Scott McIntosh considers the future – and past – prospects for using ‘off the shelf’ components in urban rail.
It is an unfortunate truism that in the modern world of bright-eyed enthusiasm, a lack of historical context and political desire ‘to be seen to do something’ can lead to decisions being made on the basis of fashion, even in industries such as transport that should be based upon science and calm deliberation. The enthusiasm for COTS (Commercial Off-The-Shelf) is one such fashion. In theory COTS items are products which are commercially available to the mass market with published pricing and specifications. Such products are expected to be used ‘as is’ and equivalent components based on industry standard specifications should be readily available from multiple suppliers at competitive prices for immediate delivery. It is claimed that published pricing and specifications facilitates the purchase of an item at market pricing and reduces the need for formal competitive bidding with the associated costs and time delays.
Given the comparatively small size of the railway industry – and the even smaller sub-sector occupied by tramways and metros – an argument can be made that it no longer makes economic sense from a time, risk, or cost point of view to design a new product from scratch for a specific application. Now, with the opportunity to take advantage of potentially multi-use COTS products manufacturers cease to be makers and become system integrators who can utilise off-the-shelf hardware, components and operating systems to meet their needs. In theory (at least) integrators can focus their efforts on the task-at-hand rather than having to design, build, and debug equivalent new products.
There are three major areas where this approach is promoted in the rail industry:
• Major vehicle sub-systems (traction packages, braking systems, heating-ventilation-air conditioning and communications)
• Control and monitoring data and operating systems
• Station auxiliary equipment (ticketing, lifts and escalators, informatics)
And while it may be worthwhile examining each in turn, it is important to recall that aspects of this approach are not new.
A little history
The transfer of know-how and indeed whole blocks of equipment into the rail industry has a long history, with varying levels of success.
In the 19th Century railways often pioneered mechanical development and expertise transferred from them to other areas, but the rapid rise of industries such as electrical and automobile engineering resulted in a change of flow. A number of specific examples can be quoted.
• Sentinel adapted its successful steam road lorry powertrain (boiler, engine and brakes) into a range of lightweight railcars that operated successfully on UK and British Colonial railways in the 1920s and ’30s. In many cases the components were taken straight off the Sentinel production line.
• Efforts were made by the noted light rail engineer Holman F Stevens to adapt road motor components (mostly Ford chassis) to provide lightweight railcars; these were less successful, with road-biased parts being
insufficiently robust to stand up to the more demanding world of railway traction.
• Walker Brothers of Wigan were successful in adapting the rugged 153hp Gardner engine, designed for heavy lorries and buses, to power railcars; first for a number of Irish narrow-gauge lines, then for the Great Northern Railway (Ireland) and later for Australia and South America. The result was a rugged and reliable unit that could stand up to light to medium railway use. The design descendants are the Stadler articulated railcars in current use across Europe.
However the biggest success was begun by the work of the Associated Equipment Co. Ltd (AEC). This company had grown out of the London Underground group and by the 1920s it was building most of the buses for the group. AEC realised that its established and successful 130 bhp six-cylinder internal combustion diesel engine that was used in London buses and numerous other commercial vehicles, was capable of powering a lightweight self-contained railcar. The vehicle featured a chassis-mounted motor and cardan shaft drive to the wheels.
The prototype was built and sold to the Great Western Railway, its solid engineering making it a success that led to the GWR building up a fleet of 38 cars between 1933 and 1939. Post-war the same basic concept resulted in development of railcars for the GNR(I), CIE and ultimately the British Railways diesel mechanical railcar fleet. This use of commercial road vehicle engines must rank as one of the longest-lived examples of the use of COTS in the rail industry.
The situation today
Over the last decade, an important trend in the US freight railroad industry has been the introduction of ‘GenSet’ (short for ‘Generator Set’, or sets of engines turning generators/alternators) locomotives to reduce fuel consumption and air pollution. GenSet technology replaces the large diesel engine and generator found in almost all existing freight locomotives with two or three much smaller diesel engines and generators. These smaller power units can be large lorry engines or off-road diesel engines and are already designed to meet appropriate Environmental Protection requirements. Advanced computer technology allows for precise ignition control, starting and stopping only as power is needed; fuel consumption and exhaust emissions can be significantly reduced by using smaller engines only when needed.
Some of the US short-line conglomerates are re-engineering older locomotives adopting such technology; using off-the-shelf components in kit form and its own workforce, one company has completely rebuilt older yard locomotives into ‘new’ GenSet locomotives. These rebuilt units are 30-40% less expensive than a newly-purchased GenSet unit. This work demonstrates how the imaginative use of COTS can de-skill a refurbishment/rebuilding project to the point where it is within the capability and financial ability of a small undertaking, railway or tramway.
In the case of first-generation tramways the major car builders adopted the practice of ceasing to be ‘makers’ to become system integrators at a very early stage, with components coming from a range of suppliers. Thus a 1937 Blackpool Brush car has a body by Brush, bogies by EMB, English Electric motors and English Electric controllers. Only the bodyshells were unique: all the other major components were supplied ‘off the shelf’ to a number of systems.
The wider adoption of components from outside the industry was often prevented by space and weight considerations and the rugged operating conditions encountered in a tramway; axle-hung motors have a particularly strenuous life and it was only in the case of the comparatively few cars to have body-mounted motors that something drawn from the wider electrical industry could be considered, as in the Adtranz/Bombardier GT6N/8N cars and in the commercially-undeveloped UK City Tram.
This approach to integrating ‘virtually standardised’ components has continued with modern tram manufacturers and whilst it may not meet the purist’s definition of COTS it has certainly helped to speed the development and production of new vehicles, reduce their cost and improve acceptance time. Specialist manufacturers are particularly strong in the fields of braking (hydraulic, pneumatic and electro-mechanical), suspension and auxiliaries (sanding, windscreen washing, door operation, lighting and ventilation).
At the same time, many large manufacturers have developed standard ranges of bogies for the different classes of vehicle they offer, (e.g. the Bombardier FLEXX range) that are available, virtually off the shelf for a rail vehicle. The one problem with these in-house off the shelf components is that the manufacturers tend to only offer them for products that they assemble themselves – don’t expect to see Alstom or Siemens cars running around on Bombardier bogies – or vice versa – at any time in the near future!
The pros and cons
Proponents of COTS claim that its adoption can meet the ‘faster, better, cheaper’ goals of leaner government procurement and the highly commercial short-termism of public/private partnerships. COTS is meant to encourage the utilisation of the best available dual-use non-development products from commercial/industrial suppliers in lieu of specialised industry components to the maximum extent practicable, including both component- and subsystem-level COTS. In the US (an early adopter of COTS), it is claimed that this approach has produced major benefits to the Defence Department, including lower costs, reduced development time, enhancement of the available supplier base and the early adaptation of state-of-the-art commercial products to maintain technological superiority.
However, not everything is as good as the proponents may claim. Undertakings such as defence and public transport often have very demanding environments; customers and operators rightly expect high levels of reliability and components have long working days and are expected to have long working lives. For example, how many people in Sheffield travel home in a 1992 Supertram and then sit down to watch the news on a 25-year-old television? And yet we expect 98% dependability out of the tramway – a very high demand on reliability and ruggedness. I well remember one metro complaining to me that the escalators they were being advised to purchase for their stations were three times the price of a locally-produced item; it took some time before they realised that they needed a product that could work 20 hours a day, seven hours a week, whereas the local product was designed for an office and would usually only work ten hours a day for five days a week. They bought the Metro-specification units.
It is essential that a rigorous assessment of performance requirements and environmental concerns be made to ensure low-risk adoption of ‘almost equivalent’ components into demanding transport applications. Many important electrical and mechanical design factors need to be considered, including: functionality, packaging, operating temperature extremes, shock levels and duration, random vibration spectra and airflow. In some cases, hardware may need to be repackaged to meet special spatial requirements or enclosures may need to be modified to survive in harsh, hostile, and extreme environmental conditions. The US has a rather unattractive word for this – ruggedisation; but whilst adopting the concept may produce a component that now meets the purchaser’s long-term needs the additional work may have eroded the cost advantage. There is also no guarantee that anyone will be willing to undertake the modifications on the component in the future when spare parts are needed.
One major area where COTS will offer significant opportunities is in the field of electronics (both system control and passenger-facing) and in supervision and control software. Here the technology is evolving so rapidly and equipment used on European railways is as rugged as it gets in civilian applications.
Requirements for these harsh environments are set out in the European Union’s specifications for embedded electronic assemblies on railway vehicles that require components of comparable resilience to those used in military applications. One European railway standard in particular, the EN50155 (Railway Applications – Electronic Equipment Used On Rolling Stock), provides an example of such a tough almost-military specification. The demanding requirements of EN50155 are used on railway systems all over Europe, an early application being the shuttle trains that provide roll-on, roll-off service through the Channel Tunnel. Nine shuttles serve the tunnel and each is some 730m long and made up of 28 vehicles and two locomotives. The development project team had to design equipment that could withstand the impact of shock and vibration on Le Shuttle rolling stock whilst meeting EN50155 requirements.
The Le Shuttle project team had just six months to design, develop and deliver the train’s control system to production. This tight timeline meant that COTS modules and assemblies had to be used extensively as there simply wasn’t the time to develop custom boards. This rapid production shows a potential advantage with COTS – the result may not be cheap, but delivery can be rapid.
To mangle a famous opening line, ‘The railway is a foreign country: they do things differently there’. One feature of the electronics industry that has become apparent over the years is the difficulty that many project managers seem to have in anticipating the problems that requirements such as Electro-Magnetic Compatibility (EMC) will cause to their project. This seems especially prevalent in industries where large projects with detailed customer specifications are the norm – such as the railway (and more particularly the urban tramway) sector. Requirements are delegated down from prime contractor to sub-contractor and the requirement to meet EMC risks gets lost or ‘forgotten’ in the process. The designer sees a functional requirement at the definition stage and believes it can be met by a commercially-available product; along with the functional requirement, there are a whole range of environmental conditions such as shock and vibration, temperature and consequent heat dissipation, ingress protection, and of course EMC. A lot of the issues arise because commercial products – IT equipment, power supplies, instrumentation, etc – are pressed into service against EMC requirements they were never designed to meet.
These problems can arise at a late stage in the design and build phase when the sub-contractor has signed up to meeting ‘the spec’ that can’t be later reneged upon, whatever the engineering practicalities, without substantial commercial penalties. To add to everyone’s difficulties, the sales negotiators may have bargained away any room for manoeuvre when they attempted to achieve closure of the contract some months – or years – previously.
The result is often an emergency patching exercise. This may be called ‘Gap Analysis’ or ‘Contract Compliance’, but in reality it is a desperate scrambling where the commercial specifications which a product is said to meet are compared with the more stringent project specifications or the railway standards, and the identified ‘gaps’ are filled by extra testing which may show the need for ‘mitigation measures’. It may then be that it is found that the component will never meet the requirements, or the mitigation measures necessary make the product unusable in its intended application – e.g. the required extra filtering might double its size and weight; or the extra shielding might mean no-one could open the door to reach the front panel.
Consequent delays make the project late and over budget, with the contractor ending up playing the all-too-familiar games of ‘Project Manager musical chairs’ and ‘pass the parcel of blame’.
While these issues would seem to militate against the use of COTS, it must not be mistaken that I am arguing against the concept. COTS can certainly help speed up product delivery – as shown with Le Shuttle – and it can help keep the product up to date in the rapidly-developing world of electronics.
A new and (one hopes) better component may be developed during the construction process and if standard hardware architecture is adopted it may be possible to upgrade to the new unit with minimum delay. It may also help the operator to upgrade systems during the long life of railway equipment, whereas a system-specific component could rapidly become an orphan and ultimately a museum piece. However, it must not be assumed that the adoption of COTS will automatically result in a better, cheaper product, delivered more quickly. Electronics engineers may consider railwaymen hopelessly conservative, but that often comes from a career of having to maintain a public service with wonder-gizmo equipment where the wheels fall off.
If one is to get the best advantage out of COTS it is important to have experienced staff who understand the potential issues; involve them at an early stage in the process and give them the time and budget to identify potential problems and ways to avoid or mitigate them long before production runs begin.
Some degree of early understanding of what the stringent specifications mean can save a lot of delay, cost and bad publicity at the back end. Searching for late-stage ‘quick fixes’ can only lead to disorder. As we used to say in the military: Order – counter order – disorder – personal abuse – physical violence.