【摘要】 采用机器学习方法将试验数据与仿真分析相结合,对熔模铸造过程中的温度、冷却速度、合金成分等进行系统优化。通过收集大量的铸造工艺数据,构建基于深度学习的工艺参数预测模型,对工艺参数进行优化设计,利用现场试验对建模型进行了验证。结果显示,采用机器学习方法优化铸造参数,可明显改善铸件组织及力学性能,降低缺陷率,提高生产效率和经济效益。在机器学习背景下,参数组对比中,最优参数组的铸造质量评级明显高于次优参数组,且最优参数组只有1组低于90分,而次优参数组均低于90分。更多还原

【Abstract】 Machine learning methods was applied to combine experimental data with simulation analysis to systematically optimize the temperature,cooling rate,alloy composition,etc.,during investment casting process.By collecting a large amount of casting process data,a prediction model for process parameter based on deep learning was constructed to optimize the design of process parameters,and the model was validated through on-site experiments.The results indicate that the microstructure and mechanical properties of castings are significantly improved after optimizing casting process by machine learning methods,and the defect rates are reduced,enhancing productivity and economic benefits.In the context of machine learning,the casting quality score of the optimal parameter array is significantly higher than that of suboptimal parameter array in the comparison of parameter groups,and only one group of the optimal parameter array gets below 90 points,while the suboptimal parameter groups all get below 90 points.更多还原