论文部分内容阅读
在使用径向基神经网络建立注塑工艺模型时,虽然能够得到较好的模型,但是建模时训练样本数量将会对模型的质量产生较大的影响。本研究对建模所需的原始样本数据首先进行信息扩散处理,然后再使用径向基神经网络建立注塑工艺参数与塑件沉降斑指数之间的模型。从结果上看,在注塑训练样本数量相同的情况下,运用该方法均可以得到优于仅使用普通径向基网络构建的模型。
When using RBF neural network to establish the injection molding process model, although the model can be better, the training sample size will have a greater impact on the quality of the model. In this study, the original sample data needed for modeling were firstly processed by information diffusion, and then the radial basis function neural network was used to establish the model between the injection molding process parameters and the index of plastic part settlement. From the result point of view, when the number of injection training samples is the same, the proposed method can be better than the model constructed using only ordinary radial basis network.