An automated optimization method for calibrating building energy simulation models with measured data: Orientation and a case study. (1st October 2016)

• Record Type: Journal Article
• Authors: Yang, Tao; Pan, Yiqun; Mao, Jiachen; Wang, Yonglong; Huang, Zhizhong
• Is Part Of: Applied energy. Volume 179(2016)
• Page Start: 1220
• Page End: 1231
• Publication Date: 2016-10-01
• Subjects: Building energy simulation -- Automated calibration -- Sensitivity analysis -- Optimization -- Particle Swarm Optimization -- Sub-metering; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2016.07.084
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: A complete and inclusive optimization automated calibration flow is developed. Sensitivity analysis is applied to determine the target tuned parameters. PSO is adopted and compiled to perform the optimization procedure. Sub-metered energy use are simultaneously calibrated through a weighted function. A case study in Shanghai based on sub-metered energy data is presented. Abstract: Due to the discrepancy between simulated energy consumption and measured data, it is essential to calibrate building energy models to improve its fidelity in evaluating the performance of retrofitting. Currently, most calibration methods are conducted manually to minimize this discrepancy, heavily relying on the knowledge and experience of analysts to discover a reasonable set of parameters. Because of the myriad independent and interdependent variables involved, the reliability of the entire simulation is largely undermined. In the presented paper, we propose a complete and fluent optimization automated calibration flow by introducing the mathematical optimization method (Particle Swarm Optimization is adopted) into the building energy model calibration process, thus leveraging the advantages of the efficiency and flexibility of the automated computer procedure. This approach is also characterized by its inclusivity, for it is compatible with other advanced manual methods and able to largely assist the experts in improving the efficiency of tuning relative input parameters. Moreover, a case in Shanghai is presented to verify the validity of the proposed method. After calibration, the simulation model demonstrates a satisfactory predicting accuracy. The calculated electricity consumption from the HVAC, lighting and equipment matches the actual monitored data with 11.6%, 7.3% and 7.2% CV (RMSE), respectively, and the total electricity consumption is within 6.1%.