Abstract
Model-based predictive control (MPC) strategies for heating, ventilation, and air-conditioning (HVAC) systems present an opportunity to lower building energy consumption and operational costs. Such approaches rely on the development of a model to precisely forecast building thermal dynamics, such as room air temperature or heating/cooling rate, and make control-related decisions. The control-oriented modeling of building energy systems should be accurate in predicting indoor conditions and present low computational complexity. These features are the key challenge of implementing advanced control methods such as MPC. Extant studies on building modeling for MPC have focused on step-ahead forecasting techniques to forecast building thermal dynamics, while multistep-ahead forecasting is essential. Moreover, machine learning model suitable in case of the domain-based engineering expertise are also not available. To this aim, we perform a comparative analysis of the grey-box model based on a resistance-capacitance (RC) thermal network and a machine learning model composed of an artificial neural network (ANN) for multistep-ahead prediction of building thermal dynamics using current and historical data. Actual experimental data obtained from the Flexible Research Platform (FRP) in Oak Ridge National Laboratory (US) are used for estimation and validation purposes. The average root mean squared error (RMSE) of the grey-box and ANN models are 0.89 °C and 1.02 °C, respectively. The results indicate that the grey-box model outperforms the ANN model in the considered validation periods in terms of accuracy and prediction stability.
| Original language | English |
|---|---|
| Article number | 107115 |
| Journal | Engineering Applications of Artificial Intelligence |
| Volume | 126 |
| DOIs | |
| State | Published - Nov 2023 |
Funding
Although AI emerged as a powerful tool in various diverse fields (Akkem et al., 2023; Čartolovni et al., 2022; Iyer, 2021). However, some knowledge gaps limit the applicability of AI-based models in building energy management applications (Khalil et al., 2022). Several papers on building performance have been published focusing on modeling and control. The applicability of AI in conjunction with MPC in building control and operation was discussed by (Afram et al., 2017), and the role of AI in energy management was discussed in (Nutakki and Mandava, 2023). Zhang et al. (2015) assessed several inverse modeling approaches to estimate HVAC hot water usage, including change point regression, Gaussian regression, Gaussian mixture regression, and artificial neural network (ANN). The Gaussian mixture regression model performed slightly better among these models in terms of root mean squared error (RMSE). Feng and O'Neill (2020) presented an ARMAX-based control-oriented model for control application. The heating, cooling, and electrical demands of five distinct buildings were examined using two alternative black-box models, namely linear regression and ANN (Gunay et al., 2017). The results obtained after assessing these buildings' performances by using different forms and various parameter values indicated that the ANN model outperformed linear regression in terms of RMSE. Dong et al. (2005) employed a support vector machine (SVM) to evaluate the energy usage of commercial buildings in Singapore. Weather data and monthly utility bills were obtained for model training and testing; the variance and error percentage coefficient was less than 3% and 4%, respectively. Cotrufo et al. (2020) compared the results of five different machine learning techniques, namely Gauss process regression (GPR), ANN, SVM, decision tree (DT), and random forest (RF). The GPR model was the most accurate in terms of RMSE, and it was further used in MPC to reduce natural gas consumption in institutional buildings. Reynolds et al. (2018) developed a zone-based ANN model to forecast heating energy demand and zone temperature. This model was combined with a genetic algorithm (GA) to lower energy costs and consumption. Ensemble models for predicting the next day's energy usage and peak demand were proposed using a data-mining-based approach (Fan et al., 2014). To map the relationship between input and output, eight fundamental models were used: multivariate adaptive regression splines (MARS), SVM, RF, boosting tree (BT), ARMAX, k-nearest neighbors (kNN), multilayer perceptron (MLP) and multiple linear regression (MLR). SVM and RF had the largest weights among the ensemble models and other models, whereas MLR and ARMAX did not perform well because of the nonlinear and complex process of the building, and therefore, their weights among those of the ensemble models were smaller. Several studies have applied machine learning for temperature and load forecasting and examples of these are summarized in Appendix (Table A.1.).This work was supported by the National Research Foundation of Korea (NRF) through a grant funded by the Korean government (MSIT) (No. 2021R1A4A1031705). This manuscript has been co-authored by UT-Battelle, LLC, under contract DE-AC05-00OR22725 with the US Department of Energy (DOE). The US government retains and the publisher, by accepting the article for publication, acknowledges that the US government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for US government purposes. DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan). This manuscript has been co-authored by UT-Battelle, LLC, under contract DE-AC05-00OR22725 with the US Department of Energy (DOE). The US government retains and the publisher, by accepting the article for publication, acknowledges that the US government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for US government purposes. DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan ( http://energy.gov/downloads/doe-public-access-plan ). This work was supported by the National Research Foundation of Korea ( NRF ) through a grant funded by the Korean government ( MSIT ) (No. 2021R1A4A1031705).
Keywords
- Artificial neural network
- Grey-box model
- Model-based predictive control
- Multi-step ahead prediction