ANN-Based Prediction and Optimization of Cooling System in Hotel Rooms

Abstract

This study aimed at developing an artificial-neural-network (ANN)-based model that can calculate the required time for restoring the current indoor temperature during the setback period in accommodation buildings to the normal set-point temperature in the cooling season. By applying the calculated time in the control logic, the operation of the cooling system can be predetermined to condition the indoor temperature comfortably in a more energy-efficient manner. Three major steps employing the numerical computer simulation method were conducted for developing an ANN model and testing its prediction performance. In the development process, the initial ANN model was determined to have input neurons that had a significant statistical relationship with the output neuron. In addition, the structure of the ANN model and learning methods were optimized through the parametrical analysis of the prediction performance. Finally, through the performance tests in terms of prediction accuracy, the optimized ANN model presented a lower mean biased error (MBE) rate between the simulation and prediction results under generally accepted levels. Thus, the developed ANN model was proven to have the potential to be applied to thermal control logic.

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

“…The initial model for predicting ΔTIME SBT is shown in For the training of the ANN model, a 0.0 minute goal, 1,000 times epoch, 0.6 learning rate, and 0.2 moment were applied, as these were found to be the optimal values in previous studies [26,27]. In total, 81 training data sets were prepared based on Equation (1) [28].…”

“…The input values for each neuron were normalized to be between 0 and 1. The normalized values were represented as +10…+30, -10…+10, -20…+40, -10…+10, and 0…+10°C for TEMP IN , ΔTEMP IN , TEMP OUT , ΔTEMP OUT , and ΔTEMP DIF , respectively.The number of hidden layers was initially designated as 3 and the number of hidden neurons was designated as 4, since these were determined to be the optimal values in previous studies[26,27]. The tangent-sigmoid function was employed as a transfer function for the hidden neurons.…”

Overview of the use of artificial neural networks for energy‐related applications in the building sector

“…They managed to reduce the energy consumption by 14% compare with a conventional control strategy employed in a building simulated using TRaNsient SYstems Simulation Program (TRNSYS). An equivalent method has been employed for setback moment determination of cooling system in other existing studies . Similarly, Macarulla et al tested a control strategy to determine the optimum starting time of an office building boiler to achieve thermal comfort at the beginning of each working day.…”

Hygrothermal Dynamic and Mould Growth Risk Predictions for Concrete Tiles by Using Least Squares Support Vector Machines

“…However, an alternative way to evaluate the influences of coupled heat and moisture transfer in building can be performed by adopting computational intelligence and machine learning techniques [14]. Moreover, this type of technology can be also used in the analysis of building energy demand and energy savings [15][16][17].…”