Assessing the Prevalence of Suspicious Activities in Asphalt Pavement Construction Using Algorithmic Logics and Machine Learning

Abstract

Quality Control (QC) and Quality Assurance (QA) is a planned systematic approach to secure the satisfactory performance of Hot mix asphalt (HMA) construction projects. Millions of dollars are invested by government and state highway agencies to construct large-scale HMA construction projects. QC/QA is statistical approach for checking the desired construction properties through independent testing. The practice of QC/QA has been encouraged by the Federal Highway Administration (FHWA) since the mid 60's. However, the standard QC/QA practice is often criticized on how effective such statistical tests and how representative the reported material tests are. Material testing data alteration in the HMA construction sector can render the QC/QA practice ineffective and shadow the performance of asphalt pavements. The American Society of Civil Engineers estimates that $340 billion is lost globally each year due to corruption in the construction industry. Asphalt pavement construction consists of several sectors, including construction and transportation, which are prone to potential suspicious activities. There is approximately 18 billion tons of asphalt pavement on American roads, which makes the costs of potential suspicious activities unacceptably large. The Idaho Transportation Department (ITD) relies on contractor-produced QC test results for the payment of the HMA pavement projects. In 2017, a case study by FHWA found some unnatural trends where 74% of the ITD test results didn't match with the contractor results. ITD's approach to track down the accuracy of mix design and volumetric test data set the off-stage of this research to mark out instances of suspicious activities in asphalt pavement projects. The first objective of this research was to develop algorithmic logics to recognize the patterns of discrepancies in agency- and contractor-produced QC/QA test results. This was possible with a unique dataset that ITD collected from several dozen HMA projects, in which all instances of data entry into the material testing report file was recorded in the background, without the operators' knowledge. My solution was bifurcated into development of an algorithm combining the logics to automatically detect and categorize suspicious instances when multiple data entries were observed. Modern data mining approaches were also used to explore the latent insights and screen out suspicious incidences to identify the chances of suboptimal materials used for paving and extra payment in HMA pavement projects. I have also successfully prompted supervised machine learning techniques to detect suspicious cases of data alterations. The second step of this research was to calculate the monetary losses due to data alteration. I replicated ITD's procedure for HMA payment calculation, and quantified payment-related parameters and associated payment for each project for two cases: 1. when the first parameter value categorized as Suspicious Alteration (S.A.) was used for payment calculation, and 2. when the last S.A. parameter value was used for payment. It was evident from my findings that there has been overpayment on construction projects across Idaho due to material testing data alterations. Overall, based on the available audit data, I found that overpayments have ranged from $14,000 to $360,000. Further analysis showed that alteration of each major material testing parameter's value can cause roughly $1,000 to $5,000 overpayment. I also note that data alteration did not always cause monetary gains. Other possible motives may include passing Percent Within Limit (PWL) criteria and precision criteria. Throughout the research, I strive to automate a suspicious activity detection system and calculate the associated excessive payment.