A hybrid Genetic Algorithm and Monte Carlo simulation approach to predict hourly energy consumption and generation by a cluster of Net Zero Energy Buildings

摘要

Employing a hybrid Genetic Algorithm (GA) and Monte Carlo (MC) simulation approach, energy consumption and renewable energy generation in a cluster of Net Zero Energy Buildings (NZEBs) was thoroughly investigated with hourly simulation. Moreover, the cumulative energy consumption and generation of the whole cluster and each individual building within the simulation space were accurately monitored and reported. The results indicate that the developed simulation algorithm is able to predict the total instantaneous and cumulative amount of energy taken from and supplied to the central energy grid over any time period. During the course of simulation, about 60-100% of total daily generated renewable energy was consumed by NZEBs and up to 40% of that was fed back into the central energy grid as surplus energy. The minimum grid dependency of the cluster was observed in June and July where 11.2% and 9.9% of the required electricity was supplied from the central energy grid, respectively. On the other hand, the NZEB cluster was strongly grid dependant in January and December by importing 70.7% and 76.1% of its required energy demand via the central energy grid, in the order given. Simulation results revealed that the cluster was 63.5% grid dependant on annual bases. In general, this stochastic algorithm is a self-learning one, i.e., at the end of each year, it utilizes the instantaneous energy consumption and generation data of each building to predict its energy balance in subsequent years. Hence, the accuracy and validity of the predictions increase over time. The simulation results are capable of modifying and readjusting the energy consumption patterns of buildings via appropriate predefined policies and well designed monitoring systems. (C) 2016 Elsevier Ltd. All rights reserved.