Abstract: A method of global sliding mode control was put forward for aeroengine with uncertainty and strong nonlinearity based on self-adaptive PSO neural network. A kind of global sliding mode surface nonlinear function was designed, which included a variational exponent function, and the function parameters were adjusted by a adaptive particle swarm optimization(PSO) study arithmetic combined RBF neural network. The global sliding mode control assured control system global robustness. At the same time, the optimization formula about parameters were established based on static state error convergent speed and sliding mode chattering value, and the adaptive PSO neural network was searching current global optimization point. Simulation results show that the devised controller has good effect and weakens chattering. 

Key words: Aeroengine; Global sliding mode control(GSMC); RBF neural network; Particle swarm optimization(PSO)

摘要： 针对现代航空发动机是一个具有不确定性的强非线性系统,提出了一种基于自适应PSO网络整定的航空发动机全程滑模控制方法。设计了一类全程滑模面非线性函数,函数中含有变参数指数函数,其参数由一种新的自适应粒子群学习算法(PSO)结合RBF神经网络来整定。全程滑模控制保证了控制系统的全程鲁棒性,同时,由稳态误差收敛速度和滑模抖振幅度建立参数优化指标,用自适应PSO神经网络快速搜索当前的全局最优点。仿真结果表明,所设计的控制器取得了良好的效果,削弱了抖振。

关键词: 航空发动机;全程滑模控制;RBF神经网络; 粒子群优化算法