Enhancing and optimization in exisiting HVAC equipment using tools of industry 4.0

Abstract

[Abstract] New artificial intelligence tools, together with those of machine learning specific to the new smart industry, represent both a great opportunity and a great challenge for older technologies not designed with the level of sensorization required by these systems. Most HVAC systems installed in buildings worldwide are still effective, but they are also old and do not have the sensory system designed for the new IoT tools that would optimize their operation. In this study, a procedure is developed based on the combination of magic learning and artificial intelligence tools, through which variables of high variability can be controlled very effectively through the design of their avatar variables in the system. The procedure achieves high precision in the estimation of the water outlet temperature, with the accuracy of values ranging between 88.5 % and 97.2 %. This allows for the creation of variables that contribute to a more precise control of the system. In addition, the methodology significantly reduces false alarms (from 12.7 % to 0.19 % with the new variables), improving the reliability of HVAC systems. These results have been validated in a real installation, confirming their success rate in detecting both operational and maintenance alerts and also validating the implementation of these systems in installations prior to the Industry 4.0 era. This study provides the refrigeration sector with a highly effective and easily implementable procedure that is aligned with circular economy policies, adding new value to old but perfectly valid HVAC equipment, which is the majority of worldwide cases