进的板坯连铸二冷控制为研究对象,针对现有控制系统由于铸坯表面目标温度是预先设定的固定值,存在二冷水量波动大、铸坯质量不稳定等缺陷,设计了基于改进PSO算法的目标温度模糊神经网络控制器,在遵守冶金准则的前提下,根据浇注钢种与拉速、中包温度变化量动态控制目标温度。仿真结果表明:该控制器控制误差小,适应范围广,可以满足生产要求。提出了模糊神经网络的改进PSO算法,阐述了其基本思想、改进之处及其实施过程。研究结果对引进的同类连铸板坯二冷控制系统的升级改造具有指导意义。

Take a introduced secondary cooling dynamic control system in casting slab of some factory as the research object,the system exists flaw that water volume undulates in a big way and slabs quality is not steadily because the target temperature is fixed values in advance.Target temperature fuzzy neural network controller is designed based on the Improved Particle Swarm Optimization （IPSO） algorithm.Consisting with metallurgical criteria,the target temperature is dynamic controlled based on casting grade,variable error values of tundish temperature and casting speed.The simulation result indicates that the controller error is small,the adaption scope is broad and may satisfy the production request.Proposed IPSO algorithm what is suitable for fuzzy nerve network,elaborates basic thought,improvement place and the implementation process.The result has guiding significance to the promotion and reformation of same kind introduced control system.