Computational Technology Resources - CCC

B. Alkali¹, M.M. Rahman¹, J.P. Gutierrez¹, A.K. Jain², C. Mcneil³ and J. Nelson³

¹School of Science and Engineering, Glasgow Caledonian University, United Kingdom
²Department of Mechanical Engineering, Birla Institute of Technology and Science, India
³, Siemens Mobility Limited, Glasgow, United Kingdom

doi:10.4203/ccc.15.8.1

B. Alkali, M.M. Rahman, J.P. Gutierrez, A.K. Jain, C. Mcneil, J. Nelson, "Pantograph Failure Analysis and Short-Term Forecasting for EMU Fleet Using Hybrid and Machine Learning Models", in J. Pombo, (Editor), "Proceedings of the Seventh International Conference on Railway Technology: Research, Development and Maintenance ", Civil-Comp Press, Edinburgh, UK, Online volume: CCC 15, Paper 8.1, 2026, doi:10.4203/ccc.15.8.1

Keywords: pantograph, forecasting, predictive maintenance, railway rolling stock, ARIMA-ANN, LSTM, decision tree regressor..

Abstract

Pantograph systems are critical to the reliable operation of electric multiple unit (EMU) trains because they maintain electrical contact with the overhead catenary. Failures in these systems can lead to maintenance issues, service disruptions, and train cancellations. This study presents a data-driven framework for pantograph failure analysis and short-term forecasting for the EMU fleet operating in Scotland. A five-year historical maintenance data was analysed, together with operational failure records, to ascertain the pantograph failure pattern at both fleet and unit levels. The methodology combines data preparation, time-series construction, multi-step forecasting, and quantile-based risk classification. Four forecasting models were compared: Autoregressive Integrated Moving Average (ARIMA)-Artificial Neural Network (ANN), Long Short-Term Memory (LSTM)-Support Vector Regression (SVR), LSTM, and Decision Tree Regressor. The results show that the ARIMA-ANN hybrid model achieved the best overall performance, with the lowest forecasting errors among the compared methods, indicating that the pantograph failure series contains both linear and nonlinear characteristics. Based on the best-performing model, fleet-level and unit-level forecasts were generated for the next 15 operational periods and level into low, medium, and high-risk. The proposed framework provides technically interpretable and practically useful decision support for proactive railway maintenance planning and service delivery.

download the full-text of this paper (PDF, 10 pages, 825 Kb)

go to the previous paper
go to the next paper
return to the table of contents
return to the volume description

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCC CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Conferences ISSN 2753-3239 CCC: 15 PROCEEDINGS OF THE SEVENTH INTERNATIONAL CONFERENCE ON RAILWAY TECHNOLOGY: RESEARCH, DEVELOPMENT AND MAINTENANCE Edited by: J. Pombo Paper 8.1 Pantograph Failure Analysis and Short-Term Forecasting for EMU Fleet Using Hybrid and Machine Learning Models B. Alkali¹, M.M. Rahman¹, J.P. Gutierrez¹, A.K. Jain², C. Mcneil³ and J. Nelson³ ¹School of Science and Engineering, Glasgow Caledonian University, United Kingdom ²Department of Mechanical Engineering, Birla Institute of Technology and Science, India ³, Siemens Mobility Limited, Glasgow, United Kingdom doi:10.4203/ccc.15.8.1 Full Bibliographic Reference for this paper B. Alkali, M.M. Rahman, J.P. Gutierrez, A.K. Jain, C. Mcneil, J. Nelson, "Pantograph Failure Analysis and Short-Term Forecasting for EMU Fleet Using Hybrid and Machine Learning Models", in J. Pombo, (Editor), "Proceedings of the Seventh International Conference on Railway Technology: Research, Development and Maintenance ", Civil-Comp Press, Edinburgh, UK, Online volume: CCC 15, Paper 8.1, 2026, doi:10.4203/ccc.15.8.1 Keywords: pantograph, forecasting, predictive maintenance, railway rolling stock, ARIMA-ANN, LSTM, decision tree regressor.. Abstract Pantograph systems are critical to the reliable operation of electric multiple unit (EMU) trains because they maintain electrical contact with the overhead catenary. Failures in these systems can lead to maintenance issues, service disruptions, and train cancellations. This study presents a data-driven framework for pantograph failure analysis and short-term forecasting for the EMU fleet operating in Scotland. A five-year historical maintenance data was analysed, together with operational failure records, to ascertain the pantograph failure pattern at both fleet and unit levels. The methodology combines data preparation, time-series construction, multi-step forecasting, and quantile-based risk classification. Four forecasting models were compared: Autoregressive Integrated Moving Average (ARIMA)-Artificial Neural Network (ANN), Long Short-Term Memory (LSTM)-Support Vector Regression (SVR), LSTM, and Decision Tree Regressor. The results show that the ARIMA-ANN hybrid model achieved the best overall performance, with the lowest forecasting errors among the compared methods, indicating that the pantograph failure series contains both linear and nonlinear characteristics. Based on the best-performing model, fleet-level and unit-level forecasts were generated for the next 15 operational periods and level into low, medium, and high-risk. The proposed framework provides technically interpretable and practically useful decision support for proactive railway maintenance planning and service delivery. download the full-text of this paper (PDF, 10 pages, 825 Kb) go to the previous paper go to the next paper return to the table of contents return to the volume description
Back to top	©Civil-Comp Limited 2023 - terms & conditions