航空发动机自适应神经网络PID控制

Adaptive Neural Network PID Control for a Aeroengine

普通PID控制器以其简单、实用、易于实现,在经典控制中倍受青睐。对于像航空发动机这样复杂的非线性系统,基于对象精确数学模型的PID控制方法的自适应性较差,难以适应具有非线性、时变不确定性的被控对象。神经网络的建立为这种问题的解决奠定了基础。文中针对航空发动机难于建立精确数学模型的特点,采用了航空发动机自适应神经网络PID参数控制方案,仿真结果表明自适应神经网络PID控制不仅不依赖于精确的对象模型,而且具有满意的动、静态性能。

Due to its simplicity, utility and easy application, PID controller is used largely in classical automation systems. Complicated nonlinear systems such as aircraft engines. PID control method which is based on precise mathematical model has poor adaptability and is not adaptive to nonlinear and time-variant plants. This problem can be solved on the basis of neural network theory. It is very difficult to formulate accurate models for aircraft engines. In consideration of this difficulty, a parameters self-adapting neural network PID control is adopted for aircraft engines. The simulation results indicate that adaptive neural network PID control is not rely on the precisely accurate models but also possesses favorable dynamic and static performances.