极限学习机在车身前纵梁装配中误差预测应用

The application of extreme learning machine to the prediction of error state in front side rail assembly

在智能制造推动下,制造业对大数据的收集与特征分析愈加重视,数据分析技术更是大数据应用的关键技术,总结现有基于数据驱动的装配偏差控制方法,提出基于极限学习机建模的车身装配偏差预测控制方法,通过对检测数据的拟合建模,实现车身产品装配质量预测,并应用于制造生产线指导。文章应用极限学习机(Extreme Learning Machine,ELM)基于采集的车身前纵梁制造装配数据预测装配过程中的关键特征质量误差状态,从结果分析角度说明ELM准确预测误差状态。

In intelligent manufacturing,manufacturing characteristics of big data collection and analysis of more and more attention,the data analysis technology is the key technology of big data applications,summarizes the existing problem of assembling deviation control method based on data driven,modeling was proposed based on extreme learning machine body assembling deviation predictive control method,through the fitting modeling of test data,the assembly quality prediction of body products can be realized,and applied to the manufacturing line.This paper applied Extreme Learning Machine(ELM)to predict the mass error state in the assembly process based on the collected manufacturing assembly data,and illustrate the accurate prediction error state of ELM from the perspective of result analysis.