基于模糊神经网络的周期性结垢预测方法研究

针对间歇换热设备的周期性结垢现象,以及由此引起的对象特性时变并造成常规控制器动态性能下降等问题,提出了一种基于模糊神经网络(FNN)的周期性结垢预测方法.把间歇换热设备的周期性结垢分解为可逆垢和不可逆垢,通过两个多入单出四层模糊神经网络分别学习结垢的短周期可逆垢增长趋势和长周期不可逆垢增长趋势,并由两者的组合得到更为精确的污垢热阻预测值.实验结果表明,使用该方法对于一类间歇式麦汁蒸发器污垢热阻的预测精度较经验估计式明显提高,利用该预测方法所构造的时变增益补偿因子在麦汁蒸发器蒸发强度控制系统中亦获得了成功应用.

Study of periodic fouling prediction method based on fuzzy neural network

For batch heat exchangers, washing and CIP (cleaning in place) processes cause periodic fouling and other undesirable changes. A novel FNN-based prediction method was developed for periodic fouling measurements. Reversible fouling and irreversible fouling comprise periodical fouling. Two MISO (multi input single output) four-layer fuzzy neural networks were trained to learn the short-term reversible fouling growth trend and long-term irreversible fouling growth trend. The combination of two network outputs provides the prediction for overall fouling. The results of experiments show that the method, compared to the experimental formula, has better performance in prediction wort evaporator fouling. Using the predicted fouling values, a compensation factor for time-varying gain was applied successfully in the evaporation intensity control system.