Applying fuzzy logic for hybrid control of grinding work

Machining dimensions fluctuate for many reasons in machining and processing. Dimensional variations are a mixture of trend, adjacent, sudden and other variations. Analyzing the causes of these variations is very difficult. Working accuracy is currently maintained by the timely correction of shop workers. We analyzed dimensional variations and found that most of these are trend variations. We also found that we can improve working accuracy by compensating for the trend variations. To improve working accuracy through correction of trend variations, we developed a system to automatically correct the sizing points of a machine control gauge based on the information transmitted from a total inspecting machine located in the post-processing end of an in-process controlled measuring instrument. The values input to this control system are discrete information sources and the variations are complex. Further, there is the factor of variable deadtime. For these reasons, we have developed a feedback system consisting of an appropriate filter to correct only the trend variations, ignoring the effect of adjacent and sudden variations, and a fuzzy calculator with non-linear quadratic control based on worker knowhow. This paper reports a working example of the system which was introduced into the crank center journal and pin grinding line.<
>