论文部分内容阅读
采用神经网络算法技术,以钒含量、钛含量、淬火温度、淬火冷却方式、回火温度和回火冷却方式作为输入层参数,以耐磨损性能和冲击韧性为输出层参数,可以构建出6×24×12×2四层拓扑结构的钒钛改性高铬铸铁热处理工艺优化模型。模型输出的耐磨损性能平均相对预测误差为2.8%、冲击韧性平均相对预测误差为2.5%。模型不仅具有较佳的预测能力和较高的预测精度,而且在热处理生产线上具有很好的应用效果,使产线上的钒钛改性(0.8%钒+0.5%钛)高铬铸铁的平均晶粒尺寸减小32%、磨损体积减小50%、冲击韧性提高62%。
Adopting neural network algorithm technology, taking vanadium content, titanium content, quenching temperature, quenching cooling method, tempering temperature and tempering cooling method as input layer parameters, taking wear resistance and impact toughness as output layer parameters, we can construct 6 Optimization Model of Heat Treatment Process for Vanadium - Titanium Modified High Chromium Cast Iron with Four - Tier Topology. The average relative prediction error of the wear resistance of the model output is 2.8%, and the average relative prediction error of impact toughness is 2.5%. The model not only has better predictive ability and higher prediction accuracy, but also has a good application effect in the heat treatment production line, so that the average vanadium-titanium modification (0.8% vanadium + 0.5% titanium) high chromium cast iron on the production line The grain size is reduced by 32%, the wear volume is reduced by 50% and the impact toughness is increased by 62%.