Predictive maintenance based on vibration analysis has emerged as an effective strategy for ensuring the reliability of mechanical systems. This work investigates the application of artificial intelligence (AI) models for automated fault detection in refrigeration chambers, with a focus on critical components, including compressors, evaporators, and condensers. The research is based on the hypothesis that subtle variations in vibrational and thermal signals are indicative of incipient failures, which machine learning algorithms can detect. The experimental methodology involves instrumenting a refrigeration chamber with temperature sensors and piezoelectric accelerometers, which are connected to a data acquisition system based on the NI CompactDAQ and Arduino platforms. To date, tests have been conducted under normal operational conditions and simulated refrigerant leakage faults, with gradual load removal. The acquired data underwent preprocessing, including noise removal and normalization, followed by the application of Fast Fourier Transform (FFT) and the extraction of statistical metrics in the time domain, such as mean, standard deviation, and kurtosis. Preliminary results revealed stable vibrational patterns under normal conditions and subtle alterations in frequencies associated with compressor operation during leakage fault scenarios, validating the capability of sensors and methodology to capture relevant anomalies. Upon completion of data collection, machine learning models will be trained using supervised approaches (such as Random Forest and XGBoost), unsupervised methods (such as One-Class SVM), and baseline-based techniques, implemented in Python using the PyCaret library. The expected outcome is the development of an automated and reliable predictive diagnostic system, aligned with Industry 4.0 guidelines, with the potential to improve energy efficiency and reduce operational costs in industrial refrigeration systems. This research contributes to the advancement of structural health monitoring in refrigeration infrastructure, demonstrating the integration of vibration analysis with modern AI techniques for enhanced fault detection capabilities.