Durham School of Architectural Engineering and Construction

 

First Advisor

Philip Barutha

Second Advisor

KyungKi Kim

Third Advisor

Jennifer Lather

Date of this Version

Spring 4-27-2022

Citation

Ding, SJ. (2022). Using Cost Simulation and Computer Vision to Inform Probabilistic Cost Estimates. M.S. thesis, University of Nebraska-Lincoln, Nebraska.

Comments

A THESIS Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Master of Science, Major: Construction Engineering and Management, Under the Supervision of Professor Philip Barutha. Lincoln, Nebraska: April 2022

Copyright © 2022 Shu Jing Ding

Abstract

Cost estimating is a critical task in the construction process. Building cost estimates using historical data from previously performed projects have long been recognized as one of the better methods to generate precise construction bids. However, the cost and productivity data are typically gathered at the summary level for cost-control purposes. The possible ranges of production rates and costs associated with the construction activities lack accuracy and comprehensiveness. In turn, the robustness of cost estimates is minimal. Thus, this study proposes exploring a range of cost and productivity data to better inform potential outcomes of cost estimates by using probabilistic cost simulation and computer vision techniques for activity production rate analysis.

Chapter two employed the Monte Carlo Simulation approach to computing a range of cost outcomes to find the optimal construction methods for large-scale concrete construction. The probabilistic cost simulation approach helps the decision-makers better understand the probable cost consequences of different construction methods and to make more informed decisions based on the project characteristics.

Chapter three experimented with the computer vision-based skeletal pose estimation model and recurrent neural network to recognize human activities. The activity recognition algorithm was employed to help interpret the construction activities into productivity information for automated labor productivity tracking.

Chapter four implemented computer vision-based object detection and object tracking algorithms to automatically track the construction productivity data. The productivity data collected was used to inform the probabilistic cost estimates. The Monte Carlo Simulation was adopted to explore potential cost outcomes and sensitive cost factors in the overall construction project. The study demonstrated how the computer vision techniques and probabilistic cost simulation optimize the reliability of the cost estimates to support construction decision-making.

Advisor: Philip Barutha

Share

COinS