Railcar condition is an important factor in the complex web of relationships between railroads, railcar leasing companies, shippers and railcar builders. The most important reasons for this are operational safety and economic considerations pertaining to equipment maintenance. In this study, an approach is presented for the diagnostics of railcar component health from vibration data, utilizing mutual information (MI) based minimal-redundancy-maximalrelevance (mRMR) feature selection and multi-class support vector machine classification. The proposed monitoring solution is a data-driven method which was developed with measurements taken at a railroad test laboratory under controlled conditions. Vibration data was collected from multiple locations on a railcar over several test runs, each utilizing wheelsets with different levels of wear. The input of controlled wheel wear levels was aimed at varying the system outputs to resemble those of cars with different levels of mileage in revenue service. The measured data sets were processed in the time domain, frequency domain and throughwavelet transforms, resulting in the extraction of a set of 687 features from the acceleration signals. A maximum-relevance minimum-redundancy feature selection algorithm was usedto find the optimal combination of features for classification. The algorithm performance was tested for the effect of feature set size, different kernels and scaling techniques on classification accuracy. The results and methods of this assessment are presented in the paper. The paper concludes with a proposal for a monitoring strategy aimed at specifically detecting faulty components and practicing predictive maintenance.
A novel embedded min-max approach for feature selection in nonlinear Support Vector Machine classification
