The introduction of advanced technologies such as the W6 edge device and machine learning offers new opportunities to improve the management of refrigeration systems: read on and find out how.

Ice buildup on heat exchangers is a significant challenge for refrigeration applications, compromising energy efficiency and system performance. Traditionally, control of this phenomenon has relied on static methods that are often ineffective and wasteful. However, the introduction of advanced technologies such as the W6 edge device and machine learning offers new opportunities to improve the management of refrigeration systems. Using IIoT sensors and cloud-based data analytics, dynamic control strategies can be implemented that optimize energy efficiency and reduce operating costs, marking a step toward a more sustainable and intelligent future for the refrigeration industry.

Ice and Machine Learning

One of the most common problems with applications using the refrigeration circuit, such as heat pumps or freezing systems, is the accumulation of ice on the heat exchangers. This phenomenon, caused by the condensation and solidification of water contained in the air on the heat exchangers, adversely affects energy consumption during operation. Ice can form on the heat exchangers of a heat pump exposed to the open air, used to heat rooms, or on the evaporators of cold rooms used to store foodstuffs.

Ice poses a problem because it is an insulator: its thermal conductivity is 5.2×10^(-4) kcal/ms°C, much lower than that of air (5.5×10^(-6) kcal/ms°C). This significantly impairs heat transfer, making it difficult to achieve the desired set points. Over time, ice accumulation increases, further worsening the situation.

Control Mechanisms and Current Limits

To detect and counteract ice formation, control systems are currently used that adjust the thermostatic valve to try to maintain the temperature set point. However, this method can trigger a vicious cycle: opening the nozzle to increase refrigerant flow rate can accelerate ice formation if the phenomenon has already begun. Usually, at predetermined time intervals, or through efficiency measurement algorithms, a defrost is initiated, but this approach can be inefficient. If the defrost is not timely, performance decreases, and if it occurs too early, electricity consumption increases.

Air exchanger fans, often controlled only in the On/Off state, could benefit from a more refined control strategy. Measuring variables such as back-emf and current drawn by the motor could allow detection of ice buildup in real time, thus optimizing defrost times and reducing consumption.

Machine Learning as a Solution

Current technology based on standard PLCs and static algorithms is not dynamic enough to adapt to evolving situations and sudden changes, such as a significant change in humidity in a refrigerated environment. However, with the use of edge devices such as the W6, advanced IIoT sensors and cloud-based data analysis, a more effective control strategy can be implemented.

Through artificial intelligence and machine learning, huge amounts of seemingly disconnected data can be analyzed and unexpected correlations discovered. This approach makes it possible to develop new management strategies based on continuously changing operating conditions. Integrating this data with machine learning systems enables experts to design innovative solutions that optimize efficiency and reduce the need for maintenance interventions.

The application of advanced technologies such as machine learning in refrigeration system management represents a revolution in the industry. By analyzing and correlating data, energy efficiency can be significantly improved and operating costs reduced. The adoption of these solutions not only improves performance, but also contributes to sustainable resource management, offering a greener and smarter future for the refrigeration industry.