Development of Weigh-in-Motion Data Quality Control Algorithms and Procedures

2020PDFTech ReportChorzepa, Mi GKim, S. SonnyDurham, Stephan A.Sinha, AnantaUniversity of Georgia. College of EngineeringGeorgia. Department of Transportation. Office of Performance-Based Management & ResearchUnited States. Department of Transportation. Federal Highway AdministrationGeorgia. Department of Transportation. Office of Performance-Based Management & ResearchUnited StatesAlgorithmsCalibrationData analysisData qualityQuality controlSensorsSpeed limitsValidationWeigh in motionQuality controlQuartzClassificationWeightComparison Data SetAxleAxle SpacingFHWA-GA-20-1836PI# 0016503Final; May 2019-August 2020This report recommends a Weigh-In-Motion (WIM) data quality control (QC) process that will assist in delivering higher quality WIM data and continuously improving WIM program performance as its user base grows within the Georgia Department of Transportation (GDOT). Specifically, this research study serves to establish a procedure for off-site office QC checks by analyzing WIM data from six permanent sites. The Office of Transportation Data (OTD) anticipates rapid growth in the volume of data seized from an increasing number of WIM sites over the next couple of years. The goal of this study is to assist GDOT\u2019s OTD with effectively managing and improving WIM data quality by providing a summary of findings that have practical implications (e.g., identifying lane numbers and site IDs needing a calibration or attention). The results indicate that a significant portion of invalid weight records exist in the 2018 data and that there are three critical steps to improve data quality: 1) replace Brass Linguini (BL) sensors with quartz sensors as soon as possible, and conduct calibration upon installation; 2) conduct an annual WIM device calibration observed by a third party, and 3) investigate means of enforcing the legal speed limit at WIM locations. Since the reliability of WIM data vastly depends on QC requirements, a thorough investigation of additional data and field validations of parameters contained in the data are necessary. A comparison data set (CDS) needs to be established immediately after a calibration for each WIM site. Additionally, such calibration should be conducted annually at three to five randomly selected sites. During a calibration, axle spacings should also be validated in association with the decision-making process employed by the vehicle classifier. Lastly, the study team recommends that calibration and maintenance logs be made available to the party conducting data QC checks.104