基于BP神经网络和NSGAⅡ算法的立式搅拌磨机能耗优化

Energy Consumption Optimization of Vertical Stirred Mill Based on BP Neural Network and NSGA Ⅱ Algorithm

针对立式搅拌磨机在磨矿过程中能耗较大的问题,采用人工神经网络的方法建立工艺参数与能耗和出粉率的回归关系,并采用多目标遗传算法作为优化算法,构建立式搅拌磨机工艺的多目标优化策略,最终确定立式搅拌磨机的最优工艺参数,从而实现降低搅拌磨机磨矿能耗和减少生产成本的目的。以矿料粒度、搅拌器转速、介质填充率、矿浆浓度作为变量,设计进行了立式搅拌磨机的磨矿能耗试验。根据试验数据构建了基于BP神经网络的立式搅拌磨机的能耗和出粉率的预测模型,并验证了其预测的准确性;然后基于该预测模型运用NSGAⅡ算法进行了多目标优化,获得了Pareto最优解集;最后分析讨论得到了最优的工艺参数,对比优化前的工艺参数,在出粉率相同的情况下,能耗降低了7%。

Aiming at the problem of large energy consumption of vertical stirred mill in the grinding process,the artificial neural network method was used to establish the regression relationship between the process parameters and the energy consumption and the milling rate,and the multi-objective genetic algorithm was used as the optimization algorithm to construct the multi-objective optimization strategy of the vertical stirred mill process,and finally determine the optimal process parameters of the vertical stirred mill,so as to achieve the purpose of reducing the grinding energy consumption of the stirred mill and reducing the production cost.The grinding energy consumption test of vertical stirred mill was designed and carried out by taking the particle size of mineral aggregate,agitator speed,medium filling rate and pulp concentration as variables.According to the experimental data,the prediction model of energy consumption and milling yield of vertical stirred mill based on BP neural network was established,and the accuracy of the prediction was verified.Then,based on the prediction model,multi-objective optimization is carried out by using NSGA Ⅱ algorithm,and the Pareto optimal solution set was obtained.Finally,the optimal process parameters were obtained by analysis and discussion.Compared with the process parameters before optimization,the energy consumption was reduced by 7%under the same powder yield.