当前位置:首页 >
基于小波神经网络的凸轮轴铸造过程数值仿真研究
Research on Numerical Simulation for Casting Process of Camshafts Based on Wavelet Neural Network
浏览(3660) 下载(0)
- DOI:
- 作者:
- 张俊,付正飞
- 作者单位:
- 襄樊学院机械工程系,湖北 襄樊 441053
- 关键词:
- 凸轮轴;铸造过程;小波神经网络;数值仿真
Camshaft; Casting process; Wavelet neural network; Numerical simulation
- 摘要:
- 利用小波分析和神经网络智能技术,针对凸轮轴铸造充型凝固过程的温度场数值求解问题,
提出了小波神经网络算法。用热电偶对凸轮轴铸造温度场进行了实测,并以实测数据为样本
进行小波神经网络学习和训练,由训练后的神经网络仿真了凸轮轴铸造过程的温度分布。实
践表明,小波神经网络数值仿真快速、准确、合理,仿真结果与实测数据相比最大相对误差
为1.83%,可为铸造工艺参数确定提供理论依据。
With the technique of wavelets analysis and neural network intelligence, aiming
at numerical calculation of temperature field in casting process of camshafts, t
he arithmetic of wavelet neural network is described. The solidifying temperatur
e of camshafts is tested by thermocouples, the data of specimens obtained by t
he testing results of temperature field is trained by backpropagation
neural network, the temperature distribution during filling and solidification i
s simulated. By contrasting the data of simulation with those of testing, the ma
ximum relative errors of simulation is 1.83%, and the theory basis is presented
for technological parameters.