摘要
设计了一个基于VisualC++平台的多输入多输出的BP神经网络程序,依据正交实验设计训练样本,实现了注射成型产品多质量指标的高精度预测,可实现对注塑产品质量的监控,提高了实际生产效率。
According to the training samples designed with orthogonal experiment, the author designed a MIMO neural network by using Visual C++, which could precisely forecast multiple quality targets of the injection molding products. The outcome showed the system improved the productivity and achieved the quality monitor of rejected products.
出处
《模具制造》
2008年第3期62-64,共3页
Die & Mould Manufacture
基金
广西科学基金资助项目(编号:桂科基0575012)