ABSTRACT
This paper investigated 2 years of WIM datasets from a Brazilian highway to compare the results obtained by a large dataset versus simplifications usually taken by Brazilian pavement designers about the traffic parameters, especially the ESAL. Since the database contained more than 3.4 million entries, it was key to use tools for Big Data analysis; Python programming with the Pandas, NumPy and Matplotlib libraries were employed for this purpose. In addition, to generate axle loads spectra from the dataset, ESAL per year were calculated on different scenarios of the fleet : (i) considering each axle weight contribution (maximum use of dataset); (ii) considering the weight means of each axle type (medium use of dataset); (iii) considering the 100% of the axles are operating on the Brazilian legal limits (minimum use of dataset). The results highlighted the adequateness of the tools used in the analysis and pointed out the importance of considering all available entries to characterize the fleet in order to improve traditional premises in pavement design.
