This study presents a machine learning approach for estimating the presence and extent of flange-face lubrication on a rail. It offers an alternative to the current empirical and subjective methods for lubrication assessment, in which track engineers’ periodic visual inspections are used to evaluate the condition of the rail. This alternative approach uses a laser-based optical sensing system developed by the Railway Technologies Laboratory (RTL) located at Virginia Tech in Blacksburg, VA, combined with a machine learning calibration model. The optical sensing system can capture the fluorescence emitted by the grease to identify its presence, while the machine learning model classifies the extent of grease present into four thickness indices (TIs), from 0 to 3, representing heavy (3), medium (2), light (1) and low/no (0) lubrication. Both laboratory and field tests are conducted, with the results demonstrating the ability of the system to differentiate lubrication levels and measure the presence or absence of grease and TI with an accuracy of 90%.

In machining processes, monitoring the condition of the tool is a crucial aspect for high productivity and quality of the product. Using different ML techniques in Tool Condition Monitoring (TCM) enables a better analysis of the large amount of data of different signals acquired during the machining processes. Different tool wear conditions were considered during the experimentation. A comprehensive statistical analysis of the data and feature selection using decision trees was conducted, and the KNN algorithm was used to perform classification. Hyperparameter tuning of the model was done to improve the model’s performance. In this research for the first time, a model agnostic approach to increase the interpretability and get an in-depth understanding of decision making is done. This research paper presents a KNN-based white box model, which allows us to dive deep into how the model performs the classification and how it prioritizes the different features included. This approach helps in detecting why the tool is in a certain condition and allows informed decision-making.

A revenue-service track evaluation of a novel, non-contact, laser-based fluorescence sensing system for assessing flange lubrication conditions is presented. The results include the system’s application onboard a hi-rail vehicle capable of operating at high speeds in field conditions. A custom-built structure enables easy installation and adjustment of the system during tests at speeds up to 30 mph (48 km/h). Flange lubrication conditions across curved sections and grease applicator zones are measured, and the results are displayed in an intuitive graphical user interface (GUI) for assessment of the flange lubricity measured at track speeds. Beyond detailing the test results, the paper provides an in-depth look at the GUI’s development in Python (Version 3.12.4), a high-level, general-purpose programming language. The results indicate that the system can consistently detect and distinguish track lubrication conditions. They further highlight the system’s ease of use onboard vehicles commonly used for maintenance-of-the-way condition assessment, such as hi-rail trucks or geometry cars. The GUI’s implementation provides a platform for easy-to-interpret evaluation of measurements on revenue-service tracks.