Energy Analysis in Green Building via Machine Learning: A Case Study in a Hospital

Electricity consumption is increasing as a result of increasing people’s needs, such as lighting, heating, and comfort. Different needs come into play day by day in the houses where people live and in places used as common areas, and this increases the need for electricity. Studies have observed that almost half of the world’s electricity consumption is made by buildings. Public buildings, shopping malls, hospitals, and hotels are typical examples of such structures. However, hospitals have an important place among all building types as they contain a wide range of devices and are of critical importance to many systems. Consumption in hospitals is a necessity rather than a desire for comfort in places such as hotels and shopping malls. Therefore, analysis of the energy consumed by hospitals is one of the important things to perform to reduce the damage caused by electricity consumption to the environment. In this study, the energy analysis of a green hospital with an installed area of 55,000 square meters in Istanbul was conducted, and machine learning techniques can be used in the analysis. Among many methods used for building energy analysis, long short-term memory (LSTM) has been chosen. The available data set was analyzed with the various LSTM methods and classification and prediction operations were carried out. Error rates for each method were compared. With the results obtained, it has been observed that the vanilla LSTM method provides acceptable results in building energy analysis.