Application of a multiple linear regression and an artificial neural network model for the heating performance analysis and hourly prediction of a large-scale ground source heat pump system

• Published in: Energy and buildings 2018-04, Vol.165, p.206-215
• Main Authors: Park, Sang Ku, Moon, Hyeun Jun, Min, Kyung Chon, Hwang, Changha, Kim, Suduk
• Format: Article
• Language: English
• Subjects: Artificial neural networks; Coefficient of variation; Data processing; Energy conservation; Energy efficiency; GSHP system performance; Heat exchangers; Heat pumps; Heat transfer; Heating; Hospitals; In situ monitoring data; Influencing factors on performance; Mathematical models; Model accuracy; Monitoring; Neural networks; Performance prediction; Prediction model by a multiple linear regression; Prediction model by an artificial neural network; Prediction models; Regression analysis; Statistical analysis; Studies; Water temperature
• ISSN: 0378-7788; 1872-6178
• DOI: 10.1016/j.enbuild.2018.01.029

•The heating performance and operating properties of GSHP system was analyzed using in situ monitoring data.•Hourly heating performance prediction models for GSHP system were developed.•The influencing factors of GSHP system performance were quantitatively analyzed by an elaborate MLR model.•Prediction powers of neural network and multiple linear regression models were compared•The prediction models can be utilized with enough good accuracy as a baseline for M&V. A ground source heat pump system (GSHP) with 450 RT capacity composed of ten heat pump units provides the heating and cooling energy to an entire hospital building. The seasonal heating performance of 3.21 and system operation properties of the system were analyzed using in situ monitoring data from Nov. 2016 to Mar. 2017. On this basis, hourly GSHP system performance prediction models applying a multiple linear regression (MLR) and an artificial neural network (ANN) were developed. The quantitative effects of influencing variables on the system performance, including the entering source and load water temperatures (EST, ELT) were analyzed by elaborated MLR model with statistical significance. The prediction accuracy was 3.56% by the MLR, and 1.75% by the ANN, based on the coefficient of variation of root mean squared error (CVRMSE) without overall bias. These prediction models can be used as a baseline for the measurement and verification (M&V) of possible future energy conservation measures and real-time performance monitoring to check malfunction of the system.