NORMAL LOAD ESTIMATION BY USING A SWARM INTELLIGENCE BASED MULTIPLE MODELS APPROACH
AbstractStatically weighing a railway vehicle is the traditional way to determine the weight of a railway vehicle. Recently, approaches for dynamically weighing railway vehicles are proposed and such approaches are based on replacing sensors to track sections. In this study, the possibility of estimating loading conditions is discussed by using only vehicle-mounted sensors. A multiple model based estimation scheme is presented. In this approach, models indicate the mathematical models of the railway vehicle which consider different normal loads. Furthermore, a swarm intelligence-based evolution of models is provided instead of using fixed models. The methodology assumes that the normal load can be estimated by using angular and translational velocity measurements taken from the vehicle. To validate the methodology, it is applied to the tram wheel test stand, which is used for research purposes in University of Pardubice (Czechia), for a normal load by considering the angular velocity measurements. In future work, it will be tested for different loading conditions with different peripheral velocities of the test stand.