用人工神经元网络对蔗糖结晶过程中的电导率建模

Modeling for Conductivity of Sucrose Crystallization in Artificial Neural Networks

针对反映煮糖结晶过程的过饱和度难以检测并难以用常规方法建模的问题,本文采用ANN模型对25锅糖膏的电导率曲线进行了拟合,并在实际生产中对其中的8条曲线(4锅)进行了验证,相关的经济技术指标表明它们的控制效果总体不错.

Supersaturation can reflect sucrose boiling and crystallization directly, but it is difficult to measure and to model in conventional method. In this paper, a few pan＇s massecuite conductivity are fitted in ANN （artificial neural network）, and the effect of eight curves （ four pans） are verified in actual production process. It is shown by relative economic technical indexes that the general control effect are fine.