论文部分内容阅读
提出以竹粉模压花盆跌落冲击响应为对象的预测模型。利用均匀设计和有限元分析技术获得试验数据,以花盆周壳厚度、底沿厚度和底壳厚度为网络输入,花盆壳体最大应力为网络输出,构建三层BP神经网络;采用遗传算法(GA)优化BP神经网络的初始权值和阈值,利用均匀试验数据对GA-BP网络模型进行训练和仿真,最后利用GA-BP网络模型预测花盆跌落冲击的壳体最大等效应力。结果表明,所建立的GA-BP网络模型具有较强的学习能力,预测值与有限元分析值相对误差小于5%,表明GA-BP网络模型可用于花盆跌落冲击响应的预测。
The prediction model of drop impact response of bamboo powder molded flowerpot was proposed. The experimental data were obtained by means of uniform design and finite element analysis. The thickness of the perimeter shell, the thickness of the bottom edge and the thickness of the bottom shell were used as the input of the network. The maximum stress of the flowerpot shell was the output of the network and the three-layer BP neural network was constructed. (GA) is used to optimize the initial weights and thresholds of BP neural network. The GA-BP network model is trained and simulated with uniform test data. Finally, the maximum equivalent stress of the shell with drop impact is predicted by GA-BP network model. The results show that the established GA-BP network model has a strong learning ability, and the relative error between the predicted value and the finite element analysis value is less than 5%, indicating that the GA-BP network model can be used to predict the drop impact response of the flowerpot.