Prediction and characteristics analysis of building heating and cooling loads at planning phase

In order to solve the problem of heating and cooling loads prediction caused by the uncertainty of single building parameters in regional building energy planning, a Monte Carlo simulation method based on Latin hypercube sampling was proposed.Firstly, the uncertain parameters and their distribution were determined, then the uncertain parameters were sampled by Latin hypercube sampling and models were generated automatically by R programming language. Finally, these models were imported into the EnergyPlus software to calculate the cooling and heating load. Meanwhile, taking a planning land in Tianjin as an example, 2 000 groups of samples were taken and predicted by the proposed method, and the uncertainty and sensitivity of the results were analyzed. The results show that the Monte Carlo simulation method based on Latin hypercube sampling realizes the fast convergence of sampling and the fast generation of model. The frequency distribution,cumulative probability and eigenvalue of peak cooling and heating load of regional buildings can be effectively calculated by this method. Six parameters of the model, including solar heat gain coefficient, external window heat transfer coefficient, number of floors, building aspect ratio, south window to wall ratio, and north window to wall ratio, have significant influence on the building cooling and heating loads. © 2020, Central South University Press. All right reserved.