This repository presents the proposed framework that automate the process of mapping BMS points to EnergyPlus variables. The framework was demonstrated using four case studies. Please refer to the notebook of Building A for the experiment procedure with thorough comments and the other three files for more results.
More details please refer to our paper "Automated recognition and mapping of building management system (BMS) data points for building energy modeling (BEM)".
The framework consists of three main parts: core information extraction, tokenization and fuzzy string matching. By using unsupervised learning for core information extraction and a universal dictionary for tokenization, the framework is applicable to different building cases.