Machine learning-based virtual sensors for reduced energy consumption in frost-free refrigerators

Aug 4, 2026

This study investigates integrating machine learning (ML) to improve household refrigerator efficiency through more effective defrost-cycle control. By addressing frost, the main source of significant energy losses, the work develops an ML-based virtual sensor that predicts frost formation on the evaporator, even in lower-end units. The paper first reviews the state of the art on ML opportunities for refrigeration performance, then evaluates environmental significance of the proposed approach for real-world appliance operation.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Introduction

Problem context: energy use of refrigerators

Motivation: limitations of external factors data

Need for sensor-based information without added cost

Impactful situations analysis and case study selection

Existing AI/ML in domestic refrigeration

Case study rationale and setup overview

Machine learning-based virtual sensors for reduced energy consumption in frost-free refrigerators Alejandro Alcaraz 1 , Dennis Ilare 1,2,, Alessandro Mansutti 1 and Gaetano Cascini 2 1 Elettrotecnica ROLD, Italy, 2 Politecnico di Milano, Italy [dennis.ilare@polimi.it](dennis.ilare@polimi.it) Abstract 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. Keywords: artificial intelligence (AI), sustainability, case study, virtual sensors, machine learning 1. Introduction The increasing number of buildings, especially in emerging economies, is one of the key factors for the continuous growth of electricity consumption by appliances. Currently, most households own refrigerators, averaging 0.9 units per household, and it has become common to own more than one television, with an average of 1.3 units per household (International Energy Agency, 2023) . Since refrigerators have 24-hour functioning, they are responsible for significant electricity consumption: for instance, it has been estimated that in India, they account for 461 kWh on average (Prayas (Energy Group), 2021), representing approximately 27% of yearly consumption in a household. The introduction of the Energy Star Program in the early 1990s, led by the United States Environmental Protection Agency (EPA), marked significant progress in improving appliance efficiency. This initiative aimed to identify and promote energy-efficient products, including refrigerators. As a result, consumers were encouraged to choose greener options, leading manufacturers to compete in producing energyefficient models. In successive years, improved insulation materials, advanced compressors, and better temperature control systems have become standard features of home refrigerators. The industry has also seen the adoption of cyclopentane and other foams with reduced global warming potential (GWP), reducing the refrigerant’s environmental impact (Faruque et al., 2022) . Although regulations and technological advances in the field have led refrigeration systems to use relatively low power for regular operation, different studies have recognised external factors that affect refrigerators’ energy consumption during everyday use, such as room air temperature, unit ageing, design practices, and user interactions (Anjana et al., 2015) . Consequently, studies that minimise the impact of these factors have gained attention in recent years (Hueppe et al., 2021), especially because the European energy system is currently addressing an unprecedented crisis. It is worth mentioning that despite the identified relevance of the influence of these factors on the additional energy consumption of domestic refrigerators, there are no robust published data in this respect (Harrington et al., 2019) . In addition to optimising control based on factors such as user interaction, room air temperature, and unit ageing to enhance efficiency, the increased appliance intelligence is driven by other trends. These include focusing on food quality, sustainable lifestyles, and healthy eating practices. To address efficiency and adapt to new trends, a potential solution involves integrating specialised sensors to enhance the information level available for the appliances and consequently improve controlsand features. While this is technically feasible, it would cause an increase in the complexity and cost of the appliances. Hence, the addition of dedicated sensors seems to be poss

FAQs

What problem does the study address in frost-free refrigerators?

Frost formation causes significant energy losses during everyday operation, reducing overall efficiency.

How does the proposed ML-based virtual sensor work?

It uses machine learning to predict frost formation on the evaporator, enabling better defrost-cycle optimization without adding physical sensors.

Why is this approach especially relevant for low- and medium-price refrigerators?

It improves energy efficiency while avoiding increased appliance complexity and cost, which would be difficult with dedicated sensors for widely used models.

Aug 4, 2026

Upload by River Wang