Category: Safety

Develop a risk based model for visual inspection considering defect occurrence rate, ability and efficiency of inspector, track utilization, etc. as a function of recorded inspection results. Walking inspectors can become complacent or may not be well trained. Defects may be hard to see or missed. This can cause the risk of derailment to increase. The type of traffic and speed of operations influence the consequences of a derailment. Historic track performance with regard to previous inspection offers an opportunity to understand the risk of defects occurring. Optimizing the visual  inspection frequency can reduce derailments and offer the railroads cost efficiencies, based on past track performance.

The proposed research will evaluate historic inspection records that include identification of defects, as well as condition observations. This data can be utilized to understand track performance. The repeated inspection data can be used to evaluate the change in condition, given the frequency of inspection. Coupling this data, along with visual inspection efficacy (modeled as a probability distribution based on previous research) allows for understanding the impact of changing the inspection frequency. This research will develop a model that provides the impact of changing the inspection frequency given these performance characteristics.

As railroads increase use of autonomous track geometry measurement systems, coupled with traditional measurement cars, hy-rail inspection and walking inspection, the reliability of inspection can increase. The additional data captured by autonomous inspection allows for a better understanding of track geometry perturbation growth. The distribution of accuracies associated with alternative inspection methods must be taken into account when defining the risk of a defect, which drives inspection frequency surfacing maintenance. This research project fuses these data sources to better understand track geometry perturbation growth, particularly in those areas where high impact forces and high lateral to vertical force ratios can be experienced, which may lead to derailment. The perturbation growth data is then used with inspection accuracy to be developed from published sources as well as the provided data. this in turn is used to optimize automated inspection.