An online monitoring methodology for grinding state identification based on real-time signal of CNC grinding machine

Abstract

The grinding state is closely related to machining accuracy, wheel wear, and material removal efficiency. Changes in the grinding state often mean instability of the processing state, which will cause additional costs such as wheel wear and a decline in the processing quality of the workpiece. Various methods for monitoring grinding conditions have been proposed in the past, but none of these methods have been universally successful due to the complex nature of the machining processes. This research presents a new method for real-time monitoring of grinding forces and workpiece surface topography during grinding processing using real-time signals from the computer numerical control (CNC) system of the grinding machine without any additional sensors. By extracting real-time signals during the grinding process for time and frequency-domain analysis, the grinding state can be identified online. Based on this method, the time–frequency domain calibration experiment is carried out. The resolution of the time-domain calibration results reached 6e-5N, which can characterize the real-time change of grinding force during the grinding process. The frequency-domain analysis can achieve real-time monitoring of the spindle state of the workpiece and the grinding spindle state and obtain the frequency-domain transmission path under different processing conditions. The workpiece surface morphology is estimated in real-time using the feedback signal of the grinding machine, and the results are verified in mm, μm, and nm scales. The test results show that the use of real-time signals from the grinding machine to monitor the grinding state has the advantages of high precision, reliability, and convenient implementation.