Evaluating Energy Efficiency of Buildings using Artificial Neural Networks and K-means Clustering Techniques

The consumption of energy in buildings has risen abruptly over the last decades. Due to less energy-efficient buildings, most of the energy is being thrown in our surroundings thus making an adverse effect on our environment. In this paper, heating and cooling loads of private or non-commercial buildings are covered. By implementing the proposed technique, which is a blend of cluster analysis and Artificial Neural Network (ANN), evaluation and prediction are performed. The estimation of heating and cooling loads of private or non-commercial buildings are performed using eight input variables in the ANN-based model. The details of variables are as follows, a relative surface area, total height, compactness, roof area, glazing area distribution, orientation, glazing area, and wall area. K-means clustering methodology is then used to cluster buildings on the basis of output variables. Stand on simulated literature data, evaluation of 768 different private or non-commercial buildings is done using the above-suggested method. Research results depicted that depending upon input variables, the above-suggested approach can efficiently evaluate heating and cooling load that is very much close to real test results.