The pursuit of higher quality services in the railway sector is a continuous process, and the availability in recent years of affordable, reliable, digitally enabled additions to traditionally mechanical-based infrastructure systems has provided a fruitful avenue for advancement. Remote Condition Monitoring (RCM) systems are one example of a tool that has been widely deployed to improve the standards of maintenance, reliability, and safety across the rail network. The advanced warnings of incipient faults provided by RCM data enable preventative maintenance to be performed before service-impacting failures arise, leading to reduced costs of disruption and increased passenger satisfaction. The perceived benefits of RCM have led the industry to install sensors on an ever-higher proportion of its assets, with a corresponding increase in the volume of data generated. In general, and according to [1], railway RCM operations can be divided into four major divisions (quadrants), which are defined by the location of the monitoring sensors and the assets being monitored: train monitoring train, infrastructure monitoring infrastructure, train monitoring infrastructure, and infrastructure monitoring train. In countries such as the UK, where the vast majority of the mainline rail infrastructure is maintained by a single Infrastructure Manager (IM), sensors that are mounted on assets belonging to one stakeholder but are being used to monitor assets related to another will, by definition, fall into the train monitoring infrastructure or infrastructure monitoring train quadrants; an example of this would be sensors mounted on the tracks that are used to detect wheel flats on the rolling stock [2]. Although this type of cross-interface monitoring of assets may be the most technically practical solution to many industry-wide problems, commercially they can prove complex as the business paying to install, maintain, and operate the sensing device is not the party benefitting from the data collected. As a result, it can be hard to generate business cases for the purchase, installation, and operation of cross-interface monitoring systems that would have recognised industry-wide benefits.

In order to address this issue, it is widely recognised within GB rail that either closer collaborations must be established between stakeholders to enable more effective cross-interface business cases to be developed or there must be a trusted audit process that can enable costs of data collection to be fairly attributed based on business benefits accrued by individual stakeholders. To investigate these issues the Rail Safety and Standards Board (RSSB) set up a Cross-Industry RCM (XIRCM) research programme, which in turn acted as sponsor to the T1010 research project [3] from 2013 onwards. The stated aim of T1010 was to overcome the barriers for rail companies to use RCM systems across company boundaries, with the first round of findings presented by RSSB and Network Rail at the IET RCM conference in 2014 [4].

A key component of business case generation for cross-interface RCM is the assignment of value to the data generated by one party but used by another. In order to address the cost issue, it was suggested in project T1010 that commercial agreements could be established between all the actors in a new condition monitoring workflow before installation of the system began [5]. However, there are issues with this approach; commercial agreements do not remove the need for a trusted third party (arbiter) to ensure compliance with the terms of the agreement, and they do not inherently include any ongoing audit mechanism that would act as evidence should issues arise. In combination, these two issues act as a barrier to the full exploitation of XIRCM data and cost sharing between stakeholders.

Distributed Ledger Technologies (DLTs) have several features which can be leveraged to address the issues outlined. The benefits offered to the industry through improved system-wide asset information and decision support are clear, but for those benefits to be realised in a privatised rail system where the separation of business functions is the main architectural driver, the commercial implications of the operation of cross-industry systems for each actor must be clear. Further to this, existing investments in specific RCM systems made by the industry are currently only in their mid-life stages, meaning a method to deliver a clear understanding of operational costs must be cognizant of, and compatible with, the methods of operation of these existing assets. DLTs are one possible solution to these issues, offering the potential for traceability of data flows between industry actors with a minimum restructuring of the current systems. By understanding the flows of data between actors, and the ultimate costs/benefits accrued by the installation and use of the system (for which mechanisms are already in place), it will be possible to accurately assign costs to the relevant parties, to cut down on the operational inefficiencies associated with manual attribution and trusted third parties, and to enable improved understanding of data provenance via the decentralised and immutable record in the ledger.

Blockchains are a specific type of DLT constructed from structured sequences of blocks connected via cryptographic hashes, providing a tamper-proof ledger that leads to a traceable and auditable log of all activities between stakeholders. In industrial environments, the implementation of this technology facilities greater integration of business processes and stakeholder data, with the blockchain delivering three major protocols: decentralisation, cryptography, and consensus [6]. Due to the censorship-resistant and tamper-proof digital networks of distributed trust created by this revolutionary technology, blockchain-driven technologies help to enhance transactions and make them more reliable and safer. Industrial deployments of the blockchain are still in the early stages of development, and further work is required to establish the full extent of the value the technology offers. However, substantial efforts have been made to investigate its applicability and future penetration in numerous industries, including the industrial sector, as the new technology continues to mature [7], [8]. The transformative potential of blockchain technology in industry settings has already been established in the literature [9], and in the rail industry specifically, blockchain-based applications for ticket sales, invoicing, and freight distribution, among others, have also been investigated [10].

In this chapter we present early findings from the European Union (EU)-funded B4CM project, a study commissioned to investigate the value that blockchain technology offers the rail industry as a ledger of RCM data transfers (section 2), along with a discussion of related work in the literature. The proposed blockchain framework will be presented in depth in section 3, with plans for future work and concluding comments detailed in sections 4 and 5, respectively.

Large volumes of data are generated daily by RCM systems installed on the GB rail network. While this data is already utilised to improve performance within the context for which the system was initially specified, in many cases, opportunities exist for the realisation of additional benefits by sharing this data between stakeholders and across system boundaries, enabling it to be used in problems that cross traditional industry interfaces (primarily the separation between the infrastructure and vehicles). The continuous improvement of system performance through RCM-informed operations and maintenance is a field of intensive research, and many projects focusing on this area have been initiated [11]. At present, the industry is still on an upward performance trend in this area, and localised sensor systems used in isolation are still providing operational benefits. However, moving forward, the industry is expecting these systems to coalesce into fewer, multiparty and sensor environments, essentially evolving the network’s current RCM capability into an “Internet of Railway Things” (IoRT) [12] requiring new ways of managing, processing, and accounting for data. This amalgamation of the state-of-the-art IT, cloud computing, and big data, presented as an Internet of Things (IoT) paradigm, will ultimately lead to a viable “smart railway” fit for the next century [13].

Depending on the nature of the sensors deployed, the data produced by RCM systems takes many forms, including audio, video, pictorial, continuous analogue measurements, and digital signals. In order for the raw datastreams to have operational value, they must first be processed, cleaned, and aligned to the point where they can be reliably used as the basis for analytics. As shown in Figure 1Figure 1:

The six processing levels of ISO 13374. Source: [14].

In this section, the authors present their proposed framework for the audit of RCM data in industrial systems. The framework replaces the TTP typically involved in these systems with a permissioned blockchain architecture, leaving data producers/data owners (providers), data users (consumers), and SC as the key actors in the system. Figure 2Figure 2:

Trust relationship between actors.

In this chapter, a proposed architecture for the delivery of a data audit chain for RCM in GB rail and other industrial contexts has been presented. The next step is for the proposed architecture to be implemented and trialled with real-world data. As there are no one-size-fits-all platforms for blockchain projects, identifying the most suitable deployment platform is critical to the success of this work. A trade-off study was carried out that compared four of the most commonly adopted blockchain platforms: Ethereum [26], [27], Fabric [28], Sawtooth [29], and Iroha [30], based on the parameters set out in Table 1Table 1:

Trade-off analysis between Ethereum, Fabric, Sawtooth, and Iroha.

Of particular interest was the fact that for any SC execution, the Ethereum chain incurred costs (gas) in its native payment currency (Ether), while the Fabric, Sawtooth, and Iroha Hyperledger systems are cryptocurrency-independent, and payment was possible in fiat currencies. Further to this, because of the voting-based consensus algorithms adopted in Hyperledger platforms, there is no requirement for time- and power-consuming consensus algorithms. The associated performance characteristic ensures quick access to provider information, which would be a key criterion for most RCM use cases.

The proposed framework requires differentiation between users to ensure the privacy of their transactions, i.e., not all agreements and payment processes are open to all network users. Any consumer may opt to have a private contract with a provider, and to keep the costs of sharing the data secret from those not participating in that agreement. The Ethereum chain treats all users identically, and all transactions are open and available to all network participants. Hyperledger networks by comparison are able to fulfil this criterion by one of several mechanisms; Fabric, for example, establishes a different channel to isolate parties requiring private agreements and cost allocations; changing the identity namespace in the transaction family on the Sawtooth chain would limit access to specific identities; and specifying guidelines for access management in Iroha would retain easy role-based access at different stages.

In the future, we seek to trial our proposal against representative use cases from GB rail and to evaluate its performance in terms of promoting trust, simplifying cost attribution, delivering a workable payment mechanism for RCM data, and implementing ad-hoc data access agreements between parties. To this end, two representative case studies will be developed, one around the Unattended Overhead Line Equipment Monitoring System (UOMS) and a second around RailBAM, an acoustic axle bearing monitoring system. Both case studies involve systems that require collaboration across the rail sector between multiple stakeholders, and the data generated is of interest to multiple actors, perfectly illustrating the cross-interface scenario that is the target of the system.

RCM is a critical technology in the evolution of the smart railway, enabling improved reliability at a reduced cost. As sensors attached to fixed and mobile assets are increasingly used to inform the operational decision making of the industry, it is becoming critical that the business processes that distribute the costs and benefits of such systems across stakeholders within the industry are aligned in a way that is fair to all parties. The ability to trade in RCM data offers a net market advantage to the industry, as this enables easy access to data by any party that believes they have a use case, while also ensuring that data providers are adequately reimbursed.

Traditional approaches to the management of costs associated with cross-stakeholder RCM deployments in rail have relied on specific business-to-business commercial agreements and predefined costs. These lack the flexibility required to fully exploit the data generated in the “big data” age, where automated model development often requires access to a wide range of data resources from across an industry. Furthermore, the specific use cases being investigated are unlikely to have been foreseen at the time the RCM systems were procured, meaning the initial agreements would need to be modified to support new usage scenarios, an expensive and time-consuming process. Some legacy collaboration arrangements are not wholly defined or explicit and are thus open to misinterpretation or may not be enforceable.

The B4CM project aims to provide the rail industry with an alternative to the traditional model for the attribution of RCM costs. This chapter has introduced a new architecture based on blockchain technology which ensures the rights to data are allocated to the data provider as long as they supply the blockchain network with evidential hash values. The architecture simplifies the mechanism for coordination between a data provider and data users, while also allowing automation of the underlying business agreements and cost distribution. A service quality agreement between provider and consumer is established enabling both actors to prove some violating behaviours; for example, a consumer may claim low service quality, prove their claim, and be paid for; otherwise, for making dishonest claims, the consumer would be fined. Fundamentally, the proposed system allows all stakeholders to contribute, and realise revenue from, their data while enabling cross-industry use cases that are currently not easily realised.

The next stage of the work is to validate the framework by trialling it with real-world industry use cases, and the results of these will be reported to the community in the near future.