Estimation and optimization of heating energy demand of a mountain shelter by soft computing techniques

In this study, the heating energy demand of a shelter located in mountainous areas (Damavand mountain) in Iran was simulated, estimated, and then optimized. The main focus was to improve the heating energy demand of existing shelter by different protective zones and heat body of occupants. In order to estimate the heating energy demand in the shelter, an expert artificial neural network (ANN) trained by back-propagation algorithm (BP) was developed in Neural Network Toolbox in MATLAB R2016a. Additionally, the performance of the ANN-BP models was enhanced by two well-known meta heuristic algorithms: particle swarm optimization (PSO) and gray wolf optimization (GWO) methods. Lastly, this research used two different optimizer engines, i.e., Galapagos and Silvereye, to minimize the heating energy demand in the proposed models. The results of soft computing techniques indicated that the performance indices of ANN-GWO model were the best compared to those of ANN-PSO and ANN-BP. Interestingly, the proposed optimization methods by Galapagos and Silvereye plug-ins showed very satisfactory results in decreasing heating energy demand. Finally, based on the local materials of the site, the designed proposed model was presented. This new shelter can allow more occupants to climb the Damavand mountain in better conditions.