Assessment of different data-driven algorithms for ahu energy consumption predictions

including AutoRegressive with eXternal inputs (ARX), State Space (SS), Subspace state space (N4S) and Bayesian Network (BN) are evaluated and compared using a case study of predictions of Air Handler Unit (AHU) thermal energy consumption. Training and testing data are generated from a dynamic Modelica-based AHU model. Four evaluation metrics of Root Mean Squared Error (RMSE), coefficient of determination (R2), Normalized Mean Bias Error (NMBE) and Coefficient of Variation of the Root Mean Square Error (CVRMSE) are used to compare the model prediction performance of different algorithms. The best algorithm is selected and proposed following the criteria recommonded by ASHRAE Guideline 14. Using the proposed data driven algorithm, the relation of AHU energy consumption with mixed air temperature, air flow rate, and supply water temperature are obtained. In the future, such correlations will be employed for an optimization analysis of AHU energy consumption.