Hybrid Bayesian-Wiener process in track geometry degradation analysis
| Author(s) | Galvan Nunez, Silvia Adriana | |
| Date Accessioned | 2018-02-09T14:34:44Z | |
| Date Available | 2018-02-09T14:34:44Z | |
| Publication Date | 2017 | |
| SWORD Update | 2017-11-10T17:22:54Z | |
| Abstract | Globally, track-caused accidents are a major factor of train derailments. Rail fatigue, rail wear, and track geometry defects are examples of track failure mechanisms. These mechanisms are usually modeled separately due to their individual characteristics, so maintenance activities are normally targeted to repair specific track structure components. Modeling track degradation and estimation of the failure time of the track is critical for safety and derailment purposes. ☐ In particular, the use of railway track geometry degradation models has played an important role in railway engineering. It helps in establishing track infrastructure maintenance policies and the output can be used to address derailment potential. Most track geometry degradation models are not stochastic and fail to account for small variations of the degradation values. On the other hand, failure time has been traditionally modeled using defect data. However, unless it is an accident due to extreme events, track geometry reaches a threshold as a result of an underlying degradation process. This dissertation focuses on the formulation of track geometry degradation and its first hitting time, in which two case studies were conducted using U.S. Class I railroad inspection data. ☐ The first case study formulates track geometry degradation as a Wiener process. The Wiener process is a stochastic process that models degradation for non-strictly monotonic increasing functions. Based on the characteristics of the track geometry data, the Wiener process appears to be suitable for modeling the degradation process. The model parameters were estimated using an adaptive Markov chain Monte Carlo algorithm. The second case study estimates the first hitting time (FHT) for each track geometry parameter and track section. The FHT is referred to as the probability distribution of the time at which the degradation path first reaches a safety threshold. The underlying degradation path is modeled as a Wiener process with drift and the FHT follows an inverse Gaussian distribution. ☐ Results from this dissertation provide a better understanding of track geometry degradation and failure by accounting for the inherent uncertainty in this process and by providing an alternative approach to identify track sections that require more attention for maintenance activities, considering each track geometry parameter. | en_US |
| Advisor | Attoh-Okine, Nii O. | |
| Degree | Ph.D. | |
| Department | University of Delaware, Department of Civil and Environmental Engineering | |
| Unique Identifier | 1022561937 | |
| URL | http://udspace.udel.edu/handle/19716/22920 | |
| Language | en | |
| Publisher | University of Delaware | en_US |
| URI | https://search.proquest.com/docview/1972753088?accountid=10457 | |
| Keywords | Social sciences | en_US |
| Keywords | Bayesian inference | en_US |
| Keywords | Railway engineering | en_US |
| Keywords | Stochastic processes | en_US |
| Title | Hybrid Bayesian-Wiener process in track geometry degradation analysis | en_US |
| Type | Thesis | en_US |