| Summary: | Energy conservation measures (ECM) through awareness, good housekeeping measures and demand side management are the first step that must be implemented to achieve energy savings in any building. The energy savings from the implemented ECM must be properly quantified to demonstrate its effectiveness and success. Quantifying energy savings using Multiple Linear Regression (MLR) model have been gaining popularity due to its simple approach but its inability to model and predict dynamic behavior that involves non-linearity relationships can be seen as its main drawback. Thus, the objective of this work is twofold, where the first fold is to implement an ECM program in a building and the second fold is to determine the energy saved from the ECM program using Non-Linear autoregressive with Exogenous inputs artificial neural network (NARX-ANN). The ECM program was executed at one floor of a university building. NARX-ANN was used to model the baseline energy consumption and determine the energy savings following the International Performance Measurement and Verification Protocol (IPMVP). A comparison was made between the MLR and the NARX-ANN model where the latter provides higher accuracy in statistical error compared to the MLR model. The energy savings from the awareness program determined by the NARX-ANN model was 170.84 kWh with Mean Square Error and Root Mean Square accuracy of 0.6068 and 0.7790 respectively. © 2020 Elsevier B.V.
|
|---|