Use of artificial intelligence in the refrigeration field

Nowadays more than 5 billion of refrigeration and air-conditioning units are employed worldwide. Internet of Things technology is becoming widely used also in these applications. Operational data can be collected from units installed worldwide and used to train artificial intelligence (AI) tools and black box digital twins. These models can predict energy consumption or detect fault and inefficiencies, allowing a timely intervention on the unit before a total lack of cold occur. In this paper a state-of-the-art review regarding the implementation of AI in refrigeration sector is presented. Works are classified according to the following topics: machine learning tools for fault detection and diagnosis (FDD), black box digital twin to predict energy consumption and performances, AI application for demand defrost, optimization algorithms for complex system.