基于混沌PSONN-PID神经网络的热连轧活套系统自适应解耦控制

为改善粒子群优化算法的寻优性能，提出了一种新的算法———混沌粒子群算法。该算法将混沌搜索机制引入到粒子群算法中来增加粒子的多样性，同时采用增加粒子交互性策略及先增后减的惯性权重因子模型来设置惯性权重因子，改善了递减策略中存在的缺陷。将改进后的算法与PID型单神经元相结合，并将其用于热连轧活套解耦控制系统。仿真试验表明：该算法较好地克服了粒子群算法易早熟和陷入局部最优的缺点，为解决活套系统高度张力耦合问题提供了一种新的有效途径。

Adaptive decoupling control for looper system of hot strip mill based on chaos PSONN- PID neural network

To improve the performance of PSO( particle swarm optimization)optimization algorithm,a new algo-rithm-CPSO(chaotic particle swarm optimization)was proposed. The algorithm chaotic search mechanism wasintroducedtotheparticleswarmalgorithmtoincreasethediversityofparticle.Inordertoimprovethedimin-ishing policy flaws,the algorithm also adopts the methods of increasing particle interaction strategy and the first-increased-then-decreased inertia weight factor model to set inertia weight factor. The improved algorithm and PID single neuron are combined,and which is used in hot rolling looper decoupling control system. Simulation results show that the algorithm can overcome the defects of PSO in prematureness and being easy to fall into local opti-mum. This research puts forward a new and effective way to solve the high tension coupling problem in looper sys-tem.