Probabilistic Feature-based Grading and Classification System for End-of-Life Building Components Toward Circular Economy Loop

Abstract

The transition toward a circular construction economy requires robust, data-driven decision frameworks to determine whether end-of-life components can be reused. This paper proposes an adaptive probabilistic framework, the Multi-Level Grading and Classification System (MGCS), which evaluates specific component types, such as precast concrete wall panels, against the performance requirements of their intended reuse scenario. The proposed MGCS integrates Bayesian modelling with scenario-dependent performance thresholds. In the first stage, the method assigns a quality grade (A–E) based on quantitative physical condition indicators. The grade then informs a corresponding intervention class (1–5), supporting strategic reuse, upcycling, or downcycling decisions. This two-step, scenario-sensitive process enables a single component to be valued across multiple potential afterlives, maximising resource efficiency and embodied-carbon retention. The framework is demonstrated using empirical data from precast concrete panels and designed to remain transparent, auditable, and extensible to other component categories. By aligning material condition with functional demand, MGCS addresses the environmental, economic, and operational challenges of end-of-life management, reducing waste, improving value recovery, and supporting workflow optimisation within circular construction practice.