K. Viswanathan,∗ A. Krishnakumari,∗∗ and D. Dinakaran∗∗∗
Prediction, grinding, wheel loading, vibration
This work presents a methodology to detect the extent of wheel loading in a surface grinding machine. Prediction model of wheel loading is done based on its secondary vibration signals. These vibration signals are captured by a piezoelectric sensor placed on the grinding spindle housing. The signals are captured for various conditions of the wheel loading, and the results are analysed both in time and frequency domain. The Lab VIEW software is used to perform the signals analysis. The results show that power spectra of vibration analysis in the range of 11–26 (m/s2) × 10−3 RMS are sensitive to wheel loading for the operating conditions in which it is tested. The regression model is developed for the prediction of wheel loading with time and frequency domain parameter as inputs which shows ±4.41% of error that is experimentally reasonable. Also a model using artificial neural network is built to predict the wheel loading through online, and the results are in good agreement with experimental values.
Important Links:
Go Back
