Development of high-speed and high-accuracy machining system based on servo data

本研究的目的是开发能够适应下一代并保持世界领先地位的高速高精度加工系统。近年来，即使使用普通机床，对模具的加工精度要求也越来越接近微米量级。目前的加工系统是由CAD、CAM、NC和机床组成的，每个加工系统都是由其他制造商单独开发的。为了实现新的高速高精度系统，需要将这些元件与数据流进行集成和融合，在本研究中，每个元件通过一个伺服数据进行标准化，作为控制机床工作台中的伺服电机的最终数据。伺服数据是每个插补时间的命令位置。从目标形状的几何模型生成这种伺服数据称为几何插补。几何插补直接计算出与目标形状几何形状相适应的加减速位置的伺服数据。另一方面，在实际加工过程中也会产生运动误差、刀具偏转误差等误差。为了补偿这些误差，提出了预测补偿方法。利用伺服数据通过数控加工仿真计算预测误差，然后根据预测误差对伺服数据进行修正，提出了几何插补和预测补偿方法，并通过跟踪位置的实验测量进行了验证。实验的物形由圆、圆弧和直线组成，补偿后的误差为运动误差。研究结果表明，所提出的系统能够有效地减小由NC程序插补误差和运动误差引起的位置误差。

The purpose of this research is to develop the high speed and high accuracy machining system that is capable of responding appropriately to next generation keeping advantage in the worldwide. Recently the requirement of machining accuracy for die and mold becomes close to the micro meter order even if using the common machine tools. Current machining system is composed with the element of CAD, CAM, CNC and machine tool and each one is developed individually in other makers. It is necessary to realize the new high speed and high accuracy system that these elements are integrated and fused with data stream.In this research, each element is standardized by a servo data which is a final data to control servo motors built in the table of machine tool. The servo data is a command positions per interpolation time. Generation of this servo data from the geometric model of objective shape is named the geometrical interpolation. The geometrical interpolation calculates servo data considering acceleration and deceleration position which are suitable for the geometry of objective shape directly. On the other hand, the errors such as motion error, tool deflection error and so on occur during actual machining processes. To compensate these errors, the prediction compensation is developed. The prediction errors are calculated by NC machining simulation by using servo data, then the servo data is corrected based on prediction errors.In this report, geometrical interpolation and prediction compensation are developed and validated by experimental measurement of traced positions. The objective shape is combined of circle, arc and line of the experiment and compensated error is a motion error. As a result of this research, it is evident that proposed system is able to reduce the position error caused by interpolation error of NC program and motion error.