工具機精度檢測技術的研發與應用

Abstract

控制與檢測調校是製造工業上非常重要的技術，控制與檢測調校之性能與品質的優劣對於加工機的精度、效率和產品品質都有直接而顯著的影響。而工具機的誤差量測與分析，已經發展了相當久的歷史。為了有效提昇工具機的精度，對於多軸同動所產生之多種誤差參數合成的誤差現象的研究，變得相當重要。 本文之研究目的，是使用D-H 齊次變換矩陣發展出的誤差模式，將其套用於工具機連續桿件上可求出輸出端與參考原點的相對位置矩陣，進而推導出三軸工具機的累積誤差模式，建立一套能夠快速計算各桿件誤差行為影響的數學系統。再藉由D-H齊次座標變換矩陣所推導出的三軸工具機誤差模式，以現成本實驗室發展之三軸工具機台，輔以實驗計劃法的直交表配置表現此系統特徵以及類神經網路的學習功能，從量測出來的終點誤差數據中反推求出各桿件機構參數的誤差量。可快速判斷誤差來源，將可提供校準相關人員參考之用，以提升工具機的調校效率。

Control and calibration is key technology in manufacture industry. It's performance and quality has notable and direct influence for machine's precision, efficiency and manufacture's quality. Error Analysis and Measurement of a Machine Tool have been developed for long years. In order to promote machine's precision, the research of error from multi-axial machine is very important. The purposes of this thesis are to utilize the D-h homogenous transformation matrix to describe the position of three axis machine and to build an error analysis model, for building a mathematics system which can rapidly calculate the influence of error behavior for multi-axial machine. Then we use the model for three-axis machine tool assisted from orthogonal array table and artificial neural network to solve for multi-axial machine's error to find the error source.