Real-time monitoring of the dressing operation is becoming increasingly important, as this process has a fundamental role in the finishing of parts produced using the grinding process. This monitoring, however, is very costly and inefficient if done visually, which is still common in many industries. A complete monitoring using sensors is still an objective of many scientists, and has been studied for years. This work aims to meet this demand, presenting a study of the variation of the width of action of the dresser through signals obtained from piezoelectric capsules, digital signal processing and neural models. The piezoelectric capsules have a very low cost and can replace widely used sensors. Tests are going to be performed using an aluminum oxide grinding wheel and a single point dresser, varying the condition of the wheel surface to also investigate its influence in the variation of the signals. The sensor data will be collected by an oscilloscope with 2 MHz sampling rate and then processed and analyzed with the aid of MATLAB software. Artificial neural networks are going to be used to estimate or predict the width of action of the dresser. Data on the width of action of the dresser (bd) are going to be obtained through high definition images of the tip of the diamond, which are going to be captured over the diamond passes on the grinding wheel. The relationship between the values of bd and the obtained signal values from the piezoelectric capsule is going to be obtained and analyzed. It is expected through this research to contribute to the research in the area in order to provide an efficient and cost-effective for monitoring the dressing tool environment.
News published in Agência FAPESP Newsletter about the scholarship: