Search results

In this paper, a data mining approach is proposed for monitoring the conditions leading to a rail wheel high impact load. The proposed approach incorporates logical analysis of data (LAD) and ant colony optimization (ACO) algorithms in extracting patterns of high impact loads and normal loads from historical railway records. In addition, the patterns are employed in establishing a classification model used for classifying unseen observations. A case study representing real-world impact load data is presented to illustrate the impact of the proposed approach in improving railway services.

Application of artificial intelligence and machine learning approaches becomes an essential tool in improving the performance of railway transportation systems. By using these approaches, the knowledge extracted from historical data can be employed in railway assets monitoring to maintain the assets in a reliable state and to improve the service provided by the railway network.

Results achieved by the proposed approach provide a prognostic system used for monitoring the conditions surrounding rail wheels. Incorporating this prognostic system in surveilling the rail wheels indeed results in better railway services as trips with no-delay or no-failure can be realized. A comparative study is conducted to evaluate the performance of the proposed approach versus other classification algorithms. In addition to the highly interpretable results obtained by the generated patterns, the comparative study demonstrates that the proposed approach provides classification accuracy higher than other common machine learning classification algorithms.

The methodology followed in this research employs ACO algorithm as an artificial intelligent technique and LDA as a machine learning algorithm in analyzing wheel impact load alarm-collected datasets. This new methodology provided a promising classification model to predict future alarm and a prognostic system to guide the system while avoiding this alarm.

The purpose of this paper is to test the applicability and the performance of an approach called logical analysis of data (LAD) on the detection of faults in rotating machinery using vibration signals.

LAD is a supervised learning data mining technique that relies on finding patterns in a binary database to generate decision functions. The hypothesis is that a LAD‐based decision model can be used as an effective tool for automatic detection of faults in rolling element bearings. A novel Multiple Integer Linear Programming approach is used to generate patterns for the LAD decision model. Frequency and time‐based features are extracted from rotor bearing vibration signals and are pre‐processed to be suitable for use with LAD.

The results show good classification accuracy with both time and frequency features.

The diagnostic tool implemented in the form of software in a production or operations maintenance environment can be very helpful to maintenance experts as it reveals the patterns that lead to the diagnosis in interpretable terms which facilitates efforts to understand the reasons behind the components' failure.

The proposed modifications to the LAD‐based decision model which is being tested for the first time in the field of fault detection in rotating machinery lead to improved accuracy results in addition to the added value of result interpretability due to this distinctive property of LAD.