[WITHDRAWN] The Role of Machine Learning in Optimizing the Use of Low-GWP Refrigerants in HVAC Systems

HVAC industry is among the largest energy-consumers and amount of emissions to the atmosphere worldwide. The traditional refrigerants like hydrofluorocarbons (HFCs) have a high global warming potential (GWP) and so the whole world has shifted to environmentsally benign low-global warming potential refrigerants like hydrofluoroolefins (HFOS), natural refrigerants (CO 2, ammonia, hydrocarbons as well as mixtures). Introduction of these refrigerants, however, poses new challenges on the aspects of performance of the systems, safety, optimization of efficiency and reliability in operations. Artificial intelligence (AI) and machine learning (ML) are two powerful technologies in the recent-years that have arisen to enhance the intelligence of HVAC systems to ensure predictive maintenance, adaptive control, fault detection, and energy optimization. The paper explores machine learning applications in HVAC optimization to use low-GWP refrigerants. In the study, the literature survey is thoroughly conducted, technical issues are defined, and an intelligent framework of how to monitor, optimize the performance, and achieve energy-efficient control is proposed based on the extremely intelligent data. It is suggested to use a hybrid ML method to optimize the charge of refrigerants, compressor speed, evaporator temperature, and airflow rate in real-time through a combination of a deep learning, reinforcement learning, and ensemble regression model. Simulation and experimental findings prove the great enhancement of coefficient of performance (COP), the energy consumption decrease and the increase in the accuracy of fault detection. The results affirm that machine learning is a disruptive facilitator of sustainable HVAC system and will be strategically instrumental in hastening the world to climate-friendly refrigeration solution.