Desenvolvimento de sistema de baixo custo para manutenção preditiva de motores de indução trifásicos via análise de vibrações com tecnologias IOT

Abstract

Three-phase induction electric motors are widely used throughout the industry due to their high efficiency, versatility, and low cost. As these are highly critical pieces of equipment, it's important to have an appropriate maintenance plan to prevent unplanned downtime due to equipment failure. In this regard, implementing predictive maintenance for continuous equipment monitoring can contribute to identifying potential failures and adequately planning maintenance actions. This work aims to develop and implement a low-cost system for monitoring and analyzing the vibration of a three-phase induction motor, using Internet of Things (IoT) technology. The system transmits data in real-time, enabling continuous vibration analysis of motor performance and the identification of possible failures. With a focus on low application cost and efficiency, the system integrates sensors and a wireless communication protocol to provide accurate and immediate information. To validate the system, tests were conducted on three motors with distinct operating conditions, and the results obtained reflected the actual condition of the evaluated motors. In Case A, the motor was in normal condition, and RMS and spectral analyses showed no signs of faults. In Case B, the motor had mechanical looseness; RMS values for all three positions remained below the alert classification, but the vertical position's value was higher, signaling a possible fault. The sub-harmonic analysis, confirmed the presence of mechanical looseness, especially in the horizontal measurement of 2.5XRPM, which reached an amplitude of 3.8 mm/s. In Case C, the motor had bearing failure; RMS vibration levels were critical according to standards, indicating an advanced stage of failure, which recommends immediate equipment shutdown for corrective maintenance. The spectral analysis of the bearings, especially in the axial position, revealed high amplitudes in the characteristic frequencies of FTF, BPFO, and BSF, with BSF close to the 0.8 mm/s limit, and a spectrum filled with noise, confirming the bearing fault. The proposed solution proved to be viable and low-cost, with the potential to be implemented in a predictive maintenance system for three-phase induction motors.