Predicting Waste Generation During the Construction of a Secondary School: A Case Study

This case study research provides unique insight into mixed construction and demolition waste generated through the construction of a secondary school project and, moreover, provides actual site data that is rare within published literature. A deterministic prediction model is subsequently developed to accurately forecast the cumulative waste generated to enable project teams to better forecast, manage and reduce site waste generated. A combination of pragmatism and interpretivism is utilised as the overarching epistemology, and deterministic bivariate models are adopted to predict waste production over time. An embedded mixed-method case study is then undertaken within the case study organisation. Model performance is assessed using Mean Absolute Percentage Error (MAPE) as a metric that characterises forecasting accuracy. A focus group consisting of a cross-section of the project team is convened to share data analysis results and accrue reasons for the waste trends observed and how these can be reduced moving forward. Emergent research results identify peaks of waste generation caused by upfront packaging brought to site and unforeseen ground conditions that increased the project’s expected waste produced. Secondary waste data collected from the case study project was used to generate a polynomial regression model which accurately predicted (producing a MAPE of 13.1%) waste year-on-year for the project over a 27-month period using only a one-year data set. The research also delves into the concept of circular economy (CE) and the theory of reduce, reuse and recycle and illustrates how packaging is a major cause of waste on a modern construction site. This novel research provides a useful future predictive model for waste generation, providing insight to project stakeholders who seek to reduce the waste produced during the construction of a secondary school project. Modelling waste generation rates accurately over repeated projects could allow benchmark measures to be developed and implemented to guide practitioners on whether excessive waste is being generated and what controls could be implemented to reduce it. Case study findings presented also provide a unique insight into real-world waste generation data.