Machine learning models for the prediction of polychlorinated biphenyls and asbestos materials in buildings - Hazardous material prediction approach for pre-demolition audits

Aug 5, 2026

Upload by Miles

Hazardous materials in buildings introduce uncertainty in renovation and demolition schedules and costs while also raising health concerns for workers. This study develops data-driven statistics and machine learning pipelines to identify patterns and the extent of polychlorinated biphenyls (PCBs) and asbestos materials in the Swedish building stock, supporting pre-demolition audits. Models are trained for four PCB and twelve asbestos components using environmental inventories, achieving strong predictive performance in residential and non-residential buildings. Geospatial prediction is demonstrated for Stockholm public housing using regional building registers, enabling cost-efficient evaluation of hazardous materials for non-inspected stocks.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Introduction

Uncertainty and limitations of pre-demolition audits

Background: PCBs and asbestos in buildings

Motivation for scalable machine learning approaches

Contributions and study aim

Resources, Conservation & Recycling 199 (2023) 107253 Contents lists available at ScienceDirect Resources, Conservation & Recycling journal [homepage:](homepage: www.elsevier.com/locate/resconrec)[ www.elsevier.com/locate/resconrec](homepage: www.elsevier.com/locate/resconrec) | Machine learning models for the prediction of polychlorinated biphenyls and asbestos materials in buildings
Pei-Yu Wu *, a, b, Claes Sandelsa, Tim Johansson a, Mikael Mangolda, Kristina Mj¨ornell a, b
a RISE Research Institutes of Sweden, 412 58 Gothenburg, Sweden
b Department of Building and Environmental Technology, Faculty of Engineering, Lund University, 221 00 Lund, Sweden | | | | --- | --- | --- | | A R T I C L E I N F O | | A B S T R A C T | | Keywords:
Hazardous material Pre-demolition audits Machine learning Prediction Probability distribution Building stock | | Hazardous materials in buildings cause project uncertainty concerning schedule and cost estimation, and hinder material recovery in renovation and demolition. The study aims to identify patterns and extent of polychlorinated biphenyls (PCBs) and asbestos materials in the Swedish building stock to assess their potential presence in pre-demolition audits. Statistics and machine learning pipelines were generated for four PCB and twelve asbestos components based on environmental inventories. The models succeeded in predicting most hazardous materials in residential buildings with a minimum average performance of 0.79, and 0.78 for some hazardous components in non-residential buildings. By employing the leader models to regional building registers, the probability of hazardous materials was estimated for non-inspected building stocks. The geospatial distribution of buildings prone to contamination was further predicted for Stockholm public housing to demonstrate the models’ application. The research outcomes contribute to a cost-efficient data-driven approach to evaluating comprehensive hazardous materials in existing buildings. | 1. Introduction The potential presence of hazardous materials creates uncertainty for time and cost estimates in renovation and demolition projects as well as health concerns for workers (Powell et al., 2015; Raˇskovi´c et al., 2020). A considerable quantity and variety of contaminants remain in the built environment nowadays, and their extent and location in buildings are rarely known due to insufficient documentation and verification (Bergsdal et al., 2014; Franzblau et al., 2020; Govorko et al., 2017; Wilk et al., 2019). As such, pre-demolition audit practice is introduced to identify the presence and quantify the amounts of hazardous materials to guide demolition contractors and waste handling companies in evaluating the contamination risk in material sorting (ECORYS, 2016; Wahlstr¨om et al., 2019b, 2019a). The audit inventories are also crucial for implementing selective demolition and deconstruction, as well as for quality assessment of reclaimed materials as a foundation for designing closed-loop circular strategies (Bergmans et al., 2017; Wahlstr¨om et al., 2020, 2019a). Nevertheless, the current time-consuming and costly practice of identifying hazardous materials on a building basis can hardly be used to estimate the remaining hazardous materials for the entire existing building stock. In addition, the quality and the completeness of inventory vary significantly between building types and regions, depending on the experience of auditors and building complexity (Wu et al., 2021a, 2021b). To prepare for and improve in-situ inspection of material quality appraisal and prevent second contamination in material recovery, exploring new approaches for enhancing the efficiency of in-situ hazardous material screening is necessary. Since the 1990s, polychlorinated biphenyls (PCBs) have been used as impregnation agents in building components for enhancing fire resistance and electrical insulation, while asbestos was commonly used for s

FAQs

Why are hazardous materials like PCBs and asbestos a problem in renovation and demolition projects?

Their possible presence creates uncertainty in project time and cost estimates and can create health risks for workers during work and waste handling.

How does the study evaluate the presence of PCBs and asbestos materials?

It generates statistical and machine learning pipelines trained on environmental inventories to predict the extent of four PCB and twelve asbestos components for inspected building stocks.

What practical application do the models enable beyond inspected buildings?

By using leader models with regional building registers, the probability and geospatial distribution of hazardous materials are estimated for building stocks not directly inspected, demonstrated for Stockholm public housing.

Aug 5, 2026

Upload by Miles