基于改进PSO神经网络的微米木纤维切削参数优化

Optimization study of cutting parameters of micron wood fiber based on improved PSO neural network

针对使用传统加工技术切削后的工件无法达到精密要求的问题,为了从工艺流程角度提升微米木纤维的加工精度和切削效果,提出了一种基于改进粒子群算法和BP神经网络的优化算法,以实现微米木纤维的精密切削加工。采用误差反向传播算法实现切削参数间复杂关系的最佳结构选择,通过改进的粒子群优化算法(PSO)解决了BP网络自身的局部极小值收敛的缺陷,给出了科学合理的切削参数输出。通过不同树种的微米木纤维切削参数仿真优化实验,验证了算法的准确性、训练精度和有效性。研究表明:提出的改进优化算法可以预测出待加工木材的切削参数,且具有较高的训练精度。

In order to improve the process of micron wood fiber cutting,an improved particle swarm algorithm and BP neural network based on the combination of optimization algorithm is proposed to achieve the precision machining of micron wood fiber.The error back propagation algorithm is used to achieve the best structure selection of the complex relationship between cutting parameters.The improved particle swarm optimization algorithm(PSO)solves the defect of local minimum convergence of BP network,and gives a scientific and reasonable output of cutting parameters.The precision and effectiveness of the precision training of the algorithm are verified by the simulation and optimization experiments of the cutting parameters of different tree species.The research shows that the improved optimization algorithm proposed in this paper can predict the cutting parameters of the wood to be processed and has a high training precision.