Development of Intelligent Key Feature Extraction System for Electrical Discharge Machining

• Main Authors: SU, TING-WEI, 蘇庭煒
• Other Authors: YANG, HAW-CHING
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/38201501820769383276

碩士 === 國立高雄第一科技大學 === 電機工程研究所碩士班 === 105 === Except to the electrode size, workpiece material, and processing parameter, accuracy of discharge machining is also affected by the variation in slagging capabilty, electrode wear, and discharge uniformity. For micro-machining, the key parts for estimating machining accuracy is how to effectively extract sensing features from the tiny varing singals. This research pessented an ISU-EDM (Intelligent Sensing Unit- Electrical Discharge Machining) to monitor the EDM process for effectively and timely collecting the signals of current, voltage, and axis position in machining. First, in the signal collecting, the proposed system included signal collection and data analysis to online monitor machining and analyzing processes. In signal processing, the effective discharge waveforms can be derived for obtaining the key feasture set by feature calculation and distribution fitting methods. Finially, the set is served as an input of the AVM (Automatic Virtual Metrology) system to evaluate final accuracy. The experimental results of machining hole with diameter 3 mm and depth 100 um indicate that the ISU-EDM can derive the key feature (1 KB) from the big signal data (13 GB) with 50 MHz sampling rate. After importing the AVM system, the MAE (mean absolutely error) of estimating roughness, diameter, and roundness of the hole bottom in free run are 0.59 μm, 8.31 μm, and 0.005 μm, respectively. Hence, the developed ISU-EDM is a promising tool for predicting accuracy of EDM process.