Predicting Track Geometry Defect Probability Based on Tie Condition Using Pattern Recognition Technique

Rail transportation plays a vital role in U.S. transportation. According to the National Transportation Statistics report from the Bureau of Transportation Statistics, railroads generate 29% of ton-miles of freight, whereas air, truck, water, and pipeline transportation represent the rest of the freight traffic. In the passenger spectrum, the 2016 National Transit Summary and Trends report stated that trips using rail transit modes increased from 2012 to 2016. These facts show the importance of rail transportation in the United States and highlight the critical importance of railroad traffic safety. Based on the FRA 2016 statistics, track-related defects are the second-largest cause of rail accidents. Furthermore, track irregularities resulting from defects in these parameters lead to an increase in dynamic forces that accelerate the rate of track deterioration. Consequently, the need to monitor and detect the presence and types of defects on railway tracks arises. The availability of track geometry cars and autonomous visual inspection vehicles has made acquiring track information easier. However, the need to study and understand these data remains unfulfilled. Machine learning has recently started to gain popularity in the field of railroad research due to increasing computational capacity and the need for such tools to provide information that is more useful. Techniques such as deep convolutional neural networks (DCNN), artificial neural networks, and support vector machines have been used for prediction problems in railroad research. This paper combines engineering judgments and statistical analysis to develop analytical models to estimate the probability of developing geometry defects as a function of tie conditions. The analysis is based on data provided by Georgetown Rail’s AURORA tie inspection system and from a major US class 1 railroad track geometry cars. The data used in this analysis relates to a geometry defect dataset and a tie condition dataset. The geometry dataset covers 125,554 geometry defects taken from several years of track-geometry inspection data. The data collection period was from 2014 to 2016. Convolutional neural network models were developed to estimate the probability of defects given tie patterns, as well as the outputs of the models used to build multiple regression models. Additionally, various data analysis issues were addressed in this paper. This paper’s contribution includes predictive models of track geometry defects as a function of tie condition and position. The models provide approaches to predicting the probability of geometry defects as functions of tie conditions and positions.

Judge Me by My Size, Do You? How Reliable Are Dent Assessments Based on ILI Data?

Pipeline dents have historically been regulated and assessed using dent depth as the primary metric. Many of the earliest analytical models for dent remaining life are based upon depth. Current assessment guidelines from ASME and the Code of Federal Regulations utilize depth as a primary metric. Consequently, ILI geometry tool capabilities are stated in terms of dent depth. However, the best modern dent assessments, including both strain and fatigue assessments, are based on dent shape. At a minimum, these models require both axial and circumferential dent profiles, or the models may utilize the full three-dimensional shape of the dent. The utilization of advanced dent assessments is expected to grow in the future as the methods are incorporated into API Recommended Practices and US regulations. While operators may have confidence in the ability of an ILI tool to confidently capture the dent depth, the shape of a dent is a recent consideration that is not addressed by current tool specifications. Unlike depth alone, dent shape is often a function of sensor coverage, speed, and caliper technology. Unfortunately, there is virtually no information available on the reliability of these assessment methods when they are based on ILI data. To-date, there have been no published comparisons examining the variation in strain or fatigue life in identical dents between multiple inspections. The reliability of these dent assessment methods is critical when choosing safety factors or reinspection intervals. This study presents a first look at the repeatability of strain and remaining life assessments based on two separate geometry inspection using different technologies. The study examines dent strain according to ASME B31.8 and fatigue life calculated using shape factors and finite element methods for 257 dents. The paper examines the variation in each of the methods and provides guidance on how users should understand the results when they are based on a single geometry inspection.

Effective time interval for railway track geometry inspection

A proper decision-making scheme for track geometry maintenance requires a knowledge of the real condition of track geometry. Therefore, the track must be inspected by measurement cars at different time intervals. The frequency of track geometry inspection plays a crucial role in decision-making and has always been a big concern for infrastructure managers. The inspection interval should be chosen properly, it means that the small period can decrease the capacity of line and affect the operation of network and the big period can result in low quality of track and in some cases derailments and possible loss of human lives. The aim of this paper is to determine the effective inspection interval such that the total maintenance cost is minimized. In the proposed cost model, the costs of inspection, preventive maintenance, corrective maintenance and the penalty for exceeding the corrective maintenance level are considered. A case study is performed on a real dataset collected from a railway line in Iran. The standard deviation of longitudinal level is considered to measure track geometry degradation. A widely applied linear model is used to model track geometry degradation over time. Monte Carlo technique is used to simulate the track geometry behavior under various track geometry inspection intervals. In addition, a set of sensitivity analyses are carried out to assess the effect of various inspection intervals on different terms of maintenance cost. The results indicate that not only can substantial costs be saved by setting effective inspection intervals, but also the time during which the track suffers from bad conditions is dramatically reduced. The result of this study has shown the appropriate inspection interval for the studied case can result in 13.6 percent decrease in maintenance cost in comparison with the current maintenance policy. Besides, it would lead to more reliable railway track by preventing the system exceed the corrective threshold.

A Case Study: Failure Analysis of Crude Oil Pipeline Rupture

More than 80% of crude oil requirement in India is met through imports. Imported crude oil is delivered to the shore tanks through Single Point Mooring (SPM) system. Generally, SPM systems are installed in the sea where water depth is around 30m and more. Crude oil tankers discharge their cargo through these SPMs and off-shore pipelines to storage tanks located in the shore. Therefore, off-shore crude unloading pipelines are a vital link to in the energy supply chain in India. Management of these off-shore pipelines is a challenging task. This paper discusses a case of mechanical damage to an Indian off-shore pipeline and how the damage is being evaluated to ensure reliability and safety of this vital link to ensure sustained and safe operation of the line. The mechanical damage discussed in this paper is in a 48″ off-shore pipeline at a depth of nearly 30m and 24km away from the shore. Owners believe that the damage was caused due to anchor hit from a ship that was buffeted away from safe anchor zone to no anchor zone during a cyclonic storm. Owner had to face considerable challenge in locating and measuring the extent of damage and evaluating its severity and probable impact on the integrity of the pipeline. Owner had done multiple geometry inspection of the pipeline to measure the length of the damage and restriction introduced in the bore due to local reduction in diameter. Possibility of presence of a crack and its likelihood of growth in the near and distant future is also evaluated. The paper also discusses the possible remedial measures to ensure long term integrity of the pipeline.

A threshold-regression model for track geometry degradation

Track-related failures are a major factor contributing to train derailments in the United States. Therefore, determining the failure time is critical for safety purposes. Traditionally, failure time in track geometry has been modeled using defect data. However, unless it is an accident due to extreme events, track geometry fails as a result of an underlying degradation process. The first hitting time is referred to the probability distribution of the time at which the degradation path first reaches a safety threshold. This paper presents the formulation and implementation of the first hitting time in railway track geometry degradation using track geometry inspection data. The underlying degradation path is modeled as a Wiener process with drift, and the first hitting time follows an inverse Gaussian distribution. The results provide a more robust representation of the failure time in track geometry using degradation data.

Single image geometry inspection using inverse endoscopic fringe projection

Fringe projection is an important technology for the measurement of free form elements in several application fields. It can be applied for geometry elements smaller than one millimeter. In combination with deviation analysis algorithms, errors in fabrication lines can be found promptly to minimize rejections. However, some fields cannot be covered by the classical fringe projection approach. Due to shadowing, filigree form elements on narrow or internal carrier geometries cannot be captured. To overcome this limitation, a fiberscopic micro fringe projection sensor was developed. The new device is capable of resolutions of less than 15 µm with uncertainties of about 35 µm in a workspace of 3 × 3 × 3 mm³.<br />Using standard phase measurement techniques, such as Gray-code and cos²-patterns, measurement times of over a second are too high for in-situ operation. The following work will introduce a new approach of applying a new one image measuring method to the fiberscopic system, based on inverse fringe projection. The fiberscopic fringe projection system employs a laser light source in combination with a digital micro-mirror device (DMD) to generate fringe patterns. Fiber optical image bundles (FOIB) are used in combination with gradient-index lenses to project these patterns on the specimen. This advanced optical system creates high demands on the pattern generation algorithms to generate exact inverse patterns for arbitrary CAD-modelled geometries. Approaches of the optical simulations in the context of the complex beam path, together the drawbacks of the limited resolutions of the FOIBs shall be discussed. Early results of inverse pattern simulations using a ray tracing approach of a pinhole system model are presented.<br />

On the Application of LIDAR Sensors for Track Geometry Monitoring

The applicability of Doppler based, Light Detection and Ranging (LIDAR) sensors for track geometry monitoring is investigated in this article. The optimum configuration of the LIDAR lenses, suitable for effective lateral track geometry measurements, is presented. In the proposed embodiment, two low-elevation LIDAR optics are employed with their beams reflected off of the rail gage face on each side to capture speed signals from both rails individually. A processing technique is developed based on the frequency bandwidth dissimilarities between the vehicle speed and track geometry irregularity to separate the two from each other. LIDAR system is body-mounted to a geometry inspection car and tests are carried out on revenue service track. The results indicate that LIDAR optics can provide a reliable track monitoring instrument for use over substantial track mileage in inclement weather and harsh track conditions with minimal operator supervision.