Measuring and Estimating Steel Drafting Productivity

This paper describes an engineering productivity measurement system and proposes a neural network modeling approach for estimating the engineering productivity. The methodology has been applied to steel drafting. Firstly, the research focused on measuring the work scope of steel drafting projects. A method of quantitatively measuring the work scope was developed based on installed quantities, which is the quantity of steel pieces in terms of their physical characteristics. With the developed consistent measurement standard, a neural network model for estimating drafting productivity was developed and implemented using influencing factors appropriate to project conditions. Historical data collected through an implemented data acquisition system in a steel fabrication company were prepared for neural network training and model validation. This predictive model streamlines and increases the accuracy of earlier estimating process, which was highly subjective.