基于神经网络和遗传算法的熔融沉积成型多目标优化
Multiobjective Optimization of Fused Deposition Modeling Based on Neural Network and Genetic Algorithm
纪良波1,2, 周天瑞1, 钟雪华1
点击:3033次 下载:0次
作者单位:(1.南昌大学 机电工程学院,江西 南昌 330031; 2.九江学院 机械与材料工程学院,江西 九江 332005)
中文关键字:熔融沉积成型;多目标优化;神经网络;遗传算法;产品精度
英文关键字:used deposition modeling; multiobjective optimization; neural network; genetic algorithm; product
中文摘要:在建立了熔融沉积成型工艺参数和目标函数的基础上,论述了人工神经网络和遗传算法在熔融沉积快速成型工艺参数优化中的应用。首先利用人工神经网络建立熔融沉积快速成型工艺参数与成型件尺寸精度、翘曲变形之间关系的数学模型,然后用遗传算法对工艺参数优化。根据多目标函数优化问题的单目标化思想,对优化后的单目标进行了分解,得到最优工艺参数条件下的成型件尺寸精度、翘曲变形,从而为建立和控制熔融沉积快速成型工艺参数提供了一种行之有效的途径。
英文摘要:Based on the construction of the process parameters and objective function of fused deposition modeling, the application of aritificial neural network and genetic algorithm for processing optimization of fused deposition modeling was discussed. First of all, the mathematics model between the process parameters of fused deposition modeling and the precision and warpage of the part was set up with neural network. Then, the process parameters were optimized with genetic algorithm. According to the idea of converting the multiobjective optimization problem into the single-objective one, the optimal single-objective was decomposed and the precision and warpage of the part under the optimal condition was obtained. So, it is an effectual means for setting up and controlling parameters of modeling.