Statistical Analysis of Axle Load Distributions in India

Abstract

In this investigation, the axle load data for a particular stretch of National Highway (NH-58) in India was collected and statistical analysis was carried out. A lognormal and Weibull distribution was found to fit the single axles whereas lognormal mixed distribution was found to fit the tandem axles. Analysis of the data revealed considerable amount of overloading. When front and rear single axles were considered, unimodal distribution was found to fit the data. When only the tandem rear axles were considered, two or more peaks were observed, the first peak signified the mean of axle loads below the legal limit and the other peaks signified the means of axle loads exceeding the legal limit. Using the moment statistics, the load spectra factor was calculated for various axle types. The sensitivity of overloading in the calculation of equivalent single axle load was quantified using the vehicle damage factor and truck factors.

Determination of load equivalency factors by statistical analysis of weigh-in-motion data

“…Previous research proved techniques to assess axle load spectra e.g. using regression analysis (Rys et al 2015), nonparametric factorial analysis (Gao et al 2005) or Bayesian model theory (Morales-Nápoles, Steenbergen 2014) or a robust aproach involving simple distribution analysis (Savio et al 2016). However there are difficulties with the use of such techniques in the everyday practice, and special abilities of the data analyst are requires.…”