Mechatronics Maintenance: Detection and Diagnosis of Bearing Faults in a Fan Using Digital Signal Processing in MATLAB

One of the most common failures in computer hardware is the slowness and lack of computing power due to dust accumulation in the fan, reducing its risk. In the face of this problem, vibrational analysis, as an area of predictive maintenance, represents a tool that is beneficial in predicting when equipment is going to fail and carrying out repair work based on it. This is why innovative research has been proposed under the supervision of the School of Engineering Mechatronics at Ricardo Palma University of a mechatronics system for the predictive maintenance industry. The mechanical design consisted of a case to support the motor fan for an isolated test, the electrical design consisted of a Hall effect sensor to detect the RPM variation in a certain period, as well as, a temperature sensor in the center and one corner to compare the variations between these points. In the electronic design, the Arduino UNO controller is selected based on the system operating voltage of 5 V to supply the complete system. The MATLAB interface displayed the results of the frequency spectrum compared to the regular operation of a good-condition fan, and simultaneously showed the variations in RPM and temperature conditions, in conclusion, favorable results were achieved. For future research, it is proposed to use a MAT LAB app to automatically incorporate all of this data while capturing the sound directly through a mobile phone and including a thermal camera to visualize the temperature distribution in the engine.