Predicting Construction Crew Productivity for Concrete-Pouring Operations

Abstract

In the construction industry, laborers generally work in a crew, and if they perform poorly, it significantly impacts on overall construction productivity. The construction crew productivity (CCP) is prone to different factors, some of which are within the engineer's control and others are not. Due to this, it is an arduous and challenging task to establish productivity claims based on the CCP. Therefore, this study aims to evaluate CCP and, based on it, signify and defend the loss of productivity claims. To evaluate CCP, a feed-forward back-propagation artificial neural network (ANN) approach is utilized. A total of 14 factors influencing the CCP are considered inputs, while CCP is considered output in the ANN model. Further, an explicit expression is derived from the final weights and biases of the trained ANN. The performance of the model is checked by statistical parameters such as mean square error (MSE), root mean square error (RMSE), average absolute deviation (AAD), square of correlation coefficient (R2), and coefficient of variation (COV). Then, to implement for practical purposes, the proposed ANN model is deployed on a project in India and found satisfactory performance. Further, the sensitivity analysis extracts the influence rate of each factor on the CCP and finds the top three significant factors: crew size, working hours, and temperature. Thus, the proposed study helps to estimate and provide caveats in the context of claims when there are losses through CCP. Herein, the presented methodology is applied to the concrete-pouring operation of reinforced concrete (RC) columns. However, it can be extended to other RC structural members like slabs, beams, foundations, etc.