Conference Paper

Proceedings of Building Simulation 2019: 16th Conference of IBPSA

Using Python to Automate the Preparation and Execution of Thousands of Daylighting and Glare Simulations on a Cloud Parallel Computing environment for Time-efficient Simulations

Rania Labib ^1,2, Juan-Carlos Baltazar ^1
1 Texas A&M University, United States of America
² Prairie View A&M University

DOI: https://doi.org/10.26868/25222708.2019.211187
Abstract: In efforts to reduce energy consumption in buildings, building performance simulations have often been used to support energy-efficient design. Evaluating the performance and energy efficiency of multiple design scenarios can be a time-consuming process due to the large number of simulations that are required to predict the building’s performance in each scenario. To overcome these challenges, this study introduces a Python-based framework for automating the preparation and execution of a large number of performance simulations in a cloudbased high-performance computing environment, which contains numerous nodes that can be used to run tasks simultaneously. The proposed framework was successfully applied to simulate hourly daylighting and glare simulations of 144 scenarios in just a few hours. The processing time was reduced over one thousand times. The proposed framework not only speeds up the simulations but also addresses other challenges, such as storage space limitations as well as results analytics and visualization.
Keywords: Python, Daylighting Simulations, Glare Simulations, Cloud computing, Parallel Computing
Pages: 1545 - 1551
Paper: