This study explores Machine Learning (ML) integration for household refrigerator efficiency. The ML approach allows to optimize defrost cycles, offering energy savings without complexity or cost escalation. The paper initially presents a State-of-the-Art of ML potential to improve functionality and efficiency of refrigerators. Since frost is the cause of significant energy losses, a ML-based Virtual Sensor was developed to predict frost formation on the evaporator also in low -level refrigerators. The results show the environmental significance of ML in enhancing appliance efficiency.

The Agenda 2030 calls for collective awareness, starting with individuals. The interaction between users and household appliances produces a relevant amount of data that can be elaborated through Machine Learning algorithms to guide users towards sustainable behaviours. In particular, the data already available on household appliances can be conveniently used to create Virtual Sensors, increasing the overall information about the system. This paper focuses on the description of the pipeline for the creation of Virtual Sensors and applies it to a no-frost refrigerator. The Data Acquisition phase is described and feeds the Model Creation phase. For the case study, the data have been discretized and labelled to train a Random Forest algorithm. The validation of the model has been done on an independent dataset. An analysis of the minimum prediction accuracy required for the model is reported. Furthermore, experimental data shows the effect of hot load positioning on the compressor's working time rate.