Dundee Tunnel Research

Applied Research - Full-Scale Measurements

Alan Vardy has participated in several programmes of full-scale measurements in tunnels embracing main-line railways, underground rapid transit systems and road tunnels.

London Underground Victoria Line

Vardy’s first major involvement in full-scale measurements was in London Underground’s Victoria Line, then newly opened.  A PhD student at the University of Leeds, Nick Higton, who has subsequently had a highly successful career in the railway industry, developed a prediction tool for pressure and velocity fluctuations in tunnel complexes. With his supervisor, John Fox, he won permission to obtain validation measurements during the night-time shut-down period.  The LUL team was led by George Birnie and Ian Cockram and the Leeds team was led by Higton with Vardy acting as supervisor.

The measurements included pressures at various locations in Blackhorse Road Station and in tunnels downstream.  Purpose designed pressure transducers were built for the project, as were optical train detection sensors at two adjacent locations in the tunnel. The train speed history along the route was monitored visually in the train cab and recorded verbally on a tape recorder.  This crude process was necessary because (a) the train speed could not be maintained exactly constant and (ii) there was no on-board speed recording system.

Higton needed the measurements primarily for validation purposes.  Vardy (1980) used them in a study designed to show what degree of geometrical simplification is acceptable when modelling pressures and airflows in tunnel complexes.

Night-time testing Pressure transducer

Tyne Road Tunnel - Flowrates and pollution concentrations

In railway tunnels, the most important aerodynamic phenomena during routine operation are fluctuating pressures and air velocities.  In road tunnels, however, pressure changes are rarely important (an exception being their influence on suspended wall and ceiling panels) and attention focuses on pollution from sources such as vehicle exhausts and tyres.  In both types of tunnel, of course, much attention is paid to the possible consequences of critical events such as fire.  It is arguable whether sufficient attention is paid to the prevention of such events, but that is another story.

The (first) Tyne Tunnel, opened in 1967, has a semi-transverse ventilation system over its whole length and two large, twin-duct ventilation shafts, each for both supply and exhaust.  The tunnel passes under the River Tyne and it has only two lanes.  Until the opening of the second tunnel, these have been used for two-way traffic - although the opening of the second tunnel will change that.  At the time of construction, the ventilation system was state-of-the-art.  Over time, however, concern arose over the possibility of high pollution concentrations, especially close to null points in the longitudinal flow distribution.  A programme of measurements desired by the Tunnel management and its consultant Mott MacDonald was undertaken by the University of Dundee.  It was led by Bain Dayman with extensive support from the tunnel staff.  Dayman is acknowledged elsewhere in this website as a truly outstanding experimentalist.  His availability for this project was highly fortuitous; his great motivational, organisational and planning talents were critical to success.

Dayman devised a trailer that could be towed through the tunnel during either routine or non-routine operation, taking continuous measurements of pollution concentrations and temperatures.  The trailer was constructed by the tunnel staff and made many journeys back-and-forth during a wide range of traffic conditions, daytime and night-time. Most of the detailed data are tunnel-specific, but the methodology was novel and of general applicability.  Also, Dayman’s insistence on measuring temperatures was rewarded by clear correlations between temperatures and pollutions.  Although the correlations did not fit a simple one-to-one pattern, they were sufficient to open the possibility of using simple temperature measurements in conjunction with computer simulations for pollution assessment purposes.

In tandem with the pollution measurements, Dayman and Vardy made extensive measurements of airflows along the tunnel and along its associated ventilation supply ducts - see Vardy (1992).  These measurements reinforced Vardy’s (low) opinions about achievable accuracy of flow measurement in typical tunnel configurations.   Access for whole-section measurements is not possible in the traffic space, of course, but many observers assume that measurements can be made accurately in supply ducts.  In reality, however, access was not possible to the few locations where well-developed flows might exist so measurements had to be made in locations with disadvantages such as proximity to upstream bends.  Also, since the ventilation configuration was semi-transverse supply, measurements along the supply duct were necessarily in a region of spatially-varied flow.  Furthermore, at times of greatest practical interest - i.e. during routine operation - there were significant traffic-induced fluctuations in the supply duct.  The practical problems were increased by persistent failure of an ultrasonic line-average velocity sensor to achieve the consistence of performance claimed by the manufacturer. [It is only fair to note, however, that this was early days for such sensors; the reliability of such instruments has increased greatly since that time.]

The strong emphasis given on this web page to difficulties encountered in the measurement of flow rates in operational tunnel environments necessarily reflects Alan Vardy’s personal experiences as well as his academic background.  Many experienced engineers and academics simply do not believe that Vardy’s views on this subject are valid.  Nevertheless, DTR recommends strongly that measurements of flowrates at full scale should always be regarded with greater scepticism than measurements of pressure fluctuations.  Whereas the latter might often be accurate to within ±5%, it is unlikely that the former will be as even as reliable as ±15%.

Tyne Tunnel - northern end CO monitors on trailer CO sensore - calibration
Tyne Tunnel - northern end CO sensors on trailer CO sensors - calibration

Stanton Tunnel - Airflow measurements (aborted)

A greatly under-played difficulty in the prediction of aerodynamic disturbances in railway tunnels is the almost total lack of reliable measurements of flowrates in real tunnels.  Pressures are measured routinely and it is widely assumed that close agreement between measured and predicted pressures will imply equally close agreement between measured and predicted flowrates.  This is a very poor assumption.  Suppose, for instance, that a mean velocity of, say, 2 m/s exists along a tunnel before a train arrives. This will influence the resulting pressure measurements, but its existence will usually be unknown and the measurements will usually be interpreted on the assumption of zero initial airflow. This is one reason for the prevailing uncertainty about train parameters such as its effective aerodynamic area and its resistance coefficients.

Airflows along tunnels are dominantly uni-directional, but they are not uni-dimensional. That is, the principal velocity component at any position sufficiently far from portals, shafts and trains is parallel to the tunnel axis, but the absolute value ranges from zero on the tunnel surfaces to a maximum at some location away from the wall.  It is quite difficult to predict the lateral variation accurately even in a steady flow along a tunnel, let alone in the highly unsteady flow regimes induced by trains.  In most measurement programmes, no attempt whatsoever is made to record flow rates.  Occasionally, velocities are measured at specific points and “assumed” lateral variations are used to estimate cross-sectional mean velocities. A better approach involves the use of ultrasonic sensors that can measure the mean velocity along a line of sight across the whole width of a tunnel.  [NB: do not fall into the trap of regarding a line-average velocity as a cross-sectional average velocity, whatever the proponents of this methodology tell you!]

Long before line-average ultrasonic methods were available, Alan Vardy led a team from the University of Leeds following an invitation to participate in a test programme planned by British Rail Research in a disused railway tunnel.  BR agreed to allow the installation of a vertical pole close to the centre-line of the double-track tunnel, but offset sufficiently to remain outside the train’s structure gauge.  The pole was duly designed, installed and instrumented.  Unfortunately, fate then took a hand.  A locomotive failure on a routine train caused the railway operators to divert “our” locomotive for revenue-earning purposes so there was no testing at all on that day.  It was not possible to shift the whole test programme by a day and the following days involved the use of both tracks.  Vardy’s team had to remove their instruments and to accept that yet another test series would need to be assessed on the basis of pressures alone. The team accepted its fate with good grace and in good spirits, but there is little doubt that the loss of reliable flow measurement was a more important set-back for tunnel analysts than was recognised at the time.

TRANSAERO Project - Terranuova Le Ville Tunnel

In the late 1990s, the European Union provided several millions of pounds in support of the TRANSAERO project, an extensive collaboration of railways and universities studying various aspects of railway aerodynamics.  DTR’s main contribution to the overall programme was the development of software predicting the rate of steepening of wavefronts in slab-track and ballast-track tunnels.  However, DTR also assumed responsibility for determining the preferred locations of measurement points for pressure transducers in the tunnels and for setting minimum standards for the instrumentation and data acquisition. This was done independently for full-scale and model-scale tests and the reports have subsequently been used in the planning of other full-scale tests in which DTR has had no formal involvement.

An initial objective of the overall programme was to measure wavefront steepening in a long, slab-track tunnel.  In practice, however, the railways were unable to identify any tunnel that was both suitable and available.  Instead, the tests were undertaken in the Terranuova Le Ville Tunnel in northern Italy.  Unfortunately, this tunnel is less than half as long as wavefront steepening tests really require and it has a partially ballast track.  Either of these characteristics would normally have deterred the railways from choosing this tunnel for the tests, but pragmatic decisions are sometimes necessary.

In the event, the Italian railways proved up to the challenge.  They proposed that some tests should be undertaken with two trains travelling side-by-side at the maximum permitted line speed of 250 km/h, thereby greatly increasing pressures (and rates of change of pressure) in comparison with single-train journeys.  This idea might seem obvious with hindsight, but it was truly inspired at the time.  Moreover, the skill of the engineers and the train drivers in actually achieving the simultaneous nose-entry of two trains deserves great admiration.  We measured our wavefront steepening! 

High-speed trains - side by side  

SBB Bahn 2000 - Airshaft flowrates

One of the most enjoyable full-scale test programmes in which DTR has participated took place during the commissioning phase of a new railway line in Switzerland in 2004.  The commissioning included pressure measurements on-board trains, with the aim of determining whether pressure comfort criteria were satisfied during train passage through tunnels.  These measurements were the responsibility of HBI AG who had designed the tunnels using ThermoTun.  SBB simply wanted to know whether compliance had been achieved (it had), but DTR saw an opportunity to obtain measurements of pressures and airflows simultaneously.  In this instance, the required airflows were in pressure relief shafts so the possibility existed of installing multiple velocity sensors without interfering with trains.

The airflow measurements project is related in some detail in this website and in Vardy & Hagenah (2006).  It is difficult to convey in full measure the astonishing extent of cooperation, assistance and pure human-kindness that poured into this project from the railway, the consultant, equipment suppliers, funding agencies, government bodies, research institutes and university technical staff.  On top of this, we were treated wonderfully by the Swiss farmer’s family in whose territory we encamped around the airshaft, by the hotel and even by vehicle-hirers. Vardy had the special joy of working professionally alongside his son Malcolm who, through Vord Consulting Ltd, provided an acoustic monitoring capability, masterminded the use of the instrumentation and data acquisition and took more than his fair share of the extensive joinery and labouring tasks needed to provide a substantial housing for controlled flow measurements on top of the airshaft.

Vord Consulting Ltd  

Selected References

Dayman B, Vardy AE & Evans K (1991) Aerodynamic investigations of the ventilation system of the Tyne Tunnel.Proc 7th int symp on the Aerodynamics and Ventilation of Vehicle Tunnels, Brighton, UK, 27-29 Nov, BHR Group, 583-627

Hagenah B, Reinke P & Vardy AE (2006) Effectiveness of pressure relief shafts - full-scale assessment, Proc 12th int symp on the Aerodynamics and Ventilation of Vehicle Tunnels, Portoroz, Slovenia, 11-13 Jul, BHR Group, 379-391

Schulte-Werning B, Grgoire R, Malfatti A & Matschke G (Eds) (2002) TRANSAERO - a European initiative on transient aerodynamics for railway system optimisation, Springer-Verlag, Berlin Heidelberg

Tokeida H, Mori E, Yamada S, Vardy AE, Mitani,A & Yokota,M (2006) Full-scale test for verification to put MPVC into practical use for tunnels with the concentrated exhaust system at portal, Proc 12th int symp on the Aerodynamics and Ventilation of Vehicle Tunnels, Portoroz, Slovenia, 11-13 Jul, Ed: AE Vardy, BHR Group, 723-733

Vardy AE (1980) Unsteady airflows in rapid transit systems. Proc Inst Mech Engrs, 194(32), 341-356

Vardy AE (1992) Flow measurement in large complex ductwork, Proc int conf on Piping Systems, Manchester UK, BHR Group, Kluwer Academic Pub, 299-310

Vardy AE & Hagenah B (2006) Full-scale flow measurements in a tunnel airshaft, Proc 12th int symp on the Aerodynamics and Ventilation of Vehicle Tunnels, Portoroz, Slovenia, 11-13 Jul, BHR Group, 343-357