Fault detection and diagnosis in refrigeration systems using machine learning algorithms

The functionality of industrial refrigeration systems is important for environment-friendly companies and organizations, since faulty can impact human health by lowering food quality, cause pollution, even lead to increased global warming. Therefore, in this industry, there a high demand among manufacturers early automatic fault diagnosis. In paper, different machine learning classifiers are tested find the best solution diagnosing twenty faults possibly encountered such systems. All sensor some relevant component simulated fidelity Matlab/Simscape model system, which has previously been used controller development verification. work, Convolutional Neural Networks, Support Vector Machines (SVM), Principal Components Analysis-SVM, Linear Discriminant Analysis compared. results indicate that detection reliability algorithms highly depends on how well training data covers operation regime. Furthermore, it found well-trained SVM simultaneously classify types with 95% accuracy when verification taken from system configurations.