Ventilation fans are an important component of any mechanically ventilated building.Poor fan performance could significantly affect the whole building performance metrics.There are several issues such as dirty blades,...Ventilation fans are an important component of any mechanically ventilated building.Poor fan performance could significantly affect the whole building performance metrics.There are several issues such as dirty blades,mechanical wear,aging of fans could impact the fan’s performance.In present work,a novel,indirect and data-driven methodology is introduced to monitor the ventilation fan unit performance.The proposed method is able to perform continuous monitoring of ventilation fan unit in real-time.The real-time performance of 3 Air handling unit(AHU)fans is examined in an academic building.Expected fan performance is modeled with the help of manufacturer data and compared against the real-time performance.Two data-driven models are developed and implemented.The first model is used to compute expected total fan pressure at a given airflow rate while second is a Support Vector Regression(SVR)model,to predict the fan efficiency.The performance monitoring of the ventilation fan unit is determined in terms of expected and actual fan energy consumption.Findings indicated a significant performance gap in three ventilation fan unit in a case building known as OU44,located in city Odense,Denmark.The advantage of this method comprises simplicity,no direct human intervention and scalability to the series of ventilation units.展开更多
文摘Ventilation fans are an important component of any mechanically ventilated building.Poor fan performance could significantly affect the whole building performance metrics.There are several issues such as dirty blades,mechanical wear,aging of fans could impact the fan’s performance.In present work,a novel,indirect and data-driven methodology is introduced to monitor the ventilation fan unit performance.The proposed method is able to perform continuous monitoring of ventilation fan unit in real-time.The real-time performance of 3 Air handling unit(AHU)fans is examined in an academic building.Expected fan performance is modeled with the help of manufacturer data and compared against the real-time performance.Two data-driven models are developed and implemented.The first model is used to compute expected total fan pressure at a given airflow rate while second is a Support Vector Regression(SVR)model,to predict the fan efficiency.The performance monitoring of the ventilation fan unit is determined in terms of expected and actual fan energy consumption.Findings indicated a significant performance gap in three ventilation fan unit in a case building known as OU44,located in city Odense,Denmark.The advantage of this method comprises simplicity,no direct human intervention and scalability to the series of ventilation units.
