Using predictive maintenance to improve the safety and efficiency of railways

Having a modern rail fleet and state-of-the-art infrastructure to complement it, that meets demanding quality standards, requires advanced predictive maintenance to be used during the decision-making processes.

The application scope of predictive maintenance in the railway sector is wide-ranging, with potential to be used within infrastructure, superstructure and rolling stock. The implementation of machine learning techniques has a bearing on both maintenance and safety, which means the optimum solution should be developed in tandem.

Solutions must be implemented when there is a business case in which improvements in availability, reliability, maintainability and safety – with regard to infrastructure and rolling stock – can be accomplished. It is of the utmost importance, therefore, to consider whether the business model of a company requires predictive maintenance tools. To achieve this, it is essential to carry out an analysis of all the processes, as well as boasting the collaboration of different stakeholders. Once it has been identified where actions must be taken, these must be prioritised in accordance with the impact and resources necessary.

Comfort, safety and efficiency are watchwords to ensure maximum performance in terms of rolling stock and on the railways across the world. Presently, Spanish firms consider it paramount to employ the most advanced predictive maintenance processes and techniques to achieve these aims.

Predictive maintenance solutions are being increasingly implemented on all train sub-systems to create patterns of behaviour, meaning any anomalies will be quickly identified; thus optimising maintenance processes for fleets. It is significant to apply predictive maintenance methodology on elements that affect safety and can help reduce the level of down-time; improving availability and accomplishing greater levels of service reliability, coupled with a reduction in life-cycle cost (LCC).

Similarly significant is the application of predictive maintenance solutions, not solely in new equipment, but (when feasible) in those that are already in use, combining methodology that is dependent on the components having the same condition with general systematic and periodic actions on rolling stock.

With regard to new-build railway infrastructures, predictive maintenance solutions must be applied from the outset of their operation, with the major challenge being the extension of said solutions to any railway line already in existence. This point includes the actual railway structures because of their direct implication in safety during operations.

Furthermore, it is key to apply predictive strategies to products whose unavailability may affect the end user. In this way, an infrastructure free of faults is obtained by staff management, and other tasks associated with maintenance, being performed in an optimum manner.

Knowledge of the state of the infrastructure and its depreciation over time will be essential. There are critical points on every line that must be monitored, through the use of machine learning and artificial intelligence to predict their response to stress and wear and tear, amongst other aspects.

The use of data is likewise a relevant aspect with regard to predictive maintenance. In many cases, companies handle vast amounts of data coming from a number of sources. Much of this data is obtained thanks to the digitalisation of productive systems. It is a priority to combine data analysis with specific technology knowledge, transforming data into information that allows for the undertaking of reliable predictions. In this manner, it will be possible to make decisions in the medium- and long-term.

Conclusion

The benefits of predictive maintenance are wide-reaching and have positive repercussions on the safety, reliability and condition of the infrastructure, as well as the life-cycle of the rolling stock, cost optimisation and reliability indexes. Furthermore, it opens the door to the possibility of scheduling activities based on the equipment’s condition, which are then assessed through data-processing techniques.

Spanish industry activity in the sector of predictive maintenance displays the advancements made in recent years. However, the industry still has many major challenges ahead, which could prove to be a major opportunity for the sector. Examples of this would be to integrate current predictive maintenance applications into the new digital environment and the management of large volumes of data. It has become indispensable to ensure the reuse of data and predictive maintenance solutions in which the client can create its own in-house knowledge or the articulation of collaborative eco-systems between operators, maintenance operators and specialist firms.

The future will also be shaped by the creation of tools composed of technologies – key enable technologies (KETs), such as automatic maintenance inspections via artificial intelligence, artificial vision or augmented reality, amongst others. In short, the challenge is to equip these new technologies with a level of maturity that allows for the full digitalisation of the railway industry.

And within all of this process, from inside Mafex, the Spanish Railways Association, there is a wide number of firms such as ALSTOM ESPAÑA, BOMBARDIER ESPAÑA, CAF, CEIT-IK4, COMSA, INECO, INSERAIL, LIMMAT GROUP, NEM SOLUTIONS, TALGO, TELICE, THALES ESPAÑA or TRIA INGENIERIA, to name just a few, that are playing an essential role in the development of innovative solutions for predictive maintenance, applied both on an infrastructure and a rolling stock level.

In the light of the foregoing, from the Association, we invite whoever shares an interest in this topic, to contact us and discover more about all of these firms and their high level of expertise in the matters tackled herein.