Physics Based Wear Model for Predictive Maintenance of Rail-infrastructure

Meghoe, A. (Speaker)

    Activity: Talk or presentationOral presentation


    The very first rails in the Netherlands were introduced around 1839 and since then the maintenance time intervals are mainly based on historic data, which means that damage prediction is not possible for time-varying loading or environmental conditions. This can be changed by using physics based models. A physical model is able to calculate the yet unknown future degradation if the current state of the system is known. To achieve this, numerical or mathematical descriptions of dominant failure mechanisms are usually used. Wear and Rolling Contact Fatigue (RCF) are identified as the most dominant failure mechanism for tracks and switches which are considered to be the most critical components in the rail-infra. To understand the governing loads of the above mentioned failure mechanisms the concept of Railway Vehicle Dynamics and Rolling Contact Theory are studied.

    As a first case, the wear model based on Archard’s law for adhesive wear, which
    indicates that the amount of wear volume loss is related to sliding distance, normal force and hardness of material, is implemented. Typically the wear prediction procedure then consists of three parts: (1) Multi body simulation to predict the contact forces and contact points; (2) Local wheel-rail contact model by using Hertz theory to determine contact conditions; (3) Local wear model e.g. by using the Archard’s law. The output of the vehicle – track simulation serves as input for the local contact model. The size and shape of the contact area depend on different combinations of wheel and rail profiles. If the distribution of the size and location of the contact area on the rail are known, the accumulated wear depth at critical points of the rail can be predicted and so can the Remaining Useful Life (RUL). In the presentation, the structure of the model and its separate elements will be discussed and preliminary results will be shown.
    Period3 Feb 2017
    Event titleMaintenance Research Day 2017
    Event typeConference
    LocationUtrecht, Netherlands
    Degree of RecognitionInternational