基于人工神经网络的充填浮选柱捕集区气含率特性研究

Study on gas holdup of collection zone in packed flotation column based on Artificial Neural Network

论述了用人工神经网络的方法对影响充填浮选柱捕集区气含率工艺参数(捕集区高径比、气体加入量和起泡剂用量)进行模拟、训练和预测的过程。结合实验结果讨论了各个影响因素对气含率的作用规律,并用基于实验数据建立的人工神经网络对气含率进行预测分析。所建立的神经网络模型的输出值与实验测量值误差在±5%以内,显示神经网络方法预测分析气含率具有较高的适用性,为充填浮选柱的放大研究提供可行的方法。

In order to further understand gas holdup in collection zone in a packed flotation column,a series of experiments were carried out to examine the effects of flotation parameters on gas holdup.In addition,a mathematic modeling method based on Artificial Neural Network（ANN） was introduced to simulate and then prognose experimental data.The errors between the outputs of neural network and experimental data concentrated in the area of±5%.The results showed that the Artificial Neural Network is applicable in gas holdup analysis of a packed flotation column,and provides a feasible way for scale-up of the packed flotation column.