Condition Monitoring of Surface Grinding Process Using Low-Cost Vibration Sensor Module

A continuous and expertise condition monitoring affects significantly to improve the grinding process performances by timely predicting the grinding process stability (dressing requirement/replace/fitting issues) and grinding wheel condition. Real-time monitoring through the high-cost industrial sensor (dynamometer, acoustic emission sensor, power sensor, accelerometer, etc.) is mostly constraint with economic concerns. Thus, the utility of low-cost sensors modules for grinding process monitoring provides a valuable option to develop a compromised/affordable/portable solution. Here, a low-cost acceleration and vibration sensor-based setup is developed, which is utilized for real-time condition monitoring of the surface grinding process. The real-time vibration signals in x-, y- and z-directions are obtained through ADXL-335 sensor which is used to predict the overall condition of the process, which is labelled as accepted, moderately accepted and not-accepted conditions. Pattern recognition-based neural network is utilized for decision-making. Process condition monitoring is performed for two different materials, i.e. mild steel and aluminium. Significant success rates (about 85%) in predications are achieved in both the case studies. The results can further be improved by improving computational robustness and by use of multiple sensors.