摘要
以GCr15激光淬火为例,将基于遗传进化的神经网络(GANN)技术应用于激光淬火技术领域,建立了激光淬火性能预报GANN模型;确定了激光淬火(多)工艺参数与(多)性能指标的函数映射关系,实现了激光淬火性能预报。试验结果表明,在试验的工艺参数范围内各项性能预报值与实测值相比,偏差在7％以内,从而验证了该性能预报GANN模型的可靠性。
Taking GCr15 laser surface quenching for example, the authors apply the neural network technology based on genetic evolution(GANN) to laser quenching field, build the prediction model of laser quenching performance, decide the function reflection relation between the technological parameters and performance indexes of laser quenching and realize the prediction of laser quenching performance. The experiment results show that the errors between the prediction value and the practical measure are less than 7%, proving the reliability of the performance prediction model.
出处
《北京工业大学学报》
CAS
CSCD
北大核心
2003年第3期377-380,共4页
Journal of Beijing University of Technology
基金
北京市自然科学基金(2032003)
关键词
神经网络
遗传算法
激光淬火
neural network
genetic algorithms
laser surface quenching
