神经网络鲁棒控制及其在燃气轮机排气温度控制中的仿真应用

设计了一种基于BP算法的神经网络鲁棒控制器。利用多层神经网络对任意函数的逼近能力和自学习功能，对运行中的PID调节系数进行在线调整，使整个系统具有良好的自适应能力和鲁棒性。燃气轮机排气温度调节仿真实验表明，这种控制器能够有效地克服传统PID调节器对经验或系统数学模型准确程度的依赖性。

Neural Network Robustness Control and It's Emulation in The Exhaust Temperature Control Gas Turbine

This paper designs a kind of neural network robustness controller based on BP arithmetic. It uses the functions of approach to any function and self study of the multilayer neural network to regulate PID parameters on line, which leads the whole control system to a good performace of adaptivity and robustness. The emulation experiment of the exhaust temperature control of the gas turbine shows that this kind of controller can effectively overcome the dependence of the traditional PID controller on the experience or the precision degree of the mathematical model of the system. Figs 9 and refs 3.