目的 通过神经网络优化不锈钢多元渗氮设计工艺,提高效率与质量。 方法 分析循环周期、乙醇+CS₂ 气体流量和氨气流量对奥氏体不锈钢 316 Ti 共渗后耐磨、耐腐蚀性能的影响,建立 BP 神经网络,通过预测,寻求不锈钢循环多元离子共渗处理工艺参数。 结果 对利用 BP 神经网络进行工艺优化设计的工艺参数进行实验验证,理论数值与试验数值的偏差在依5% 以内。 结论 新方法脱离传统的经验法与试验法,实现了循环多元离子共渗的优化设计。

Objective To improve the efficiency and quality through the neural networks optimization of multiple ion nitriding process on stainless steel. Methods The influences of cycle, ethanol, CS₂ and ammonia flow on the wear resistance and corrosion resistance of austenitic stainless steel 316Ti in nitriding were analyzed. A BP neural network was established, and parameters for the stainless steel circular multiple ion nitriding process were seeked through prediction. Results The process parameters error was optimized within 依5% by using BP neural network to validate the technical parameters of the process optimization. Conclusion Different from the traditional empirical method and test method, the new method realized the optimization design of multiple ion nitriding.