Impact of accuracy of weigh-in-motion data for pavement analysis and design: influence of temperature variations

This paper investigates the effect of temperature on WIM-collected data using the steering axle load spectrum (SALS) and two datasets. The first dataset, provided by the WIM system operator, contains detailed raw data, while the second, supplied by the road authority, includes less detailed information but covers 46 stations with various sensor types and configurations: single-line bending plates and two-line quartz piezoelectric sensors. A new methodology and parameter, the Spectral Bias Index, were introduced to evaluate WIM data accuracy. This approach effectively identifies stations operating correctly or incorrectly, with bias percentages of 6% for correct stations and 15% for incorrect ones. These biases significantly influence equivalency factors for pavement design, ranging from 1.28 to 2.52, particularly increasing for stations using single-line bending plate sensors. Pavement design datasets should span at least a year to reduce temperature effects and be checked for accuracy using the proposed steering axle load spectrum approach.