Railway bogies are generally maintained preventively within certain time periods. Vehicles run for a long time (up to thirty years) so that about one-third of lifecycle costs are caused by maintenance. Condition-based predictive maintenance strategies offer economic improvements of up to 15 percent and an increased availability of up to 100 percent. By implementing a sensor based diagnostic system using artificial intelligence techniques and business analytics, maintenance can be optimized.
This paper presents a concept for the detection of faulty mechanical components, such as dampers and springs, of railway bogies. On the one hand, the work deals with the analysis of time series data by evaluating power spectral density and transfer functions. On the other hand, it shows results of the training of machine learning models based on features. Various multi body simulations have been done to develop a physical understanding of the effects of the individual fault modes. Beside this, the available data set for the analysis also includes real measured data from test rides with faulty components.
The developed algorithms have been deployed on a Siemens commuter train. In addition to validation results, this work also shows an example of the successful detection of real failures in operational mode.
Railway Bogie Diagnostics Using Machine Learning and Bayesian Net Reasoning Approaches
Lect.Notes Mechanical Engineering
The IAVSD International Symposium on Dynamics of Vehicles on Roads and Tracks ; 2021 August 17, 2021 - August 19, 2021
2022-08-06
11 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Railway vehicle bogie frame, bogie and railway vehicle
Europäisches Patentamt | 2023
|Railway vehicle bogie frame, bogie and railway vehicle
Europäisches Patentamt | 2023
|Railway vehicle bogie frame, bogie and railway vehicle
Europäisches Patentamt | 2023
|