The Tusenfryd Thundercoaster

Size: px

Start display at page:

Download "The Tusenfryd Thundercoaster"

Bonnie McCormick
5 years ago
Views:

1 The Tusenfryd Thundercoaster Synopsis This paper outlines key aspects of the design and construction of a timber roller-coaster in Norway. The ride was tailor-made for the site and the paper describes how this was done. It provides a description of the various forces acting and how they are carried by the timber supports. Fabrication and construction of the track are also described. Introduction Structural engineering can be fun and the fun is rarely more intense than when designing a roller-coaster. Coasters attract a fanatical following of enthusiasts keen to follow developments, always looking for faster, better and just that extra thrill. Among this group there are those who swear a trip on a wooden coaster still can t be beaten. Woodies are rougher than their steel counterparts, but the sensation of uncertainty, the noises and general clatter all add to the experience. Babtie, Allott & Lomax has worked on a number of wooden coasters mostly for inspection and repair but has designed two in the UK, so its team was enormously pleased when tasked with designing a brand new ride for the Tusenfryd Park, near Oslo. The ride was started in summer and completed before winter so as to be available for the next season. It was to be the park s star attraction so a lot depended on its success. Prior to opening it was tried out by the President of the European Coaster Club (riding at 10 C!), who wrote in his club magazine (First Drop 1 ) I prayed I was in for a good ride. This didn t happen: I had a fantastic ride. The train crests the second hill at a fair rate, but as the train starts its rapid descent, the world is literally pulled out from under you. The profile of the hill is of a parabolic arc, which is the profile the train would follow if the track weren t (sic) there. What this means is that you feel zero g all the way over it. I have never experienced air time like it on a wooden coaster. There were actually times that I had to pull the bar down a few extra notches I have never had to do this before! All accolades are gratifying, rave feedback from an informed client is especially welcome. At Tusenfryd, the 30m first drop is the steepest in Europe allowing cars to reach 24m/sec (54mph). But the coaster s sting is its surprise second drop only fractionally lower than the first and set to provide unanticipated thrills. During their trip, passengers are subjected to maximum accelerations of about 4g vertically and 1g sideways; these being generally reckoned tolerable limits for a fast ride. Overall the track is just over 1000m long and the ride lasts about a minute. Fig 1 gives a general impression of the layout. Track design Coaster design is more than just calculation; it requires imagination to lay the track out on the available site so as to give the ride a combination of sensation and experience attractive to customers. No one has to go on a ride, so customer experience is key. The passenger base may be kiddies or it may be ride-experienced teenagers, two groups who clearly have different needs. Wild and fast rides attract a distinct following of those who will pay to go on through a mixture of thrill seeking, bravado, showing off and a host of other personality traits. The customer base has to be known before the design starts and the Tusenfryd ride was targeted at the faster end of the market. A second major determiner is the site. Rides involve constant changes in track alignment and the designer must take full advantage of available topography. This involves maximising alignment and elevation potential plus looking at the line with an eye to creating an illusion in the rider s mind of what is to come. It s no accident that tracks may be set out to provide a sensation of imminent danger or the sensation of flying off into space: that s all part of the fun. The designer also has a portfolio of experiences to weave into the trip that customers might like to enjoy. Some of these are intangible like sensation, anticipation, thrill. Others are physical qualities that can be measured: acceleration; rate of change in acceleration; speed, positive or negative g etc. This last aspect of negative g is the key to air time or the feeling of floating over the crest of hills or dropping in free fall down a slope. Like structural engineering in general, coaster design is both an art and a science. In preparation for the Tusenfryd design, our client was taken on a series of UK coasters to appreciate their feel and he then picked out certain features to his liking such as bends, crests and amounts of air time. The team s task was to take these demands coupled with the available site and mix them into a ride experience that also met limitations on cost and ride duration. The time factor governs the overall track length. Obviously customers need to have a ride lasting long enough to believe their trip is value for money. But an over long ride may be physically upsetting, plus it limits the throughput and rides are, after all, built to make money. It follows that ride durations are actually quite short and the ride at Tusenfryd lasts just 60sec. This is short enough to achieve the commercial objectives with capacity being about 1000 persons / hour. Track layout must also recognise the passenger needs of boarding and exiting along with the operator s needs for maintenance. Thus all tracks include a station for passenger transfer, whilst for maintenance, tracks include a spur where cars can be driven off for inspection and repair. Velocity profiles Once a layout has been planned schematically, the next task is to compute the velocity profile since that governs everything else. Cars gain speed by being taken to the top of a hill and rolled over. At the top of the hill they have potential energy and at any lower part of the track this converts to velocity through the conservation of energy (i.e. mgh = 0.5 Allan Mann PhD, BSc, CEng, FIStructE, FREng, MICE Doug Dadswell BSc, CEng, MIStructE Babtie Group Received: 04/04 Accepted: 06/04 Keywords: Rollercoasters, Timber, Tusenfryd Thundercoaster, Norway, Design, Construction work Allan Mann & Doug Dadswell Fig 1. Track layout 30 The Structural Engineer 7 September 2004

Fig 2. (Right) Train length Fig 3. (Far right) Track camber too slow too fast average velocity mv 2, where m is the mass of the car and passengers and h is the height lost).

2 Fig 2. (Right) Train length Fig 3. (Far right) Track camber too slow too fast average velocity mv 2, where m is the mass of the car and passengers and h is the height lost). Naturally some energy is lost through friction and head wind and allowances have to be made for this. An analysis of the ride s vertical profile then follows, plotting velocities and accelerations at all points. At the bottom of dips, increased vertical acceleration is felt because the curve s radial accelerations are additive to those from gravity. For structural design, the computation is essential since the combined peak acceleration defines the forces transmitted to the track and thence to the underlying supports. This is one way in which coaster track differs from that on any other elevated structure. In these more standard (horizontal) structures, g is constant and the computation of vertical force simply follows from the mass of the object and its position on the span. In coaster track, vertical accelerations are not constant and will vary approximately: 0.5g < vertical acceleration < 4g. (Lateral and horizontal accelerations also vary. Lateral values are kept ~<1g. Forward accelerations may be 1g and under severe braking may be 0.7g). Tracks are never aligned in straight lines, they are threedimensional curves and must incorporate bends in plan and elevation (steel coasters also have twist). These geometrical configurations give rise to radial forces and require the introduction of counteracting camber. Herein lies one of the first complications. In highway design, a car is short in relation to a road s curvature but on coasters this is not so since the train of cars may be long in relation to any of the vertical or Fig 4. Typical support frames 7 September 2004 The Structural Engineer 31

converted to an average velocity somewhere towards the train middle.

3 horizontal curves. Considering the vertical direction (Fig 2), the speed of the train (which is the same all along its length) is determined by the integration of the potential energy lost along the train length converted to an average velocity somewhere towards the train middle. What this means is that unless the train is all at the same level, the back and front ends are travelling at the wrong speed for their particular height. Thus camber (Fig 3) may only be perfectly correct for the speed for one car somewhere towards the centre of the train. At the train s back and front ends, the camber is incorrect. This is similar to a car going around a motorway bend at the wrong speed. Travel too fast and the car has to be steered into the bend to keep it in lane, drive really too fast and the passengers feel uncomfortably as though they are being thrown out. Of course there are really dedicated coaster riders who have personal preferences for front- or back-end exhilaration. But the significance for structural engineers lies in the need to compute a range of lateral forces at each section of track derived from the velocity profile and from the length and mass of the cars. It cannot be assumed that the camber will be accurate. Rather at each section, there are overturning forces to be resisted both locally and on the supporting frame as a whole. In common with highway design, coaster track shares the need to consider the change of lateral force from zero on a straight to radial around a circle. On steel coasters that ride fast, smooth and sophisticated transitions are incorporated. On a steel track these might include two planes plus twist to cope with the multidirectional pattern of motion. However, on slower wooden track, and recognising the nature of wood, building in transitions is difficult and traditionally ignored. This explains part of the roughness and ride experience. Nevertheless, the designer still has to calculate the full acceleration profile around the track to ensure it stays within tolerable limits for the passengers. Humans can accommodate high vertical g better than high lateral g and the amount they tolerate with comfort is linked to the intensity duration. Instantaneously, 4g may be tolerated but prolonged exposure would be uncomfortable if not unhealthy. Timber design Timber design for wooden coasters relies heavily on tradition. A typical support frame is shown in Fig 4. It will be observed that bracing setting out lines do not coincide at nodes with zero eccentricity; but all rides are built like that. Vertical forces are transmitted directly from the train to the track and thence to strong cross beams supported on uprights. The forces to be resisted laterally are those from the train with radial accelerations combined with those from operating in the governing wind speed. Thus there are normally two wind cases to account for. Firstly the standard site wind speed acting on the structure as a whole and secondly a reduced one known as the operating speed that acts on the increased area of the frame plus the cars. The wind speed restriction is partly for structural reasons and partly to ensure the train will make it around the track back into the station. (Commonly rides are closed when wind speeds exceed 15 ~ 20m / sec). Vertical loads are those from the cars, amplified as necessary to account for the dynamic effects at the bottom of dips. Longitudinal forces also exist related to dynamics, braking and longitudinal wind; these are resisted by bracings to ground. The structural design of any particular frame is straightforward enough but there are many frames at Tusenfryd (about 500) and the geometry varies continuously around the track to provide correct elevation. Frame centres are fairly standard but adjusted slightly to position them closer together where larger vertical forces have to be carried. Since much of the capital cost comes from the need to support track at height, layout ingenuity in utilising the same structure to support different parts of track on the same frame saves money. Figure 4 shows how this may be done. The layout also has to ensure preservation of safety clearances as shown in Fig 4 and Fig 5. Given the number of frames deployed, along with their bracing systems, the final impression of the finished ride is one of a timber maze as shown in Fig 6. Altogether the piece count was about and upwards of bolts were used in the joints. The volume of timber used was about 700m 3. For analysis, the frames are treated as plane frame, pinjointed trusses though in reality there must be some lateral load redistribution between adjoining frames via the complex bracing systems and the stiff interlinking of the track. This three-dimensional arrangement certainly gives robustness, highly desirable since the cars bulldoze their way around the track for 9h / day in season. By analysis, the taller frame sizing is dominated by wind load and the shorter ones by the train imposed lateral forces. Since member section sizing is uniform, any lack of capacity is counteracted by simply providing double members. Frame joints are made by bolting through galvanized steel cover plates using large washers to make sure the bolts do not bed into the timber and shake loose (bolts working loose is a frequent occurrence on most rides and has to be checked in routine inspection). A typical joint is shown in Fig 7. The grade of timber adopted is traditional. For the supporting structure, Southern Yellow Pine is used and for the track itself Douglas Fir. Pine is specified for its strength, resilience and availability in long lengths. Both timbers are reasonably durable (the extent is not readily definable but Fig 5. Safety envelope Fig 6. Maze of timber 32 The Structural Engineer 7 September 2004

Fig 7. Typical joint job (local contractors were just not used to the structural format) and simply cost. As a concession to home usage, all trees felled on site were reused in some fashion.

The need to use Douglas Fir for the track stems from the need to bend the members to specific profiles whilst simultaneously having the material hard enough to sustain concentrated loads from the

4 Fig 7. Typical joint job (local contractors were just not used to the structural format) and simply cost. As a concession to home usage, all trees felled on site were reused in some fashion. there are plenty of examples of old coasters lasting up to 80 years, albeit with some repair). The need to use Douglas Fir for the track stems from the need to bend the members to specific profiles whilst simultaneously having the material hard enough to sustain concentrated loads from the wheels. Track is steel surfaced to disperse the wheel point loads over a wider area of timber, but the timber is nevertheless still subjected to high rolling loads both vertically and horizontally. In the vertical direction, the timber is of course being loaded across its grain and in this configuration performs well. Laterally, the timber takes load from radial forces and is sheared parallel to the grain. Even though protected by steel facings, timber performs less well in this direction and requires periodic repair. The twin tracks are set out to a precise gauge matching the bogie wheels. Each side piece is a laminated section made by nailing successive layers to each other (Fig 8 and 9). A laminated form is needed to permit development of track curvature, which has to be skillfully set out on site. Board sizes were generally 10in 2in and 12in 2in but on bends, sizes were increased to 12in 2in and 14in 2in. Joints are staggered and the nailing pattern prescribed. The track has to carry large vertical and lateral loads with the laminated section acting as a continuous beam between adjacent support frames. Horizontally the track is not actually fixed down, but floats on the bearers largely retained in place by friction but capable of some side slip under the imposed forces. The omission of transitions or even the recognition that those defined may be imperfect requires the track to include a strong lateral restraint strip faced with steel to cope with hammer from trains entering a bend (the resilience of the freely supported track helps). Clearly the beam s sectional strength and stiffness is less than that of a fully glued laminate but better than a collection of individual boards. To gain some appreciation of what these structural properties might be, tests were carried out off site (the properties are, alas, much closer to those of individual boards than to the full composite). Since this ride is in Norway it might be thought logical to have used Norwegian timber and local labour but the labour was British and actually the timber was all sourced and imported from Canada! This was largely because of contractual nervousness about the unusual nature of the turnkey Fig 8. (Right) Laminated track section Fig 9. (Below) Laminated track section Foundations The foundations are simple mass concrete pads and about 1500 were required. As the elevated structure is light but tall and subjected to sideways force, support tension is predicted so bases needed sufficient mass to remain stable. The site at Tusenfryd is mostly rock which tended to make the base sizes governed by practical considerations of buildability at the depth required or by the demands of sufficient mass to prevent overturning. In some cases, bases were anchored into bed rock as a cheaper alternative than casting large volumes of concrete. Trains There is strong interaction between track design and the ride s passenger cars. The car s mass obviously affects the forces imposed and the combined length of the train affects the out of camber forces that result from the train being long in relation to track curvature. The car s wheel base governs the track gauge. It is normal for only the rear bogie wheels to be articulated so there are elements of imperfection as the car moves along between the side stop rails. Moreover at certain speeds, there is a tendency for cars to shuffle along in an unstable manner. This is a common phenomenon on wooden coasters and is a vibrational effect partly governed by the clearance magnitude between the car s side wheels and the guide rails. Such vibration adds impact to the calculated lateral forces at bends and is another reason for adhering to traditional sizings known to work. Work in this sense is the balance between usage and repair since on all coasters significant measures of repair are 7 September 2004 The Structural Engineer 33

Some of these can be seen on the illustrated profiles. Fig 5 shows the safety envelope and Fig 11 shows one of the Tusenfryd cars with its passenger lap bars.

(Left) Cars with lap bars, the anti-rollback device is under the train Construction The strategy for construction was heavily governed by climate.

5 Some of these can be seen on the illustrated profiles. Fig 5 shows the safety envelope and Fig 11 shows one of the Tusenfryd cars with its passenger lap bars. Fig 11 also shows the anti roll-back device using on the starting hill plus the side track installed for passenger emergency evacuation. Fig 10. (Above) Acceleration profiles Fig 11. (Left) Cars with lap bars, the anti-rollback device is under the train Construction The strategy for construction was heavily governed by climate. At most parks, rides are built in the winter months between operating seasons but in Norway this would have meant construction in strong winds, ice and snow and with temperatures as low as 20 C. This was too risky, so the whole ride had to be built in the summer one year early to ensure it could open in the following Spring. Based on cost, largely due to local labour rates, it was found cheaper to fabricate the frames in Canada, dismantle them and ship them over to Norway where they were reconstructed by a UK-based labour force. The timber pieces were detailed in much the same style adopted for structural steelwork. Members were cut to size, drilled and then preserved in a Canadian workshop. Frames were then bundled up for shipping as a unit. On arrival at site, each frame was sub-assembled in a lay down area prior to lifting into position complete. Difficult access over the hilly terrain determined that two tower cranes were deployed to cover the ride line. As adjacent frames were erected in position, they were immediately interconnected by longitudinal horizontal members and by a system of vertical bracings set in a diagonal diamond pattern. This method has been used on many coasters and is seen as the standard way of ensuring longitudinal stability during construction. Fig 12 shows what the ride is all about, the thrill of the drop and customers having a good time. The team The team included John Roberts, Phil Hill, Abdul Ghalbon and Robert Casey. It was their combined enthusiasm that made the ride such a success. Fig 12. What it s all about Contractors Main contractor. Vekoma International BV Vlodrop. Holland Timber erection: Bellingrath Ltd UK. routinely accepted. Two trains were included in the contract, each with 24 seats made up by 4 6 seater cars. Safety The standard issues for ride safety are: keeping the accelerations within tolerable limits making sure the safety envelope is not infringed ensuring the train can be braked by the control system checking that passenger containment is adequate having regard to the nature of the ride. Fig 10 shows one trace of site tests used to check that actual accelerations matched the design ones (this is done by riding the track with accelerometers). Normally, if velocities / accelerations are plotted against track profile, the results are fairly close to theory with maximum and minimum velocities matching track dips and crests. Acceleration values also lie in reasonable agreement with theory but there are local perturbations of short duration corresponding to impacts. REFERENCES 1. Garvanovic, Justin: Tusenfryd s Thunder Coaster report. European Coaster Club magazine: First Drop, Winter Issue The Structural Engineer 7 September 2004

NASA Connection Free-Fall Rides

NASA Connection Free-Fall Rides A free-fall ride, like the one pictured here, lets you fall for about 1.5 seconds. Once the car is lifted to the top and released, the force of gravity pulls it toward the