Relearning necessity determination method and relearning necessity determination device of diagnostic model in machine tool, and readable storage medium

Abstract

The invention provides a relearning necessity determination method and relearning necessity determination device of a diagnostic model in a machine tool, and a readable storage medium. In the machining abnormality diagnosis of a machine tool, relearning necessity of a diagnostic model is determined when there is suspicion of an over-detection. A control device (2) executes the following steps: a tool usage amount storage step in which the accumulated cutting time or the accumulated cutting distance of the tool (3) mounted on the machine tool (1) is stored in a tool usage amount storage means (12) as a tool usage amount; an abnormality diagnosis tool usage amount storage step for storing, in an abnormality diagnosis tool usage amount storage means (14), the tool usage amount when a machining abnormality diagnosis means (13) diagnoses the machining abnormality; and a relearning necessity determination step of determining, by a relearning necessity determination unit (16), whether relearning is necessary on the basis of the frequency distribution of the tool usage amount stored in the tool usage amount storage step at the time of abnormality diagnosis.

Description

technical field [0001] The present invention relates to a method and device for judging whether a learned diagnostic model needs to be relearned in a machine tool equipped with a machining diagnosis function using machine learning technology, and a readable storage medium. Background technique [0002] When machining a workpiece with a cutting tool, tool damage such as cutting tool breakage may occur due to overload, biting of cutting powder, increased wear of the cutting edge, etc. In this case, the desired processing cannot be achieved and time is required for reprocessing. Not only that, there may be losses due to defective products, but there is also a danger of the machine colliding with the workpiece due to the generation of cutting residue. In order to avoid the above-mentioned danger, a technique of detecting a processing abnormality and stopping the machine has been proposed. [0003] Patent Document 1 discloses a technique in which a motor load waveform during nor...

AI Technical Summary

Problems solved by technology

If this situation persists, it will lead to an increase in the cost of the tool, and the possibility of over-testing will only be discovered at this point

However, when the tool cost has risen significantly, a considerable period of time has passed since the processing suspected of being over-inspected. At this time, it is difficult to discern the frequency of unnecessary tool replacement due to over-inspection. The tool cost increases due to the increase, or is the damage of the tool actually increased due to the processing conditions or the state of the machine, etc.

If re-learning is forcibly carried out when it is impossible to judge whether it is a threshold problem or an incomplete diagnostic model, in addition to the loss of time and cost, it may also lead to over-learning or wrong learning, resulting in a further increase in losses

Images