Jobsite condition-based construction manual work modeling for activity duration estimation

Abstract

Estimating the duration of manual operations is a key step for construction project scheduling. However, due to uniqueness of a construction project, it is often difficult to accurately estimate an activity duration affected by various job conditions such as the number of workers, the type of tools and devices, and the quantity of materials. This paper thus presents a job-level simulation modeling approach that extracts site information from drawings (e.g., BIM) and estimates a motion time-based production rate of an individual worker, by which various operational scenarios can be tested to determine the optimal activity duration given the job condition. For the modeling, a case study is conducted for a concrete pouring activity of slabs in building construction. Worker motions involved in the processes of concrete pouring, vibrating, and flattening are modeled to determine typical motion ranges with the specific tool used, and the motion time required to complete the job is estimated by applying a Predetermined Motion Time System (PMTS) to the modeled motions. The resulting motion time per unit working area is eventually used to determine the motions required for specific jobs given the site conditions and estimate the total duration; such site information (e.g., working area and floor shape) are extracted from a drawing of the slab. For validation, the simulated result is compared with an actual activity duration collected in the case study, and scenarios using different numbers of workers and different tools (e.g., bull floats, hand floats, and both) are compared to evaluate the productivity improvements. The experimental results show that the proposed model can be used to accurately estimate the activity duration and perform optimal scheduling through simulation. This approach thus explores a systemic way to estimate activity durations of building components using only information available during the design phase (i.e., BIM model).