FROM HORSE TRAMROADS TO STEAM RAILWAYS
Like many other primitive tram roads – or plateways as they are sometimes called – the Gloucester and Cheltenham Tramroad offered a smoother alternative to contemporary highways for horse-drawn waggons. It also guided plain wheeled carts by means of upright flanges on the inside edges of its metal running plates, which themselves rested on stone blocks.
A typical train consisted of two wagons ( carrying a four ton load ) hauled by a single horse and inclusive of loading, unloading and stops in the many passing loops along the single line a round trip from Gloucester to Cheltenham would have taken all day.
The Gloucester & Cheltenham Tramroad experimented with a steam locomotive in 1831 – just a year after the opening of the Liverpool and Manchester Railway. Patriotically named “Royal William”, this six wheeled machine was built at Neath Abbey Ironworks in Wales. However, it was so heavy that it kept breaking the iron plates and its boiler expired before it could run beyond the City boundaries.
More importantly, when true railways of the type we know today arrived in 1840 it was difficult for plain wheeled carts running on flanged rails to compete. The combination of steam locomotives, flanged wheels running on plain rails and the possibility of longer routes without unloading pushed the Gloucester and Cheltenham Tramroad – by this time part of the Midland Railway – into the red during the 1850s. An Abandonment Act was passed by Parliament in 1859 and on April 16 1861 the tram plates themselves were auctioned off for scrap.
By using flanged wheeled vehicles running on plain rails supported by wooden sleepers a modern edge railway allowed fast, powerful and relatively heavy steam locomotives to be introduced, outperforming horses at a stroke. Even more fundamentally, flanged wheels running on flat topped rails offered better vehicle riding characteristics due to continuous wheelset self-steering.
At the kind of speeds that a steam locomotive of the 1830s could generate, a plain wheeled trailer vehicle would have a tendency to oscillate against the tramroad flanges on either side of it – leading to track damage and ultimately to derailment. A flanged rail however only relies on its flanges for vehicle guidance in extreme circumstances. For most of the vehicle’s journey it is kept in place on the track by the conicity of its wheel treads.
Resistance to constant forward motion of any wheeled vehicle comprises several components including bearing friction ( reduced in the case of Timken, SKF or other brand of roller bearings replacing traditional white metal axleboxes ) , air resistance and rolling resistance. On a true railway vehicle, rolling resistance is of primary concern and occurs at the “contact patch” – the tiny coin-sized area at which the wheel tread is actually touching the rail.
The contact patch moves laterally along the wheel tread according to the radius of track curvature and experiences a slight elastic deformation of both contacting surfaces due to the high vertical, longitudinal, lateral, traction and braking forces applied to it.
The rolling resistance of a steel wheel on a steel rail is far less than that of a rubber tyre on a tarmac road. That is why a diesel locomotive hauling loaded vehicles on a railway track is more energy efficient than a diesel lorry hauling an equal load on road trailers. It must of course be remembered though that a tarmac road, although not always perfectly smooth, can be successfully built for and used by road vehicles at inclines much more severe than those found on Britain’s main line railways. As such, roads will always have a cost benefiit advantage over railways where traffic is necessary but light.
This fact was practically illustrated on the Paris Metro system in the 1950s when new trains running on rubber tyres decreased noise levels and increased traction, but at the cost of a dramatic rise in energy consumption. Such rubber tyred railway vehicles are now confined to low speed lightweight vehicles such as airport shuttles.
In the early 1960s meanwhile, British Railways introduced continuously welded rail (CWR) on many main lines and shortly afterwards the number of derailments of short-wheelbase two-axle freight vehicles began to increase. Consequently, freight train speeds were limited to 40 mph while investigations were carried out.
The Railway Technical Centre (RTC) in Derby became involved , and discovered that CWR was allowing an increase in the level of a phenomenon known as “hunting”. This action is caused by continuous changes in rail contact position on the wheel tyre during rotation. The profiles of wheel tyres are cone shaped at an angle of 1 in 20 to match the curve of the rail top edge : and as the wheels are mounted on a fixed axle, slight differences in wheel diameter will produce a yawing or side-to-side movement as the wheelset tries to find its own rotational equilibrium.
The reason for the prevention of many derailments in the past was jointed track. The presence of the rail joints every 60 feet had in fact been “killing” the hunting oscillations by allowing the yawing forces to unload and thereby prevented the wheel flanges from climbing the railhead.
It is important to bear in mind that wheelsets are not a tight fit within the track gauge – they have to have lateral clearances between the flanges and the inside of the rails to allow room for negotiating curves, to facilitate continuous “self steering” and to prevent constant flange contact, which would greatly increase rotational friction.
But although there was some “play” in the typical two-axle freight wagon of the 1960s, the lateral and longitudinal movement of the wheelsets was restricted by the axleboxes and vertical suspension system of the period. That had the effect of causing the vehicle body to yaw as train speed increased. Energy transferred into the body was leading to violent hunting, flange climbing and eventual derailment. Light Alloy wagon 21 built by Gloucester Railway Carriage and Wagon Company for the Allied Portland Cement Marketing Group was a technical advance on previous designs, but could still “hunt” when not coupled sufficiently tightly – as the inquiry into the Thirsk disaster of 1967 discovered..
Further research at the RTC revealed another related phenomenon. In the early 1960s – except for slight differences in flange thickness – there had been only one type of wheel tyre profile in normal British use with a fixed conicity of 1 in 20. But engineers found that worn – lower conicity – tyre profiles actually reduced the level of wheelset hunting. Coupled with improved vehicle suspension systems, it was thus possible to significantly improve lateral ride performance.
Consequently, passenger and freight vehicles are now fitted with different profiled wheels according to their operational duties and design speeds. These discoveries helped lead to the development of the Advanced Passenger and High Speed Trains, enabling faster, safer services.
However, practical engineering frequently involves a compromise, and there is always a negative side effect. To control bogie hunting at higher speeds, the rotation must be made very stiff. Therefore large yaw dampers have to be fitted between the bogie frames and coach body. These dissipate velocity energy in the form of heat, thus stabilising the bogie. However, increased rotational stiffness also increases in wheelset tyre and railhead wear at speed.
The tendency for wheelset hunting transmits energy into the bogie frame, which in turn causes bogie hunting. The suppression of bogie hunting means the wheelsets themselves are impeded in lateral movement. Consequently, wheel tyre and railhead fatigue are increased by the increase in rotational stiffness, even on straight track. As the yaw dampers cannot absorb all of the hunting energy, the tyre/rail contact patch absorbs the remainder. This mode of wear is exacerbated on curved track, especially at speed.
When a train enters a curved section of track, it is subject to a number of new phenomena. In the very worst case vehicles could derail due to flange climbing or the track could be physically shifted sideways, causing misalignment. In extreme case on tighter radius curves, trains grossly exceeding the permitted speed can topple off the track and overturn, as happened on the London & South Western Railway at Salisbury in 1906.
FURTHER ROUND THE BEND
Looking back into history, some of the earliest investigative work into vehicle wheelset curving behaviour took place on the Liverpool & Manchester Railway in the 1830s. Engineers and scientists of the day were trying to understand why it took more energy to haul a train over curved track. The answer is, of course, that more energy is required to overcome the resistance caused by wheelset steering forces and flange friction against the rail. It was not until the 1970s that measurement and instrumentation technologies became sufficiently advanced to enable the various theories to be tested on specially prepared railway vehicles.
In an ideal world, curves would be negotiated by vehicles with negligible flange contact. To enable this to happen, axles would need the ability to take up positions radial to the centre line of any curved track rather than remaining parallel to each other at all times. In other words, each individual wheelset would be self steering.
In the real world, however, the self steering ideal cannot be easily realised because increased yaw flexibility results in high levels of hunting at speed, which also causes instability and derailment. So wheelsets are constrained to be parallel, even though this means increased tyre and flange forces.
On tight radius curves, the flange of the leading wheelset nearest the outside of the curve tends to contact the outer rail, and the opposite flange on the trailing wheelset tends to contact the inner rail. This is called “constrained curving”.
Constrained curving is usually accompanied by a high pitched squealing noise, created as the tyre “crabs” on the railhead and the flange contacts the inside edge. This situation producs a great deal of friction and energy loss in addition to increased wheel and track wear( often indicated by arced scuff marks on the head of the outer rail).
Of course, the further apart the wheelsets are, the greater the problem, which is why relatively long wheelbase two axle vehicles such as “Pacer” diesel hydraulic multiple units are particularly noisy when negotiating tight curves. Their wheelsets give what is known as high flange “attack angles” in constrained curving conditions
FAILING TO DANGER
On the larger radius curves found on most main lines, a situation called “free curving” occurs, in which the wheelsets are normally prevented from reaching flange contact by by the treads of both wheels developing a force known as “creep yaw torque”. This force attempts to drive each wheelset towards radial alignment, but is kept in check by the yaw-restraining system on the suspension. An effect of this force is to slightly misalign the bogie, pointing it moderately outwards at the leading end relative to the curve alignment. As a consequence, the leading wheelset always sees the greater curving force.
The effect of higher speed with increased centrifugal force on the rear wheelset is to laterally drive it outwards towards the centre line of the curve. This moderately improves the steering, reducing the lateral force on the leading wheelset, but increasing it slightly on the trailing.
So, if wheelsets were able to freely yaw into radial alignment, tread guidance would provide the steering movement, minimise creep forces and significantly reduce wear and lateral forces on curved track. But this is at present impractical on main line vehicles due to the dangerous hunting it allows at speed.
The resultant increased wheel tread and rail surface wear – the latter known as “rolling contact fatigue” or “gauge corner cracking” – is kept in check through regular maintenance by Network Rail. The failure of Network Rail’s predecessor, Railtrack, to keep abreast of such maintenance was one of the factors that led to the Hatfield derailment of 2001. 91 107 “Newark on Trent” – pictured here- was a sister unit to the GNER engine involved in the 2001 Hatfield disaster.
SEEKING A SOLUTION
Various attempts to solve the wheelset misalignment problem have been attempted, and two significant ones took place in the 1980s. Firstly, the RTC developed the self-steering or “cross braced” bogie, designed to improve wheelset alignment and reduce track forces. Secondly, Italian State Railways modified the bogies of several coaches by fitting them with “independent wheelsets.”
In the Italian experiment, the wheels were mounted on to stub axles fitted with roller bearings so that they were free to rotate independently. The results of this trail – subsequently verified by RTC research – showed that there was a much improved lateral ride and lack of hunting on straight track at speed. However, on curved track the wheelsets shifted laterally outward into flange contact. This was due to the lack of a solid axle connection, as creep forces could not be generated to produce a turning moment and steer the flanges away from the outer rail edge. Centrifugal forces therefore predominated, causing hard flange contact, noisy,inefficient friction and excessive wear.
The cross braced bogie was more effective, providing improved alignment on curves by allowing controlled yawing of the wheelsets driven by tread guidance. Unfortunately, although this was an ingenious design, it was never adopted by British manufacturers, possibly due to high production costs. Similarly, conventional bogies and wheelsets were used under the famous Italian “Settebello” electric multiple unit, pictured here.
TILT – GAME OVER?
More recently, schemes using active ( power driven ) steering of wheelsets by complicated control systems have been postulated. These schemes utilise electro-hydraulic actuators connected to the vehicle axleboxes, thus enabling yawing of the wheelsets into radial alignment when negotiating curved track. The yaw electronic control system is fed with information from sensors and traducers on the bogie suspension components.
Increased cost and sophisitication are involved, not to mention possible fail-safe implications at high speed, so an uncomplicated, cost-effective, practical solution to the wheelset curving force problem still eludes railway engineers, especially on fast main lines.
On very high speed routes such as those carrying TGVs in France, of course, very gradual curves have been very deliberately constructed to eliminate the worst of the forces mentioned above. Even so, the TGVs have to have their wheel tyres re-profiled or trimmed at frequent intervals. In addition, frequent railhead grinding is necessary to remove any surface fatigue cracks and to ensure that correct profiles are maintained.
The general opinion is that tilting trains overcome much of the problems associated with curving forces, but they do not, Tilting vehicles exert the same force as any vehicle of equal weight. Tilting is mainly for passenger comfort and, in many engineer’s opinion, is an expensive luxury for the small amount of reduced journey time. Pictured here is five car Super Voyager 221 115 “Sir Francis Chichester” heading north through Oxenholme on Tuesday 31 July 2007.
Increasing vehicle speeds above the norm on curves will substantially raise the level of lateral force applied to the rails by the wheelsets. The lateral/ centrifugal force is proportional to the speed squared, so a 30% increase in speed means a 69 % increase in lateral track force. This applies to all vehicles, tilting or non tilting, depending on their gross weight.
Essentially, what you gain on the straights, you lose on the curves.