加和函数及其在Ni基喷涂材料硬度研究中的应用

Study on Ni-based Spraying and Coating Material Hardness Via Additive Function

提出了加和函数的概念,根据Ni基喷涂材料的研制经验,给定一些基本概念和合理假设,并以此为基础对加和函数的表达形式进行理论上的推导,探讨人工神经网络在材料研究中存在的问题及其局限性,而后利用人工神经网络得到的影响函数对Ni基喷涂材料性能加以研究和预测。计算结果表明,该加和函数的表达形式比较准确地刻画了Ni基喷涂材料成分-性能相关性,预测效果明显优于人工神经网络。对于检验样本而言,绝大多数预测误差小于5.0HRC,满足工程要求。为材料性能研究和预测提供了新的计算处理工具,可以在新材料开发的过程中,使试验目的明确,从而降低开发费用,并缩短研制周期。

Based on the experience of researching and manufacturing of nikel-based spraying and coating material, some basic concepts and reasonable hypotheses are put forward. Additive function has been put forward and deduced to investigate the regularities of element influence on spraying and coating material hardness.The results of computation indicated that additive function to predict material hardness was obviously better than neural network,almost every error was less than 5HRC. So that ,additive function is a new effective assistant tool to research and rehabilitate this complex relation between elements and quality of material.