Machine Learning and Building Information Modeling for optimization Energy and acoustic performance

Costantino Carlo Mastino

;Raffaello Possidente;Andrea Frattolillo;Juozas Vaiciunas

2025-01-01

---

Abstract

The integration of Machine Learning (ML) with Building Information Modeling (BIM) is revolutionizing the optimization of energy and acoustic performance in buildings. ML algorithms analyze large data sets to identify patterns and predict energy behavior, optimizing HVAC system use, reducing energy consumption, and improving thermal comfort. Additionally, ML performs detailed acoustic simulations, evaluating noise impact and suggesting solutions for better acoustic insulation. Tools like Autodesk Revit, ArchlineXP, IES VE, and DesignBuilder integrate these technologies to provide advanced analyses and support informed design decisions. The combined use of ML and BIM creates more efficient, sustainable, and comfortable buildings. European regulations, such as Directive 2014/24/EU, encourage or mandate BIM use for public works and design competitions. The European Union Public Procurement Directive also promotes BIM in publicly funded projects. In Italy, Legislative Decree 36/2023 introduces innovations, including the digitization of processes through interoperable digital platforms (BIM) and digital construction information management tools. This work describes how ML analyzes BIM data, identifies patterns, and makes predictions to optimize HVAC use, improve energy efficiency, and reduce consumption, considering often underestimated acoustic aspects. ML enables detailed simulations based on measured data to evaluate energy and acoustic performance, using parameters from main reference standards, many of which are being digitized by various working groups. These procedures suggest solutions to improve thermal and acoustic insulation.