Browsing by Author "Soufiane, Kenza"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
Item Impact of adjacent support condition on premature crosstie failure(University of Delaware, 2021) Soufiane, KenzaRailroad cross-ties are a crucial element of the track structure as they contribute to the distribution of loads through the track. Determining the life of cross-ties, and particularly timber cross-ties is of importance in both design and maintenance decisions and activities. This includes mathematical modelling of the average crosstie life and predicting how different factors affect its lifespan. It also includes the use of automated tie inspections to provide information on ties condition, which in turn allows for the determination of the ties life and the making of decisions on tie maintenance and replacement operations. ☐ This thesis addresses the issue of the use of automated tie condition inspection data to predict and model tie life based on support condition, as defined by the condition of adjacent cross-ties. The thesis presents several analysis approaches, all e based on the use of tie condition data from two different inspections performed over a span of 3 years. The dataset used consisted of tie inspection data from approximately 100,000 crossties), from multiple inspections carried out on the same track during the period 2016 to 2019. ☐ The approach taken is a comprehensive data analysis approach starting with data preprocessing and exploratory data analysis. A macro-analysis was then performed in order to determine the mile-by-mile behavior of ties, followed by a microanalysis that allowed for the study of each individual tie and its degradation behavior. Noting that tie loading was dependent on the condition of adjacent ties on both sides of a given tie, ties were grouped based on their adjacent tie condition and their corresponding loss of adjacent tie support as defined by the Beam on Elastic Foundation theory. In the initial analysis, a simplified formula was developed that described the life reduction as a function of this loss of adjacent tie support. ☐ The second part of the thesis consisted of a more detailed investigation of the tie condition changes over the three years study period. In this analysis a piecewise reconstruction of the average tie life was performed and used to compare the tie degradations rates with respect to loss of adjacent tie support. As such, regression functions were developed for the different support groups, based on the distributions of the different tie condition score transitions from 2016 to 2019 . These functions were then used recursively to predict the tie score change over time. Then, Dijkstra’s algorithm was applied to model each group’s average tie life based on adjacency matrices developed using the tie score transitions. In a third analysis approach, Markov chains were utilized for the determination of the probability of tie failure as a function of loss of support. ☐ The results showed different average tie lives for different support conditions confirming the fact that loss of support contributes significantly to premature tie failure with the higher the loss of support, the shorter the average tie life. Life reduction formulas were then generated based on the three analysis approaches.Item THE DYNAMIC INTERACTIONS OF ADJACENT CROSSTIES DEGRADATION RATES: A THEORY GUIDED MACHINE LEARNING FRAMEWORKSoufiane, KenzaUnderstanding the deterioration behavior of crossties (also known as railroad ties or sleepers) is of paramount importance due to the major safety concerns associated with their failure. Studies have been conducted to assess ties degradation and their maintenance planning considering different parameters: tonnage, climate conditions, type of tracks, frequency of trains, and degree of curvature, among others. However, very few studies determined the relation between a crosstie’s degradation rate or failure and that of its adjacent ties. Degradation rates of adjacent ties are interconnected and interact in a dynamic manner over time. On track, ties have different conditions which leads to an imbalanced load distribution: as crossties deteriorate, more load is exerted on their adjacent ties, causing them to degrade faster. Consequently, as these adjacent ties deteriorate as well, their capacity to withstand additional loads diminishes, resulting in the transfer of more load to nearby ties, accelerating their degradation. The work performed in this PhD thesis investigates this complex dynamic interaction between changing crosstie condition over time (through degradation and replacement) and the support it provides to adjacent crossties. In fact, the objective is to use a theory-driven machine learning framework incorporating real data to understand, model, and predict cross-tie degradation behavior and the corresponding cross-tie life under actual in-field conditions. Using different machine learning techniques, tie degradation trends were investigated to not only determine the dynamic impact of adjacent support condition on a center tie, but to also quantify the load distribution effect over time. Then, using automated tie inspection data, theory was integrated into the machine learning framework using three key approaches: constructing the model's architecture in accordance with a comprehensive mathematical formulation to ensure it adheres to tie degradation mechanisms; designing a loss function that aligns with the underlying tie degradation trends, in addition to theory-guided calibration of the tie condition input data. The established framework allowed for the use of real tie data, supported by well-defined railroad engineering relationships, to forecast the tie condition as a function of its adjacent ties and their corresponding degradation rates over time. Theory Guided Machine learning models were developed with the domain knowledge from an engineering-based Beam-On-Elastic-Foundation track model with variable local stiffness. The models were able to determine the load variation and associated rate of degradation on a given cross-tie, based on the condition of both the cross-tie itself and that of the adjacent ones. Using three years of railroad cross-ties inspection data, the resulting models forecast the fourth-year condition. The models showed excellent correlation with the test data set, exhibiting strong performance indicators, and outperforming more conventional traditional neural networks. They also accurately represented tie degradation behavior and effectively captured the complex dynamics introduced by tie replacements, suggesting that the incorporation of domain knowledge into the machine learning model leads to demonstrably better tie life prediction results.