| 摘要 |
<p>PURPOSE:To shorten the period of a roughing step by learning and inferring a working condition which allows an electric discharge machine to maximally exhibit its capability for a workpiece during a roughing step of an electric discharge machining work, and by automatically advancing the work in accordance with thus inferred working condition. CONSTITUTION:With the use of a neural network which follows the result of learning, an IP (maximum working current) value in a final roughing condition at the time when a desired working depth is achieved, is inferred in view of an input working area, input electrode decreased dimensions and an input working depth (by a working condition inferring part 2). However, thus inferred PP value is in general not an integer, and accordingly, the maximum inferred current value cannot be set in an electric discharge machining device. However, of IP values which can be set in the machining device, the one which does no exceeds the inferred IP value and which is neared thereto is selected as a value which is used at a final roughing step.</p> |
