AI system developed to predict the energy rate of buildings

Computer scientists at Loughborough University have worked with engineering consultant Cundall to develop an artificial intelligence (AI) system that can quickly predict the energy rate of a building.

Image: Shutterstock

The building emission rate (BER) is an important factor in calculating the energy performance and efficiency of a building and is required to complete the Building Energy Performance Certificate (EPC).

The current method of generating a BER is generated by manually entering hundreds of variables. This can take hours to days to generate, depending on the complexity of the building.

A research team led by Dr. Georgia Cosma of the Loughborough Science Department and graduate student KareeM Ahmed now claims to have designed and trained an AI model that can predict BER values ​​in non-domestic buildings in just one second. 27 variables with little loss of accuracy.

Dr. Cosma described the study as “an important first step in using machine learning tools for energy prediction in the UK” and showed how the data “improve the current process of the construction industry.” ..

Created with support from CundallEdwin Wealend, Head of Research and Innovation, reportedly trained AI models using large amounts of data from the UK Government’s energy performance appraisal.

More energy and environmental news is here

Researchers use a “decision tree-based ensemble” machine algorithm and use 81,137 actual data records containing information on non-domestic buildings such as stores, offices and restaurants across the UK from 2019 to 2019. He said he built and verified the system. Information on building capacity, location, heating, cooling, lighting, activities, etc. was included.

“There are studies that apply machine learning to building energy projections, but these are limited and account for only 8% of the total building, while non-domestic buildings account for 20 of the UK’s total CO2 emissions. It accounts for%, “said Dr. Cosma.

Wealend added that the team ultimately wants to build on methods developed to predict actual operational energy consumption.

“By quickly and accurately predicting energy consumption and emissions of non-domestic buildings, we can reduce energy consumption and focus our energy on the more important task of reaching net zero.” He commented.

The results of the project survey were announced at Chartered Institute of Building Services Engineers (CIBSE) Technical Symposium 2021 The treatise will be published on the CIBSE website later this year. AI system developed to predict the energy rate of buildings

Back to top button