Evaluating Construction Equipment Accident Risk by Analyzing Utilization and Costs Using Regression Models

Construction vehicles and equipment are a vital resource for all construction projects, with its demand expected to increase alongside technological advancements. While the use of such equipment reduces manual labor, it also introduces new risks, potentially leading to accidents. This study quantitatively analyzes the likelihood of accidents by examining utilization rate, subcontractor types, and construction costs. A regression-based prediction model for accidents involving construction equipment is proposed, utilizing data augmentation techniques with multivariate normal and Poisson distributions to improve prediction accuracy. The study is structured around three main steps: (i) Data collection and classification, (ii) calculation of hourly operating costs (HOC) and construction costs, and (iii) data augmentation and regression analysis. Regression analysis showed high R² values exceeding 0.6 for seven types of equipment, with loaders, bulldozers, and air compressors as exceptions. Although dump trucks had the highest frequency of fatalities, the prediction model identified excavators as having the highest predicted fatality count in the case study. The proposed model emphasizes safety management by categorizing risk groups based on operating costs and construction costs. It also offers a practical process for field application, providing a valuable tool for developing regulations and making investment decisions related to safety management in construction equipment.