糖厂澄清工段生产指标建模及操作参数优化

Modeling of Key Production Indices and Operating Parameters Optimized Set for Sugar Clarification Process

针对亚法糖厂澄清工段清汁色值和清汁残硫量难以在线测量的问题,提出了一种基于人工蜂群优化的在线极限学习机软测量方法;先用核主元分析法确定影响清汁质量的关键参数,建立基于在线极限学习机的软测量模型;同时利用人工蜂群算法对在线极限学习机的隐层参数进行寻优,优化所建模型;最后,使用带约束的粒子群对软测量模型进行优化求解,得到典型工况下的最优操作设定值,为后续工况操作提供参考依据;仿真结果表明,基于人工蜂群优化的在线极限学习机模型能够准确地预测清汁色值和残硫量,同时基于此模型优化的操作参数设定值能够达到期望的指标。

For the problem that the color value and residual sulfur content of clarified juice in the clarifying process of sugar factory are difficult to get online, put forward a kind of online extreme learning machine soft sensor based on artificial bee colony optimization method. First, using kernel principal component analysis method to determine the key parameters affecting the quality of iuice, establish the soft measurement model based on online extreme learning machine, and using artificial bee colony algorithm optimal initial weights and threshold for the extreme learning machine. Later on, using constrained particle swarm optimization algorithm of typical working conditions in past production process, to take the optimal set point hitting the target of technical indexes in the corresponding working conditions, this provide a reference basis for the subsequent condition of operation. The simulation results show that online extreme learning machine based on artificial swarm optimization model has better precision and optimization of operating parameters set value based on this model can achieve the desired index.