Pratt Truss Girder Bridges on a curve - advice needed, please.

[NWJ]

Hopefully somebody can answer this query, and apologies for the long preamble...

 

I am working on a project loosely based on Padstow and one of the features I want to include is the famous Petherick Creek bridge just before the station:-

 

This is built on a slight curve, so I am happy that I can do the same although mine is on a much sharper curve:

 

Having constructed the basic bridge using the PECO sides, I am now in a slight dilemma regarding the two central pairs of support cylinders. On the prototype, these are arranged to be parallel with the flow of the river which is logical, and the end of each truss girder is central on a concrete pad on each of the cylindrical support. In my situation, this would require me to engineer a sharp meander to direct the flow of the river parallel to these. These photos indicate the sort of angle that he supports would have to be at if I did this. 

 

Fiddling around this evening, I tweaked the supports to reduce the angle under the bridge and straighten the river a little more. This is more aesthetically pleasing but would mean that the trusses no longer meet at the centre of the concrete pad but rather to one side. The right hand support shows this more clearly than the left in the pictures below.

 

Would this be possible in a prototype situation? The trusses are a scale 22m (72') long and I am guessing that the downwards force exerted by a pair of trusses on this support would balance sufficiently to avoid any potential disaster in a real life situation but I am hoping that somebody with a bit more knowledge may be able to confirm this. 

 

Many Thanks

[uax6]

I'm no engineer, but I would expect that the side trusses would also be of differing lengths to cope with the curve. That would then help to get the ends of the girders on to the centre line of the piers.

 

But, of course they aren't actually sitting on the centre line of the piers, because they will be fixed at one end and be sat on bearings the other to allow expansion......

 

Andy G

[Dunsignalling]

I'm no engineer, but I would expect that the side trusses would also be of differing lengths to cope with the curve. That would then help to get the ends of the girders on to the centre line of the piers.

 

But, of course they aren't actually sitting on the centre line of the piers, because they will be fixed at one end and be sat on bearings the other to allow expansion......

 

Andy G

I tend to agree with this, though I'm not an engineer either.

 

The issue is that a real bridge is tailor-made to fit the landscape and you have tried to follow this approach. However, using off-the-peg bridge components forces one to do things the other way about. To make a model that "looks right" using equal-length trusses throughout, the procedure would be to:

 

1. Set out the alignment of the curve, ensuring that the space allowed for the river is a bit wider than the intended bridge length. I'd say about 3-4 inches in this case.

 

2. Position the centre-span and its supports in relation to the curve and roughly in the middle of the gap.

 

3. Add the end spans, skewed to fit the curve and then the bridge abutments .

 

4. Shape the river and its banks to fit around the completed viaduct.

 

John

Edited August 17, 2017 by Dunsignalling

[34theletterbetweenB&D]

...Fiddling around this evening, I tweaked the supports to reduce the angle under the bridge and straighten the river a little more. This is more aesthetically pleasing but would mean that the trusses no longer meet at the centre of the concrete pad but rather to one side. The right hand support shows this more clearly than the left in the pictures below.

 

Would this be possible in a prototype situation? The trusses are a scale 22m (72') long and I am guessing that the downwards force exerted by a pair of trusses on this support would balance sufficiently to avoid any potential disaster in a real life situation but I am hoping that somebody with a bit more knowledge may be able to confirm this... 

 

Part of the joy of pillar and beam construction is that what works in model form, essentially works full size until factors like the bearing strength of the materials are approached. This is how cathedrals were designed. But what was learned was that vertical lines of force should be put well inside the pillar, and visually in this kind of construction that means each beam end occupying half the pad with the pillar centred beneath. Call me stupid, but what prevents moving the pillars slightly to be centred under the truss ends? Visually that will give the right result.

 

Don't be concerned for the flow of the sloppy stuff, it will not be constantly neatly parallel with the pillars in any arrangement, as most of the time it is turbulent.

 

...apologies for the long preamble... 

It was a lovely preamble, I can see why you fancy this for a model subject.

[Gordon A]

As you are modelling unprototypically sharp curves, I think you will have to accept some alterations that are not in the original.

 

Gordon A

[Dave John]

This is a picture of the CR crossing the Kelvin adjacent to what was Partick Central station. It is a girder bridge with a support formed by two circular stone pillars. The girder ends are equally supported on the pillars. The pillars themselves are not in line with the flow of the river, water just swirls about them. You can still see the bridge on google earth, the angle of the pillars is about 45 deg to the water flow. So I'd say fit the pillars to suit the girders not the water. 

 

 

Sorry about the picture quality, I wanted to capture the Kelvin in spate which tends to mean in the pouring rain. 

[NWJ]

Thank you for all of the replies so far, they have been most helpful. 

 

Andy G, I had not taken into account that the trusses may be mounted on bearings simply because I could not see them in the original drawings:

http://www.northcornwallrailway.co.uk/pics/Viaduct-Drawing-1.jpg

http://www.northcornwallrailway.co.uk/pics/Viaduct-Drawing-2.jpg

However, you are absolutely right in this and I will need to mount the end of each truss centrally on the support.

 

John, you're right in the fact that I have resorted to off-the-shelf products so have effectively created more problems by not making my own. However, I have been attacking the landscape a little today to widen the available valley and am now happier with the overall appearance.

 

34theletterbetweenBandD, as I have already said, the correct approach would be for me to mount centrally and indeed it is more visually pleasing - the reference to ecclesiastical architecture is a particularly relevant point too.

 

Gordon A, in an ideal world, I would also prefer the radius to be less tight. However it is not so noticeable from the normal viewing position. I am sure that you will agree that many model railways are forms of compromise and I will be compromising on this structure in relation to the original, however what I am seeking is something that would look as if it would work as a prototype.

 

Dave John, that photo was the sort of evidence that I was looking for to justify the positioning of the support columns, thank you.

[John_Miles]

If you have unequal length girders, that creates its own set of problems from a structural point of view. They are by no means insuperable but they would put up the weight and cost of the girders. Also, it wouldn't look as good and believe it or not, some civil engineers do care about aesthetics. In a tidal estuary such as that near Padstow, the current velocity is probably relatively low so the forces on the piers won't be too bad. Where velocities were high, cut waters tended to be used - these are just lumps of stone or concrete shaped to smooth the flow of the water around the pier. I am a civil engineer.

[NWJ]

If you have unequal length girders, that creates its own set of problems from a structural point of view. They are by no means insuperable but they would put up the weight and cost of the girders. Also, it wouldn't look as good and believe it or not, some civil engineers do care about aesthetics. In a tidal estuary such as that near Padstow, the current velocity is probably relatively low so the forces on the piers won't be too bad. Where velocities were high, cut waters tended to be used - these are just lumps of stone or concrete shaped to smooth the flow of the water around the pier. I am a civil engineer.

Yes, I suppose that the flow is generally fairly low, and although it is tidal, the velocities in either direction will be fairly constant during the rise and fall. Thank you too for the info about cutwaters - I will remember that for another bridge on another project.

[rovex]

The problem with water isn't so much the force on each pier but the scour that occurs round the base which can wash out the ground from under the pier causing it to collapse. 

 

Again I am no engineer and so will refrain from comment as to whether the arrangement of the supports (with the flow or across it) affects the degree of scour, but presumably the more disturbance caused to the flow the greater the scour.

Edited August 17, 2017 by rovex

[NWJ]

The problem with water isn't so much the force on each pier but the scour that occurs round the base which can wash out the ground from under the pier causing it to collapse. 

 

Again I am no engineer and so will refrain from comment as to whether the arrangement of the supports (with the flow or across it) affects the degree of scour, but presumably the more disturbance caused to the flow the greater the scour.

 

That is a problem that was common to many river bridges. I know from the plans that in this case it was alleviated on the original by sinking the cast iron cylinders over 70' through to the bedrock. I will not be doing this on the model though!

[The Johnster]

Yes, I suppose that the flow is generally fairly low, and although it is tidal, the velocities in either direction will be fairly constant during the rise and fall. Thank you too for the info about cutwaters - I will remember that for another bridge on another project.

 

The velocities of tidal currents depend on the rate at which the tide is rising or falling, but are anything but constant and are at their greatest at the 'mid tide' point half way between the low and high.  Moreover, since tides vary in height between neaps and springs, and are further varied on the day by atmospheric pressure, storm surges in exteme events, and supertides caused by planetary alignments (particularly involving Jupiter, whose gravitational pull is very considerable), the rate of rise or fall and hence current velocity at mid tide can vary considerably in comparison to what has been predicted.

 

In an estuary like the Camel, which is to all intents and purposes dried out sand except for the freshwater steam of the river at low water, this is further modified by the mid tide point on the seaward side of the Doom Bar being earlier than in the estuary and possibly around the same time as the water starts rising up at Little Petherick.  So mid tide at Little Petherick is not the mid point between low and high tide just over 3 hours before the high.  Only the local fishermen know this stuff in detail, and they tend to be secretive if it involves them getting better catches than their mates!

 

But the angle of the piers relative to the water is not vital in such a situation in any case, and I would go for round piers, as these are likely to be the result of sinking caissons into the river sand and mud when they were being built in order to get down to bedrock for foundation; 70 feet was mentioned, and sounds not unlikely.  The actual piers at Little Petherick are pairs of round iron pillars connected by steelwork, the original caissons filled with backfill.  No cutwaters are needed, and the main point is to prevent scour from undemining the piers.  The way the sand and mud moves in an estuary is notoriously unpredictable, and an alteration such as a new bridge or quay miles upstream can radically and unpredictably alter matters downstream as new currents and eddies are set up.  When the piers for the original M4, now M48, Severn Bridge were built in the 1960s, a whole new sandbank appeared a dozen miles away off Portishead!   Moreover the sand and mud moves, more slowly but nonetheless, inside the sand or mud banks, in a way about which very little is known but which can occasionally return ancient shipwrecks to the surface centuries after they were last seen.

 

All of which is of moot relevance; your model seems to me to capture the essence of this bridge very well.  I would follow John Dunsignalling's advice about 'fitting' it into your scenery, and I look forward to the photos of a T9 and a couple of malachite green Bullieds crossing it!