Automatic generation of energy conservation measures in buildings using genetic algorithms

Building energy simulations are key to studying energy efficiency in buildings. The state-of-the art building energy simulation tools requires a high level of multi disciplinary domain expertise from the user and many technical data inputs that curb the usability of such programs. In this paper an IT tool is presented, which has the capability of predicting a building's energy utilization configuration based on the reported annual energy and a few non-technical inputs from the user; and correspondingly generates cost effective energy conservation measures for the intended savings.The approach first identifies the system variables that are critical to a building's energy consumption and searches for the combination of these parameters that would give rise to the annual energy consumption as reported by the facility. Genetic algorithms are utilized to generate this database. A statistical fit is formulated between the system variables and the annual energy consumption from the database. Using this correlation, system configuration for the target energy efficiency is determined with corresponding energy conservation measures. A cost analysis is carried out to prescribe the most cost effective energy conservation measures. Competency of the tool is demonstrated in the paper through case studies on three geographies with different climate conditions.

• Building system variables critical to the energy consumptions are identified. • Building system is predicted from its energy use and non-technical physical data. • Genetic algorithms are utilized in this approach. • The approach does not need human intervention to physically analyze building systems. • Competency of the tool is demonstrated through three case studies.