灰色神经网络控制器在液位控制中的应用

Application of Grey Neural Net Controller in Level Control

针对实际工业生产过程中被控对象的时变和纯滞后的特点,本文根据神经网络具有能够充分逼近任意复杂的非线性关系和能够学习严重不确定系统的动态特性,具有适应性、智能性较好的特点并结合灰色预测控制的超前预测的特点,提出了基于BP神经网络的灰色预测控制算法。仿真结果的对比分析表明:本文的控制算法与传统控制器相比,具有控制简便、自适应性较强等特点,适用于对纯滞后和模型时变对象的控制。最后采用xPC技术在实际设备上进行了算法的验证,取得了满意的效果。

Aimed at the controlled object＇s character of changing with time and time delay, we based the NN＇s merit of highly accord with all kinds of non-liner, learning the uncertain system, good adaptability, excellent intelligence and join the characteristic of forecast of the algorithm of grey prediction control , we put forward the algorithm of grey predict control based on the theory of BP NN.The comparing and analysing of simulation result shows：compared with traditional controller, our algorithm become simple and convenient, stronger in adaptability, it works well in resolving the problem of time delay and model of time varying. Finally, we do real-time control verification by proposed algorithm in the method of xPC technology on real device.