摘要
为了提高数控机床对测试误差的补偿效果,开发一种通过遗传算法(GA)来完成BP网络的优化过程,并加入坐标参数、运动速度指标,建立工作台的定位误差仿真模型。先通过Matlab软件构建得到GA-BP模型,得到优化权值与阈值后,再将结果移植至DSP内开展建模与预测,由此促进预测速率的大幅提升。研究结果表明:以DSP构建的预测系统对各定位误差残差分布进行预测得到的范围是-0.69~0.51μm。采用GA-BP网络构建的模型进行预测时达到了更高精度。
In order to improve the compensation effect of CNC machine tools for test errors,a genetic algorithm(GA)was developed to complete the optimization process of BP network,and the coordinate parameters and motion speed indicators were added to establish a simulation model of positioning error of the worktable.The GA-BP model was constructed by Matlab software,and the optimization weights and thresholds were obtained,and then the results were transplanted to the DSP for modeling and prediction,which greatly improved the prediction rate.The results show that the range of the residual distribution of each positioning error is-0.69~0.51μm by the prediction system constructed by DSP.The model built by the GA-BP network achieves higher accuracy in prediction.
作者
庞晓霞
PANG Xiaoxia(Intelligent Manufacturing College,Zhengzhou City Vocational College,Zhengzhou 452370,Henan China)
出处
《锻压装备与制造技术》
2024年第1期75-77,共3页
China Metalforming Equipment & Manufacturing Technology
