M963镍基高温合金疲劳性能的神经网络预测

Fatigue performance prediction of M963 Ni-based superalloy based on neural network

基于神经网络技术,根据M963铸造镍基高温合金疲劳试验数据,建立了高温合金疲劳性能预测模型,并进行了M963铸造镍基高温合金相关疲劳验证和分析。结果表明:从提高M963高温合金疲劳性能的角度出发,合金元素添加量(质量分数)优选为7%Cr+5%V+0.5%RE;熔体过热温度优选为1 030℃;表面超音速火焰喷涂合金粉末配比优选为Cr-20Al-30V-0.5Y。

According to the experimental data of fatigue test, prediction model of fatigue performance of M963 Ni-based superalloy was built by using neural network technology, and experimental validation and analysis were carried out. The results show that for improving the fatigue performance of M963 superalloy, the optimized amount of alloying elements is 7%Cr＋5%V＋0.5% RE;the optimized melt superheat temperature is 1 030 ℃ ; the optimized matching ratio of alloy powder for high-velocity oxygen fuel spraying is Cr-20A1-30V-0.5Y.