Productivity Analysis of Earthmoving Operations

Heavy construction and mining general contractors record on a daily basis large amount of operational data. Nevertheless, this information is rarely used to enhance the knowledge and capabilities of the companies that spent great amount of money and resources recording it. This research presents different approaches on how to process this data to convert it in useful information. The prime goal of this analysis is to determine a suitable and convenient method to obtain and present historical productivities of key equipment, in order to provide a tool to aid estimating and generate reference information to support decision making.;Data mining, artificial neural networks and summarization tools proved to assist effectively in the assessment of historical productivities and in the identification of the attributes that most influence the results. Multiple ANN configurations were evaluated in the determination of hauling trucks and excavators productivities. Ward net architectures that include different activation functions applied to hidden layer slabs performed better than standard backpropagation nets since they are able to detect different features in a pattern processed through a network.;Estimating construction operation productivity is mostly experience-based due to the complexity involved. However, predominantly empirical practices do not secure a reliable estimate because of the absence of a binding mechanism that relates the present case to past patterns (Chao and Skibniewski 1994). This study involved the analysis of the historical productivity of more than 230 hauling units, 160 excavator units, and 150 units of support equipment. The historical data has been recorded for about three years and represent the operations of one of the largest contractors on the Alberta Oil Sands in eleven different projects.