灰色预测自适应内模控制在电厂过热汽温控制中的应用

Gray Prediction Adaptive Internal Model Control,Applied Superheated Steam Temperature in Power Plants

针对电厂过热汽温的大惯性、大滞后和参数时变等特性,提出了基于灰色预测的自适应内模控制方法。使用Adaline神经网络实时辨识对象的增益和时滞,克服参数时变影响。在误差反馈通道上增加灰色预测模块,根据辨识出的对象时滞的大小,动态确定误差反馈灰色预测的长度;通过超前调节,可以有效克服模型失配和外扰的影响。

Adaptive IMC （Internal Model Control）, based on gray prediction, is proposed for dealing with problems, characterized by large inertia, temperature lag and time varying parameters, as exhibited by superheated steam systems of boilers. Adaline neural networks are used to identify on-line object＇s gain and delay, in order to overcome effects caused by variation, with time of relevant parameters. With a gray prediction module, inserted into the error loop, the grey predicted error in feed-back time is dynamically determined according to the identified objects time lag. The influence of the model＇s mismatch, as well as that of external disturbances, can then effectively be overcome by anteceding control actions. System simulations show that this method is featured by good control performance properties.