The International Light Rail Magazine
+44 1733 367610
Geoff@lrtap.com

Understanding the wheel/rail interface – part two

Figure 1. BS 2, Rev 1 (1921) railhead sections 8 and 8C. The latter, shown dotted, for radii less than 45.7m (150ft). Sections 6 and 6C and 7 and 7C differed only in the profiles of the groove bottoms. Subsequent BS revisions, the last in 1944, did not alter sections 6-8 and in 1981 the British Standards Institution withdrew BS through lack of demand. Remaining users by then preferred European standard profiles.

In 1903 the Board of Trade, the then UK regulatory body for (amongst other industries) railways and tramways, helped to create the Engineering Standards Committee, which included a Rails Committee. It met only once, its members recognising that railway and tramway technologies had conflicting needs and thus splitting into the Railway Rails Sub-Committee and Tramway Rails Sub-committee.

British Standard (BS) 2 – 1903 identified complementary pairs of sections, one each for straight and curved tracks. The curved sections had a check of sufficient thickness to survive wear from flange back contact. In 1921, BS 2 – Rev 1 withdrew all five pairs in favour of three replacements; Figure 1 illustrates BS 2 Rev 1, Sections 8 and 8C. Inventors patented several renewable checks: I illustrate one developed by Hadfields, a well-known Sheffield-based switch and crossing manufacturer.

In 1905 Germany introduced a similar series of five matching rail standards, although with a slightly wider groove to accommodate a broader wheel-flange that remains the continental norm. The current Ri range is descended from this German standardisation [5]. German Ri sections have become internationally adopted profiles and 60Ri1 has become the choice for urban tramways throughout Europe and beyond. To my knowledge, none of the 60-plus German urban street tramways use anything other than 60Ri1, although a few heavier metros with thicker flanges, normally confined to off-road operation, lay 59Ri1 where occasionally their tracks stray onto paved thoroughfares that, of necessity, are relatively straight. A composite drawing illustrates the difference and, coincidently, the available sacrificial metal embodied in 60Ri1 section.

Check and girder rail design

In 1885, the Board of Trade directed that any railway track carrying passengers, or adjacent to such a track, of radius ten chains or less (200m) must have a check rail. A 21st Century version requires all passenger carrying railway lines curving at 200m or less to have a continuous check rail, engaging with the back of the wheel-flange.

Curves on second-generation British tramways are typically up to ten times as sharp as heavy rail curves. Regardless of whether plain rail suffices for tangent tracks, experienced designers invariably specify grooved rail for such curvature for exactly the same reasons as for railway practice, because acute bends in urban streets require low rail flange-back guidance.

Tramcars frequently need lateral force available only by matching wheel and groove contours. Transmitting thrust through a large surface, rather than point contact, reduces wear, which demands that girder rail gauge-faces have a pronounced downwards incline.

Unlike railways, the track gauge is narrower at the bottom of the gauge-face. German Normalprofil profiles all had symmetrical V-shaped grooves with sides inclined 1:6 (about 9.5 degrees from vertical), as do their present day derivatives and most other girder rails. My impression is that UK engineers’ lack of understanding of girder rail design differences has led to inappropriate use of 59Ri1. It is no coincidence that the 15mm check thickness of 59Ri1 is analogous to earlier tangent track profiles, 9/16in (14.3mm) for BS 2 sections and the 15mm early German straight track Normalprofil sections 1-5.

London Tramlink initially laid 60Ri1 then, presumably after acceptance of the ‘keep theory’, replaced worn-out curves with 59Ri1 on the assumption that its wider groove would reduce wear. Since the gauge-face must then take all the wear, it effectively halves rail life and, coincidentally, creates a less pleasant ride around curves, as illustrated on Manchester Metrolink.

Engineers to long-established tramways know that interaction between wheel and rail depends upon a relationship between gauge, wheelbase, wheel diameter, flange profile, track radius, rail profile and groove width. The equation appears too complex to define with accuracy and, in practice, four-point contact rarely occurs until slight wear achieves the desired equilibrium. A continental permanent way engineer explained that reality to me some four decades ago.

Imagine wheelsets that are set on girder rail. If the flange profile properly matches the groove then all four flanges share lateral thrust, the inside faces of diagonally opposite wheels effectively spreading the turning force between four contact points. To some extent, the high rail check guides the trailing axle away from centre, forcing wheelsets to align reasonably squarely to the track, both axles then approximating radial orientation. That eases the attack angle and reduces stress on the leading outer wheel-flange. Wheel flanges taper outwards from the tip, meaning that on railways, interaction occurs only after severe wear or if something else is seriously amiss. Properly laid girder rail gauge and check faces wear simultaneously from the outset, with checks on curves being sufficiently robust to accept wear, a fact that anyone can verify by examining continental tramway infrastructures.

As mentioned earlier, Germany has more than 60 urban tramway systems. Most have operated for more than a century, which prompts a reasonable assumption that their operators have a long-developed awareness of ideal design parameters. A typical German wheel profile running on 60Ri1 encounters flange-back/check contact at about 40/45m radii.

British tramway engineers may well benefit from examining why those undertakings have a near-universal preference for Ri60 and, on every French second-generation tramway, a very similar SEI 35G (or related slightly taller checked SEI 35 GP). Observers will have difficulty identifying an urban tramway beyond the UK that has adopted Ri59, SEI 41G or other wide grooved profile.

The North American continent has its own standards, although current policy documents clearly state that gauging should achieve four-point contact.

Nottingham Express Transit (NET) developers chose rail section SEI 41GP, a unique profile used by no other European urban tramway and what I believe is inappropriate application of railway technology on NET. I also suspect that engineers with experience only of heavy railway track specified the embedded specialwork where one track crosses another in Noel Street. It was hardly a ‘state of the art’ design, comprising eight separate pieces fishplated together.

Despite having to impose a 5km/h (3mph) speed limit within days of opening, residents in adjacent properties endured a decade of severe noise and vibration before the operator at last adopted what has been standard tramway practice for more than a century. The new (2013) Noel Street crossing incorporates flange-tip running through railhead gaps, where visual evidence illustrates flange-back contact by both checks.

Elsewhere on that tramway, embedded pointwork similarly causes noise and vibration, indeed the design team made special provision for noise mitigation beneath a crossover outside the Theatre Royal, which did not work. This suggests that only a few years ago British tramway professionals had yet to study either history or what is standard practice elsewhere in
the world.

Recent developments

Around the Millennium, numerous continental tramways began adopting section 62Ri1, a related profile to 60Ri1. It has a 25mm check thickness and a nominally reduced groove width, the latter feature encouraging flange-back contact at less severe radii. Four-point contact occurs from the outset on properly designed curves, as illustrated on relaid nine-day-old track in Gent, Belgium, where both high and low rails immediately experience flange/gauge face and flange-back/check contact. Figure 8. Similar observations in Zürich illustrate flange-back guidance on new first-class trackwork. Does anyone have sufficient chutzpah to dispute Swiss engineering principles?

The probable encouragement for adopting 62Ri1 is that wheel/rail interface of low-floor, multi-section trams on rigid trucks (or at best, almost rigid bogies) is more complex than the long-proven PCC and its derivatives. The former steer less well than conventional bogie trams and appear to wear the high-rail check more rapidly than its low rail partner, a problem that their engineers probably now understand.

Accepted specifications

Common practice is to quote the distance between wheelset inner faces, a measurement known as flange-back or back-to-back, whereas when passing through railhead gaps at crossings the all-important dimension is the back-to-front, the distance between one flange front face and the back face of its partner; any change to this dimension will affect guidance.

Figure 9 depicts wear on ‘A’ railway wheels and ‘B’ tramway wheels. Railway flanges rarely encounter check-railed curves, so alter that crucial back-to-front by wearing principally on the outer face, whereas tramway girder-rails react with both sides of the flange to produce a completely different profile. Effectively a symmetrical trapezoid has the advantage that its virtually unchanged back-to-front allows tramcars to operate with wheel wear well beyond acceptable railway limits.

It is also worth considering the German BOStrab Guidance Regulations (March 2004) [7]. The English translation has an illustration of wear on gauge-faces, wheel-flanges and check. In the translation from German, added footnotes appear to presume that flange wear profiles are the result of worn tracks, a rather unkind (and unfair) inference that German operators do not maintain their infrastructures.

The only rational way to interpret the BOStrab diagram is that Germany, a country having a significant number of the world’s tramways, regards check-flange guidance as accepted practice. Anyone taking issue should include a reasoned condemnation of operators throughout the world who have always relied on the principle.

Summary

I have offered sample evidence from undertakings around the world and would suggest that British engineers involved in tramway development need to choose between data from long-established experts, or newcomers whose first specialism is not tramways. I have attempted within these two articles to show clear evidence of flange-back/check guidance, which I hope makes everyone think carefully about his or her practice. As the renowned US Democratic Senator Moynihan once said: ‘Everyone is entitled to their own opinions, but not to their own facts’.

There is still much to learn. It is now more than 25 years, about half the working lifespan of most people, since the UK began reintroducing tramways. How much longer must we wait before British designers grasp what for decades elsewhere has been elementary?

 

Footnotes:

  1. Ri is an acronym of Rillenschiene, translating directly as grooved rail, a suffix number denoting the weight in kg/m.
  1. www.modernstreetcar.org/pdf/circulator_trackway_report_final_3_30_07.pdf
  1. Regulations on the Guidance of Rail Vehicles in accordance with the German Federal Regulations on the Construction and Operation of Light Rail Transit Systems, BOStrab – Guidance Regulations (SpR) – (March 2004) – translated into English in 2008
    (www.orr.gov.uk/__data/assets/pdf_file/0018/5076/ttgn5-bostraben-main.pdf)

 

Feature originally appeared in Tramways & Urban Transit – August 2015 issue (932).