BEFORE you start any research it is vital to understand what the problems are and what your vision is. The programme being put forward by the European Railway Research Advisory Council (Errac) is designed to do this and will guide railway research and innovation efforts up to 2020 and beyond. It will also be effective in winning more money for the railway.

By 2050, rail's share of both the freight and passenger markets will double, so capacity and reliability must double, while costs and energy consumption need to be halved.

DohertyThe 'double' challenge of increasing capacity and improving customer service while reducing costs and carbon emissions - the four 'Cs' - is that there are in many cases conflicting outcomes. For example, doubling capacity will normally just push costs up - especially for infrastructure. Increasing capacity usually makes the railway less resilient to failure which in turn affects performance, while the introduction of new trains can have a massive impact on the infrastructure.

Technological innovation, especially from other sectors such as aviation, offers the railway the opportunity to deliver all the objectives at the same time but there are costs associated with this.

The new railway system is a product of new trains, train control and traffic management, infrastructure, information, communication, ticket purchase, and stations. To imagine we are going to achieve these objectives without real innovation is unthinkable.

Innovation, initially at least, needs a catalyst. The digital railway, which is already a major initiative for both Network Rail and German Rail (DB), is a European catalyst and Shift2Rail is another. The problem is that we are probably 15 years behind so we need to catch up.

The core operating step-change components will involve new concepts. This means using mechatronics on the train running gear, and switches and crossings, including the introduction of self-adapting and adjusting capability (as in aerospace) to dramatically reduce wear and damage which will enable us to move closer to a maintenance-free railway. Train control for convoying, together with train-assured braking will assist in dramatically increasing capacity.

A complete rethink of stations, their design, and features is needed.

Infrastructure also needs to digitise, involving things such as data, measurement, information, maintenance, and machinery. In effect the whole railway needs a digital architecture.

The €920m Shift2Rail programme, which has five innovation pillars, became statute on July 7 2014. The eight founding members, including Network Rail, are named in the regulation while the associate members are currently being selected. We are very keen for some of Europe's major railways, such as French National Railways (SNCF) and DB, to join, and we hope to open calls in early 2016, which will be a massive opportunity for the rail sector.

The infrastructure pillar will seek to solve many of the current problems in areas such as the track substructure, rail and sleepers, and switches and crossings.

Improve performance

Several measures are needed to improve the performance of the track and substructure. Ballast shoulder retention needs to better as the shoulder slope angle is probably too steep for ballast stability. The ballast should be strengthened using fibre reinforcement, but we need to finalise the design and develop the implementation process. Embankments need stabilisation and remediation, including water proofing. Development of a continuous refurbishment process is required, with material re-engineered without removing the track.

Sleeper development has had a chequered history so we need to come up with an improved sleeper design, and refine the ballast/sleeper system design. Weather-resilience is fundamental to the design and in renewal systems and processes.

Why are sleepers longitudinal? Idealised support for the rails suggests different shapes and forms of sleepers. There is an opportunity for combined slab and ballasted track designs, with an under slab to provide primary geometry stability and ballast retention. Slabs could be precast in 25m sections for rapid installation, with drainage and cable provision already built in.

Why not have a combined ballast supporter, noise reducer and anti-derailment device? These things are not necessarily the answer, but we need to look for common solutions.

The ballast should allow micro adjustment of the track geometry while being optimised for ease of long life maintenance and a reduction in noise and vibration.

The work stream also needs to consider the whole system's need for axle counters/train detection, and rail maintenance.

We certainly need to rethink rail design and metallurgy. If the rules of conicity change by applying mechatronics we can rethink rail and wheel profiles. Do we need inclined rails or embedded rails?

New rail steels will be required to take advantage of the new wheel rail interface.

We must reduce and lower the carbon footprint, and take resilience to fatigue to a new level.

We have already reduced rolling contact fatigue to nearly zero in Britain. We should consider multi-use rail surfaces, such as old bull head rail, to reduce carbon demand, while self-healing and lubricating rail steel should be possible by using nanotechnologies.

Asset management in this context includes asset data collection, RCM, asset condition assessment, predictive tools, and intelligent maintenance and renewal.

The digital enablement of our people, and maintenance, renewal and design processes will have data collection, harvesting and utilisation at its core.

Reliability and maintenance are clearly part of the digital railway both in concept and content, and Shift2Rail in IP3 majors on developments in this area.

Energy harvesting will reduce the need for running supply cables beside the track.

We need to automate inspection, data collection and analysis, data harvesting and transfer, and integrate and improve the connectivity of all devices and sensors.

In order to switch to proactive maintenance, we need to integrate non-destructive testing and thermography into automated predictive tools and models, and expand their use to all assets.

There needs to be a radical increase in the use of robotics and automation in the delivery of routine maintenance.

We need to add products and systems which have embedded sensors and RCM capability in IP protocols. The use of mechatronic adaptive systems will avoid significant maintenance activity for mechanical systems and assets that wear. In the future systems will automatically predict and compensate themselves to load and demand, while fully-automatic robotic systems will repair many parts of the railway infrastructure. For example, aircraft engines already have remote automatic adjustment.

By using asset measurement, data and information management, and prediction algorithms, we will achieve through-life management as described in the aerospace and automotive areas.

The fundamental design concept for switches and crossings (S&C) has not changed greatly for many years. S&C is a heavy maintenance burden, with facing point locks tests. Throughout Europe, if not most of the rest of the world, tests are required every 12 weeks, which represents a huge staff cost.

Current S&C design assumes active or stored energy clamping forces and effective assembly mechanisms - tight bolts, huck riveting, and welding - to achieve safe operation. It also suffers from having the highest wheel rail forces applied on the weak unsupported leading edge of the switch blade. S&C accounts for as much as 20% of all railway failures and is therefore a major cause of unreliability. Current S&C design also leaves the open switch blades as a hazard to derailed trains.

There are numerous possibilities to eliminate some of these design flaws. We could improve the inherently failsafe locking philosophy by using technology and machine science already in use in other sectors like aerospace. We could introduce self-adjusting switch blade movement actuators and/or stretcher bars, using embedded and differential detectors to assure safety. Why do we move the front of the switch rather than moving other parts of the switch and in different planes?

Positive train steering capability would radically change the dynamics of the train and the switch interaction. Do we need any moving parts at all if the train has sufficient and safe steering capability? Can we remove the open switch blades to reduce the enhanced consequential risk to derailed trains?

In the future, embedded sensors will enable all assets and functions to be measured for condition and performance. The data would feed into the intelligent maintenance work stream and enable 'predict and prevent' maintenance scheduling. This data and information will be open and transferable across the industry.

Mechatronic adaptive self-adjustment techniques will be used to maintain safety limits and provide continuous availability with no loss of detection. This would significantly reduce the need for very regular testing. Perhaps test intervention could be lengthened to 12-24 months, with remote monitoring providing FPL-type assurance. This could include individual blade movement and gauge calculation, while adjustable stretcher bars using worm screws could compensate for movement and wear. There could also be alternative movement techniques. We need to aim for zero failure to switches and crossings.

Infrastructure resilience

Railways should be the most weather resilient form of transport but they are not. They suffer from water ingress and flooding, snow and ice, and heat. The control and effective management of water and heat in infrastructure is recognised as a design target of Shift2Rail.

To sum up, in the future the ultimate autonomous digital railway will have fully-autonomous trains. There will be virtual convoying with a virtual signal overlap, and we will have a new train architecture. Mechatronics will play an increasingly important role. Trains could be self steering, with assured braking, and the ability to couple and uncouple on the move. Control systems could be self-learning. We might have fixed points with no moving parts, where self-steering trains could negotiate at line speed, and a fully-digitised infrastructure with automated/robotic maintenance.