Year of Publication
Doctor of Philosophy (PhD)
Dr. Nikiforos Stamatiadis
Dr. Mei Chen
The context of this research is the investigation and application of an approach to develop an effective evaluation methodology for establishing the investment worthiness of a range of potential Light Rail Transit (LRT) major system improvements (alternatives). Central to addressing mobility needs in a corridor is the ability to estimate capital costs at the planning level through a reliable and replicable methodology. This research extends the present state of practice that relies primarily on either cost averages (by review of cost data of implemented LRT projects) or cost categories in high and low cost ranges. The current methodologies often cannot produce accurate estimates due to lack of engineering data at the planning level of project development. This research strives to improve current practice by developing a prediction model for the system costs based on specific project alignment characteristics.
The review of the literature reflects a wide range of estimates of capital cost within each of the contemporary mass transit modes. The primary problem addressed in this research is the challenge associated with producing capital cost estimates at the planning level for the LRT mode of public transportation in the study corridor. Furthermore, the capital cost estimates for each mode of public transportation under consideration must be sensitive to a range of independent variables, such as vertical and horizontal alignment characteristics, environmentally sensitive areas, urban design and other unique cost-controlling factors. The current available methodologies for estimating capital cost at the planning level, by transit mode for alternative alignments, have limitations. The focus of this research is the development of a statistical theory-based, capital cost-estimating methodology for use at the planning level for transit system evaluations. Model development activities include sample size selection, model framework and selection, and model development and testing. The developed model utilizes statistical theory to enhance the quality of capital cost estimation for LRT investments by varying alignment characteristics.
This research has identified that alignment guideway length and station elements (by grade type) are the best predictors of LRT cost per mile at the planning level of project development. For the purpose of validating the regression model developed for this research, one LRT system was removed from the data set and run through the final multiple linear regression model equation to assess the model’s predictive accuracy. Comparing the model’s estimated cost to the projects final construction cost resulted in a 26.9% error. The percentage error seems somewhat high but acceptable at the planning level, since a 30% contingency (or higher) is typically applied to early level cost estimates.
Additionally, a comparison was made for all LRT systems used in the model estimation and the percent error range is from 2.4% to 111.5% with just over 60% of the project’s predicted cost estimate within 30% or better. The model appears to be a useful tool for estimating LRT cost per mile at the planning level when only limited alignment data is available. However, further development of improved predictive models will become possible when additional LRT system data becomes available.
Digital Object Identifier (DOI)
Catalina, Anthony J., "Development of a Statistical Theory-Based Capital Cost Estimating Methodology for Light Rail Transit Corridor Evaluation Under Varying Alignment Characteristics" (2016). Theses and Dissertations--Civil Engineering. 48.