摘要
采用神经网络技术建立了沉积坯特征尺寸模型,该模型描述了喷射成形关键工艺参数对沉积坯尺寸的影响规律,模型输出的相对误差为6.58%,RM S(均方差)为0.372mm。模型的仿真结果给出了沉积坯尺寸的变化规律,其中稳态仿真结果可用于预先确定喷射实验中所采用的合适工艺参数;而动态仿真结果表明,雾化气体压力和沉积器平移速度对沉积坯几何尺寸都有较大影响,其中沉积器平移速度具有调节范围大的优点,成为调节沉积坯几何尺寸较合适的工艺参数。
Neural network technology was applied to establish a modeling of the characteristic dimension of spray deposited preform, which described the influence of the spray forming parameters on the deposit dimension. The relative error of the modeling output was 6.58% and the RMS （root-mean-square） was 0. 372mm. The relationship between deposit characteristic dimension and processing parameters was given by simulation results of the neural network modeling, and suitable parameters were defined according to simulation results of static spray forming processes; the simulation results of dynamic spray forming processes showed both atomizing gas pressure and translating speed of substrate were important factors influencing deposit characteristic dimension. Furthermore, it was preferred that the translating speed could be adjusted in a wide range, and became more suitable and effective parameter to control deposit dimension.
出处
《材料工程》
EI
CAS
CSCD
北大核心
2005年第8期15-19,共5页
Journal of Materials Engineering
基金
金属精密热加工国家重点实验室开放课题资助项目(51471040101JW0301)
国家自然科学基金资助项目(50174022)
关键词
喷射成形
神经网络模型
沉积坯
特征尺寸
spray forming
neural network modeling
deposit
characteristic dimension
