基于多误差源耦合的五轴数控铣床加工误差综合预测及评判

在综合考虑机床动静态多种误差源的基础上，建立了各运动轴伺服运动模型和多体联动模型，给出了刀具的实际运动位置和姿态，基于包络理论求解了曲面加工实际成形面，对比理想数学模型，对加工误差进行了综合预测和评判。以复杂非可展曲面--S试件为例，给出了S试件的铣削精度构建方法，分析了机床动态因素(位置环、速度环等)对零件铣削精度的影响，并通过切削实验后的数据回归分析予以验证。建立了基于神经网络的机床铣削误差辨识模型，用于评估机床加工后的状态。该平台的搭建为实现大型、关键零件的加工精度预测和保障提供了技术支撑。 

Prediction on Machining Accuracy of Five-axis Milling Tool for Multiple Error Source Coupling

 A method integrated with dynamic error factors and static geometric errors was presented to build the surface data of actual workpiece. Each servo axis movement was simulated and composed by kinematics of the joint bodies of the machine for calculating the actual position and attitude of the tool. The actual milling point was solved by the envelope theory and the final part of the surface was obtained by the surface forming method. A case study was analyzed through the S specimen sample, which was verified the composition of surface methods and get the milling errors caused by influences of dynamic factors, such as the gap, the position loop and speed loop. The results were verified by experimental data through regression analysis. Finally, the error track model was established based on neural network for condition assesment after machining. The developed platform may provide technical support for the realization on precision forecasting and security for large-scale and the key structure.