TECH TALK


Explaining Explainable AI


Klas Berglöf, head of R&D at ClimaCheck Sweden, discusses how Explainable AI (XAI) drives the change to predictive maintenance (PdM) in air conditioning, refrigeration, and heat pumps.


Most


failures and downtime are caused by a lengthy period of poor operation detectable by data-driven PdM long before it causes high energy bills and failures.


paradigm shift is approaching the RACHP industry. Digitalisation has resulted in sensors that are often factory-installed with systems connected to Building Management Systems and cloud platforms. IoT is entering the RACHP industry, which will change our industry. However, commissioning and maintenance practices have not yet been upgraded to take advantage of the data. The RACHP industry uses 20% of the global electrical consumption, and a 25% saving potential in existing equipment is achievable at a low cost. Poor operating effi ciency is caused by a lack of focus on commissioning at varying operating conditions and a wide range of undetected faults. Prevailing commissioning and maintenance practices do not


A


ensure effi cient and reliable systems. A ‘business as usual’ approach can lack performance


verifi cation and optimal control during regular operation; instead, it concentrates on inspection at handover, warranties, and ticking boxes for delivered goods. Product standards are expected to deliver effi cient operation but often fail due to the lack of performance monitoring and understanding of the diff erence between static design and dynamic operation. Awareness among equipment owners is increasing,


and regulations are focusing more on effi ciency and benchmarking. The increasing focus on sustainability drives a paradigm shift made possible through digitalisation and new analytical methods. RACHP system maintenance rarely uses installed sensors


to reduce maintenance and energy costs. Thermodynamic analyses combined with Digital Twins drastically enhance the capabilities of Automated Fault Detection and Diagnosis (AFDD).


The Internal Method, based on thermodynamics, has been


proven for more than 35 years and identifi es any changes in performance on the component level. Required sensors are often installed from the factory, and if any sensors are missing, the cost of adding them is low compared to local troubleshooting, cost of energy, and failures. As the RACHP sector uses 20% of the global electricity,


the pressure is increasing to reduce energy consumption. There is a 10-25% saving potential based on ClimaCheck’s


28 April 2025 • www.acr-news.com Figure 1 Download the ACR News app today


experience from thousands of sites and many other reports. These savings are ‘low-hanging fruit’ for property owners. Poor operating effi ciency results from poor commissioning, lack of optimisation to varying operating conditions and a wide range of undetected faults. New practices are required to ensure effi cient operation in dynamic applications and to detect degradation before failures occur. Current maintenance practices, with scheduled visits, do not reduce energy waste and failures in an acceptable way. A paradigm shift is rapidly approaching the RACHP industry


as regulations and customer awareness drive the transition to monitoring-based predictive maintenance (PdM) and optimisation.


Performance analytics


1.1 Thermodynamic analyses make information out of data. Figure 1 shows the Pressure Enthalpy chart for a standard refrigeration process with the three enthalpy points required to defi ne COP, isentropic compressor effi ciency and a fl ow chart with required sensors marked. Electrical power is measured to establish cooling and heating capacity. The method of defi ning performance based on thermodynamics is well-proven theoretically and through hundreds of thousands of tests in developing and producing OEM test rigs. 1.2 Key Performance Indicators KPIs for real-life systems. Product ratings based on COP, EER, SCOP, kW/RT, and weighted seasonal values contribute towards pushing the effi ciency of products but not in the same way as operational


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40