基于神经网络和粒子群算法的硅钢工艺参数优化

Process parameter optimization of silicon steel based on neural network and particle swarm optimization(PSO)algorithm

结合BP神经网络与粒子群算法,提出了一种降低硅钢铁损的工艺参数优化策略。首先,采用BP神经网络建立了对硅钢铁损的预测模型,模型具有很高的拟合精度和预测精度。然后在工艺参数的优化方面,以BP神经网络预测模型作为适应度函数,选取连续退火RTF炉段的各段炉温作为优化变量,采用粒子群算法优化这些工艺参数。结果显示,基于BP神经网络,采用粒子群算法对部分工艺参数进行优化后,硅钢铁损明显降低,具有一定的指导意义。

A process parameter optimization strategy for reducing silicon steel iron loss was proposed by combining BP neural network and particle swarm optimization(PSO)algorithm.Firstly,a BP neural network was used to establish a prediction model for silicon steel iron loss,which has high fitting and prediction accuracy.Then,in terms of optimizing process parameters,the BP neural network prediction model was used as the fitness function,and the furnace temperatures of each section of the continuous annealing RTF furnace were selected as optimization variables.PSO algorithm was used to optimize these process parameters.The results showed that using BP neural network and PSO algorithm to optimize some process parameters can significantly reduce the iron loss of silicon steel,which has certain guiding significance.