Monitoramento do desgaste da ferramenta no processo de fresamento de aço P20 através da medição da potência do motor

Abstract

This master's thesis proposes to develop a tool wear monitoring system during the machining process by measuring the electrical power of the spindle motor. Direct measurement methods are more reliable, but can only be used when the tool is out of operation. The solutions for online wear monitoring can be the measurements of the following items: cutting force, acoustic emission, vibration, electric current, cutting temperature, part dimensions, ultrasonic and optical signals, surface finish, which can be used individually or jointly, for better monitoring of the process. Initially, a review is carried out on the milling process, tool wear, wear measurement methods, in particular, monitoring by measuring the electrical power. Subsequently, the method for calculating electrical power, the operating principles of an induction motor, the control of motor speeds, digital signal processing and artificial neural networks (ANNs) are highlighted. Then, the work methodology is presented, that is, how the data acquisition board was developed, what were the materials, equipment and methods used in the tests. The experimental results show the construction and calibration of the data acquisition system, the tests to determine the electrical power consumed by the spindle motor as a function of wear using different machining conditions. With these data in hand, ANN training is carried out and the results are checked, comparing the measured wear value with that estimated by the network. The ANN used is a Feed-Forward, with training through the BackPropagation algorithm Levenberg-Marquardt. This process obtained satisfactory and promising results, identifying a great potential for the use of this measurement of electrical power to predict the wear of the machining tool.