How the EULYNX data prep standard can improve railway robustness

Standardisation

EULYNX is the initiative of European Infrastructure Managers to standardize interfaces between signalling systems and their periphery. Standardisation creates a world-wide market for peripheral device controllers ranging from train control systems to point controllers. In the foreseeable future, IM’s can pick and mix signalling systems from supplier A that controls field devices from supplier B. This will do for the railway signalling industry what the USB-standard did for the IT industry and the GSM-standard for telecommunications.
EULYNX also standardises data preparation. This consists of designing the data structures that capture the information a supplier needs to build a signalling system from the ground up. This is a radical improvement from the present where the parties involved in (re-)signalling projects exchange heterogeneous and proprietary datasets, often on paper. In the future, the signalling industry ingests standardised EULYNX data into their proprietary design toolset. As before, the signalling industry process the data and then return the enriched data, in EULYNX format, to the IM who absorb this as-built data into their asset management systems. IMs and signalling industry retain their proprietary formats and tooling but EULYNX harmonises data exchange. The transfer of data must not reduce the quality, in other words, the probability of data being corrupted in transit must be acceptably low. Manual transfer is notoriously error-prone and the cost of finding and removing errors is high. The case for automating the data transfer process through well-defined standard data structures is obvious.
This paper describes how EULYNX defines classes of data that describe the objects we find in railway signalling, no easy task given the fact that European railways have nearly 200 years of technological traditions and jargon that must be reconciled. Answering a seemingly simple question like “what is a route” is fiendishly difficult once one starts drilling down. The participants from the various European IM’s in the data prep working group construct UML class models that analyse tangible and intangible signalling objects, define the semantics and most importantly, define the relations between objects and their states. We explain how we handle the different national signalling concepts and requirements to construct UML class models.
Finally, the UML models will be transformed into XSD schemata that harness XML data exchange between data-producers and consumers. Whilst the primary use case for this UML model remains the lossless data transfer between IM’s and signalling industry, one can think of many other use cases such as capacity analyses that need accurate information about signalling systems to forecast traffic. Another prospective use case is to improve the robustness of the railway system. The data model allows automated analysis of the relations between objects. One can design stress tests to answer questions like “what routes are affected when a given signal lamp fails” or “what is the impact on throughput if this train detection section fails” ?