论文部分内容阅读
采用5×30×6×1的四层拓扑结构,以铝合金牌号、模具预热温度、始锻温度、终锻温度、锻压速度作为输入层参数,以抗拉强度作为输出层参数,构建了6系铝合金接头锻压工艺神经网络模型,并对其进行了训练、预测和验证。结果表明,模型预测性强,精度性较高,平均相对预测误差值仅为3.3%。和生产线传统用工艺锻压的试样抗拉强度相比,运用神经网络模型优化工艺锻压的试样抗拉强度增大了16%,拉伸性能得到了显著的提高。
A four-layer topological structure of 5 × 30 × 6 × 1 was used as the input layer parameters based on aluminum alloy grade, mold preheating temperature, initial forging temperature, final forging temperature and forging speed, and tensile strength was taken as output layer parameters 6 Department of aluminum alloy forging process neural network model, and its training, prediction and verification. The results show that the model has strong predictability and high accuracy, and the average relative prediction error is only 3.3%. Compared with the tensile strength of the traditional forging process sample in the production line, the tensile strength increased by 16% and the tensile property was significantly improved by the neural network model optimization forging process.