This paper aims to provide a summarized classification of fault detection and diagnosis (FDD) methods in Heating Ventilation and Air Conditioning (HVAC) systems by dividing them into knowledge-driven-based, data-driven and hybrid approaches, and then subdividing each category to more detailed categories. Considering the advantages and disadvantages of each method, it is concluded that knowledge-driven approaches require noticeable expertise, high number of input variables and consequently sensors to be installed, also having scalability issues. On the other hand, data-driven methods provide more precise results, while they require reliable labeled fault free and/or faulty data which is hard to access especially in real-world Building Automation System (BAS) data. Considering the disadvantages of knowledge-based and data-driven approaches and following a brief explanation of current studies based on hybrid methods, this paper highlights the necessity of hybrid FDD approach expansion in the future studies specifically in fault diagnosis.
Overview on Fault Detection and Diagnosis Methods in Building HVAC Systems: Toward a Hybrid Approach / Piscitelli, Marco Savino; Hooman, Armin; Rosato, Antonio; Capozzoli, Alfonso. - 378:(2024), pp. 709-719. (Intervento presentato al convegno SEB 2023: Sustainability in energy and buildings 2023 tenutosi a Bari (Italy) nel 18-20 September 2023) [10.1007/978-981-99-8501-2_61].
Overview on Fault Detection and Diagnosis Methods in Building HVAC Systems: Toward a Hybrid Approach
Piscitelli, Marco Savino;Capozzoli, Alfonso
2024
Abstract
This paper aims to provide a summarized classification of fault detection and diagnosis (FDD) methods in Heating Ventilation and Air Conditioning (HVAC) systems by dividing them into knowledge-driven-based, data-driven and hybrid approaches, and then subdividing each category to more detailed categories. Considering the advantages and disadvantages of each method, it is concluded that knowledge-driven approaches require noticeable expertise, high number of input variables and consequently sensors to be installed, also having scalability issues. On the other hand, data-driven methods provide more precise results, while they require reliable labeled fault free and/or faulty data which is hard to access especially in real-world Building Automation System (BAS) data. Considering the disadvantages of knowledge-based and data-driven approaches and following a brief explanation of current studies based on hybrid methods, this paper highlights the necessity of hybrid FDD approach expansion in the future studies specifically in fault diagnosis.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/11583/2987473
Attenzione
Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo