Algorithm for optimal application of the setback moment in the heating season using an artificial neural network model

• Published in: Energy and buildings Vol. 127; pp. 859 - 869
• Main Authors: Moon, Jin Woo, Jung, Sung Kwon
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.09.2016
• Subjects: Artificial neural network; Heating energy; Optimal controls; Setback temperature; Thermal comfort; Thermal comfort; Artificial neural network; Optimal controls; Setback temperature; Heating energy
• ISSN: 0378-7788
• DOI: 10.1016/j.enbuild.2016.06.046

•Initial ANN model was developed for predicting the optimal setback application.•Initial model was optimized for producing accurate output.•Optimized model proved its prediction accuracy.•ANN-based algorithms were developed and tested their performance.•ANN-based algorithms presented superior thermal comfort or energy efficiency. The objective of this study was to develop an artificial neural network (ANN) model to predict the optimal start moment of the setback temperature during the normal occupied period of a building and to suggest an algorithm employing the developed ANN model to enhance indoor thermal comfort and building energy efficiency. To achieve this objective, three major steps were undertaken: the development of the initial ANN model, optimization of the initial model, and development of control algorithms and performance tests. The development and performance testing of the model and algorithm were conducted by employing numerical simulation methods using transient systems simulation (TRNSYS) and matrix laboratory (MATLAB) software. The results of the development and tests revealed that the indoor temperature, outdoor temperature, and temperature difference from the setback temperature were the three major variables predicting the optimal start moment of the setback temperature. Thus, these variables were used as input neurons in the ANN model. In addition, the optimal values for the number of hidden layers, number of hidden neurons, learning rate, and moment were found to be 4, 9, 0.6, and 0.9, respectively, and these values were applied to the optimized ANN model. Comparative performance testing of a conventional algorithm and two ANN-based predictive algorithms demonstrated that the ANN-based algorithms were superior in advancing indoor thermal comfort or building energy efficiency.