基于神经网络算法的飞机部件制孔末端执行器结构设计方法研究

Research on structural design method of drilling end-effector for aircraft components based on neural network algorithm

制孔末端执行器是飞机部件自动制孔的主要功能集成设备,其采用的设计方法对制孔工艺有重大影响。从结构设计部分出发,基于神经网络拟合算法,采用80组包含3类框架组件材料、位移和质量的5维向量样本数据,训练获得均方值MSE为0.06的网络模型,从而通过预测获得末端执行器框架的关键尺寸参数,并进而采用拓扑优化方法对框架组件进行轻量化设计。拓扑优化前框架结构在实际工况下最大变形量为0.04 mm、最大应力值为44 MPa,刚强度符合要求;同一工况下,拓扑优化后的框架结构最大变形量为0.026 mm、最大应力值为42 MPa,在质量降低43%的同时,制孔末端执行器框架结构的性能也得到改善。基于神经网络拟合算法对框架关键尺寸的预测,使得设计不再简单依靠经验判断,为末端执行器的结构设计开辟了新思路、新方法。

Drilling end-effector is the main function integrated equipment for automatic drilling of aircraft components,and its design method has a significant impact on the drilling process.Starting from the part of structural design,based on the neural network fitting algorithm,80 groups of 5-Dimensional vector sample data including three types of frame components material,displacement and mass are used to train the network model with a mean square value MSE of 0.06,and the key dimensional parameters of the end-effector frame through prediction are obtained,and then the topology optimization method is used to design for lightweight of the frame components.Before topology optimization,the maximum deformation of the frame structure under actual working conditions is 0.04 mm,the maximum stress value is 44 Mpa,and the requirements of the stiffness and strength are met;Under the same working condition,the maximum deformation of the frame structure after topology optimization is 0.026 mm and the maximum stress value is 42 MPa.While the mass is reduced by 43%,the performance of the frame structure of the drilling end-effector is also improved.The prediction of the key dimension parameters of the frame based on the neural network fitting algorithm makes the empirical judgment is no longer simply relied on to design,and the new idea and method for the structural design of the end-effector is developed.