A predictive maintenance project commissioned by Arneg aims to develop a data acquisition system using accelerometers to monitor the vibrations of electric motors used in refrigeration machines by Arneg. The goal is to predict potential mechanical failures using data analysis methodologies.

The technological solution developed involves an integrated system with two accelerometers for each motor, allowing the measurement of vibrations in both axial and radial directions, and an IoT gateway for data collection, pre-processing, and transmission to a remote service center dedicated to data storage, analytics, and visualization.

How it works:
The system monitors the vibration intensity from three motors in a refrigeration unit at a large-scale distribution center (GDO). The accelerometers continuously capture the vibration signals, and a Discrete Fourier Transform (DFT) is applied to estimate the mode (amplitude and frequency) of the vibrations.

The processed data is collected and transmitted via an IoT gateway to a remote data collection platform, where it is displayed in an amplitude-frequency graph via a dashboard. A Graphical User Interface (GUI) has been developed to enhance the flexibility of database access, enabling detailed visualization of all acquired data.

Based on the collected data, heuristics can be applied by analyzing the available spectra, or inferences can be made by examining the system’s historical behavior to predict potential failures (e.g., mechanical imbalances, bearing issues).