神经网络PI控制算法在燃气发生器上的应用  

Application of Neural Networks PI Control Algorithm in Gas Generator

作　　者：何坤[1] 陈雄[1] 余业辉 严登超[1] 

机构地区：[1]南京理工大学机械工程学院,南京210094

出　　处：《兵工自动化》2017年第6期42-45,共4页Ordnance Industry Automation

摘　　要：为解决燃气发生器喉部时变性使得控制系统对燃气发生器压力控制较为困难的问题,采用神经网络PI算法、模糊PI控制算法和PI算法进行压力控制仿真对比的方法,研究神经网络PI控制算法对容积时变的燃气发生器压强控制的有效性。结果表明:当燃气发生器自由容积变化15倍时,模糊PI算法响应速度是PI算法的8.0倍,且神经网络PI算法控制的压力超调量为0;3种算法中,神经网络PI算法响应速度最快,压力产生的超调量最小,可以较好地解决自由容积时变性带来的控制超调和响应慢的问题。In order to solve the problem of time-varying of gas generator throat which makes control system is more difficult to control gas generator＇s pressure. The control simulation results of neural networks PI algorithm, fuzzy PI algorithm and PI algorithm are compared to study the effectiveness of neural networks PI algorithm when controlling the pressure of time-varying free volume gas generator. The result shows： When the free volume changes 15 times, the response speed of neural networks PI algorithm is 8.0 times faster than PI algorithm, and the overshoot is 0. It has the fastest response speed and smallest overshoot when it takes the neural networks PI algorithm as its pressure control algorithm instead of traditional PI algorithm or fuzzy integral algorithm, so the problems of pressure overshoot and slow response speed caused by the time-varving free volume can be solved hv fuzzy PI algorithm.

关 键 词：燃气发生器 自由容积 神经网络PI算法 模糊PI算法 

分 类 号：TJ711.03[兵器科学与技术—武器系统与运用工程]