Talk for The Engineering Club, January 2022

The recording is now available here.

The Q&A is here

There will be the remaining chat questions and my answers soon.

We didn’t get this list completed before the start of the talk! Check back for updates.

Bill’s email address is bill@billharveyassociates.com.

Sign up for the Bridge of the Month newsletter using the pop up form on this page (bottom right). There is an archive of past editions.

Sample models from Clifton can be viewed on sketchfab. The vault 12 model has some annotations highlighting interesting features.

There were a lot of questions in the Chat which I tried to answer in the text doc below.

17:46:03 From Andrew Scoones to Everyone:

Please feel free to start to propose your questions in the ‘chat’

17:47:45 From Nina Baker to Everyone:

Is this being taught to civil and structural eng students?

I don't think so Nina not in any determined way anyway.

17:48:17 From Brian Duguid to Everyone:

Do we need to start training engineers in psychology, particularly regarding cognitive biases? Key issues are behavioural, not technical

I think the most important thing is to start emphasising the issue of honesty. Of recognising when you've got something wrong and knowing how to own up to it. That of course requires a corporate culture that encourages such admission.

17:48:23 From Ciaran Malik (he/him) to Everyone:

Admitting you're wrong can often get you in trouble, has this discovery gotten you in trouble?

Not noticeably yet though I've tried pretty hard. The starting point for me is realising that no one else has done any better.

17:48:54 From John G to Everyone:

Hi Bill, It's very interesting problem. With the viaducts showing movement, do you think that is caused by the condition causing the viaduct to start to move or is it the original construction parameters of the arch which is to blame?

This issue of condition is always raising it's hard. If the condition has deteriorated then it is the result or progressive damage. So the answer is no I think

17:49:38 From Fergusons to Everyone:

 Is ArchieM still valid for use on our load assessments?

Well that is an interesting question. And my first answer is what do you think. what Archie set out to do was to allow engineers to explore the effects of the code quickly and easily. I don't think I have ever described it as an analysis. The thing to be careful about particularly is short span bridges with shallow fill. But Archie still satisfies the code.

17:49:50 From Mike Hales to Everyone:

 Often passed these bridges looking at damage wandering about the stiffness of those spandrel panels and their interaction with the arches. Definitely an illuminating talk!

Thank you Mike.

17:49:58 From Chris Jones to Everyone:

 I am not a structural engineer, but is one of the aggravating factors the advent of larger longer goods wagons and longer heavier trucks ?

I believe so but we don't have good records of the state of the bridges before those big trucks were introduced. My concept is capable of explaining why things changed with big trucks but we have no real evidence think they are the cause.

17:50:13 From Konstantinos Voulpiotis to Everyone:

Bill you are amazing as always. I'm curious on your take on the Engineering culture and business models and how this prevents admitting errors, or even looking for them.

That is a hard one Konstantinos. The culture is very variable and depends more on individuals then on company culture. There are big companies worthy overall behaviour is about chasing money but some individuals pursue knowledge and understanding. I believe the culture in Network Rail is driven too much by financial directors who choose to believe that engineering is deterministic. The poor engineers at the coalface make a list each year of urgent repair jobs and then they are given a budget to deal with maybe 1/3 of them and they have to decide which third. They then have to pass the work over to a framework contractor who employs a consultant to do the design. The pressure to complete the work within the financial year militates against any detailed consideration of what has been planned and whether it is actually useful.

17:50:29 From Joanna Buckingham to Everyone:

Where there are cracks and stitching isn't going help, what are more sensible options to repair/strengthen

Of course this depends on where the cracks are and what is their nature. I believe that at 9 wells the spandrel walls had split vertically about 250 millimetres behind the face. Before you can do anything about that you need to recognise it. longitudinal cracks in an arch can only sensibly be dealt with by prestressing the arch across its width. Under no circumstances should bars be drilled through the arch itself. The horizontal cracks in the spandrel walls present much greater difficulty and frankly I haven't worked out what to do with them yet.

17:50:44 From Tracey to Everyone:

Is the damage more akin to a fatigue issue rather than just a ULS failure

Certainly that is true in a sense. Though I think it's more closely akin too low cycle fatigue such as we use everyday for breaking wire for example. Fatigue begins with changes in the crystal lattice as energy is put into the system. In masonry the cracks formed from the start as cracks and it is working of the cracks that does the damage.

17:51:00 From Richard Lorch to Everyone:

It takes a brave person to admit (to themself and others) that they are wrong. How do you suggest that people overcome their fear of being wrong?

Consider the alternative of keeping quiet and someone dying as a result.

17:51:02 From Derec Hickman to Everyone:

Why not seek funding for investigating this from those most affected and whom stand to benefit the most...… the tenants landlords- Blackstone (Arch Co), Network Rail and TFL?

We have had some funding from Network Rail but it was tide to impossible deadlines. The field measurement required it's more difficult than anything done in normal practise at present. We need to understand patterns of behaviour and that means a lot of measurements of deflections at sub millimetre level. We would love to be able to continue but we lost a small fortune on the last contract and now I am unlikely ever to be able to do field work again.

17:51:04 From Hong to Everyone:

Thanks Bill. Is Network Rail, as the owner of a lot of viaducts, aware of this concern?

 I hammer this home to everyone who will listen. Network Rail have funded some research but what is really needed is a steady flow of money over a number of years.

17:51:26 From Patrick Barry to Everyone:

How dynamically sensitive is this effect? Can the piers and train resonate (at reasonable speeds, and high speed rail)

Those who have examined dynamic behaviour of Archers have found no evidence to suggest the behaviour is any different from static.

17:53:01 From Aaron Moison to Everyone:

Have you noticed any difference in the cracking in arches constructed with cement or lime mortars?

where cement was used on the railways it was Roman cement not modern Portland. So much softer and more porous than modern.

17:53:06 From kevin clark to Everyone:

Re masonry bridge analysis are you suggesting that a qualitative structural assessment is more appropriate than a quantitative one?

Absolutely. What is needed is for properly observant and knowledgeable inspectors to look at these bridges regularly and to record effectively where cracks are and how they develop over time. The current code for recording such things divide the viaducts at the piers. The cracks which I regard as diagnostic are there for recorded as separate pieces. And the records of the piers separate the two faces. So the records are not capable of identifying systematic behaviour centred on the piers.

17:53:29 From Tracey to Everyone:

Marsh Lane bridge generally has quite low fill levels so it isn't really surprising that there are issues with it

No viaduct carries enough fill to make any difference.

17:53:31 From David Hoyle to Everyone:

Lower bound plastic analysis being a poor guide to serviceability states is, I think, something to consider in design of even modern bridges with no masonry or arches involved

indeed, force follows stiffness whatever the engineer may think.

17:55:34 From Mike Hales to Everyone:

Maglev…

I rather like Gareth Dennis's term for this and other types of modern shall we say systems. He refers to them as gadget Bahns. I don't think they will provide anything useful in my son’s lifetime. A system that needs energy to lift the freight as well as move it is going to be expensive.

17:55:49 From Patrick Barry to Everyone:

make rails 10x deeper :) 

Nowhere near stiff enough to stop the cantilever behaviour.

17:57:13 From Tracey to Everyone:

In the past the loads were quite heavy and also there was a hammer effect due to the driving wheels

Yes but there is only one loco per train. When most heavy loads were carried in two axle 50 tonne waggons the axles Were typically 3 metres and four metres apart. For most viaducts that is close to a uniformly distributed load.

17:58:18 From Samuel Djanie to Everyone:

Bill, do you think the complex behaviour of masonry bridges suggests it may be prudent to allow them to fail and replace with other bridge types?

There are probably nearly 10,000 spans of viaduct in London alone. There is no space for diversions. Do you think replacement is viable.

17:58:46 From o.diamond to Everyone:

The problem seems to be prolonging the life of infrastructure that is past its best, but we are reliant on it's continued unaffected use. Better understanding of the models will help in designing new structures more reliably but as you've shown, repair of existing arches is incredibly difficult because stiffness dictates load paths. I think we are destined to continually repair and retrofit these arches forever!

I think it's possible to provide better stiff load paths. But it needs to be tried and tested effectively before it goes into widespread use.

17:59:18 From Peter Laidler to Everyone:

Thankyou Bill. I'm not a bridge engineer, but it's a reminder to question the norms and stay interested!

Understanding of physics is only good until the next idea.

17:59:57 From Mike Hales to Everyone:

Another option must be to change the design of the rolling stock. 

I agree, although it would be necessary to demonstrate that it really made a difference.

18:00:08 From Tracey to Everyone:

On several viaducts we have noticed that damage often start where there is a change in stiffness often adjacent to abutments or a change in span

I would be interested to see photos! But there are massive changes of stiffness between arch and arch + backing and between arch and walls.

18:00:44 From Marcus Beale to Everyone:

As an architect/conservation specialist whose father was a linear programmer in the 1950s-1980s FRS he taught me over many heated and earnest discussions:  we make simplified representations of reality and the only use is if they can predict, so I completely resonate with the comment “all our models are wrong”. This such an important lesson from your talk. Don’t rely on the calculations and you should always have in mind approximately the answer… I loved your reference to Occams razor, a great book by JohnJoe McFadden just published on this which others may find inspiring.

Thank you for the reference I will look for it.

18:01:34 From JakS to Everyone:

Thanks mr. Harvey, as if my nightmares as a bridge engineer weren't bad enough. Just kidding, big thank you for your contribution to the community, its the people like you who brought me in.

Thank you.

18:03:53 From William York to Everyone:

How many engineers are doing something wrong because "that's the way I was taught/ that's the way it's always been done"?  Do we need to set up better platforms for engineers to get advice from the rest of the industry, rather than just asking within their companies?  I think there is an understandable reluctance to write into the ICE or IStructE journal saying "do you think my approach is right?", but might be much less daunting on an Engineering Club online forum...

I interact with many young graduates providing support for the Archie programmes. Many of them have been given an inspection report and a copy of Archie and told to analyse that. When they get into difficulty they ring me presumably because their boss either fob them off or is more threatening. Things get really fraught when I say those dimensions are clearly wrong. I can say that because the bridges were built in imperial units and spans et cetera quoted in metric often do not produce round feet. The usual answer I receive is that they have been told they must accept the report and do the analysis. To me that says the boss does not have enough understanding of the problem in hand.

18:04:43 From Khuram Hussain to Everyone:

Is there any benefit in making masonry structures stiffer by pressure grouting the fill behind the arch? / is the risk of the arch barrel losing its compression worth it?

I don't think routing will provide enough strength to be useful. Digging out the fill and replacing it with concrete might help but would be very expensive requiring the closure of the railway for long periods.

18:04:59 From Derek Taylor to Everyone:

Did the Romans know something that we don't?

I don't think so. Roman bridges were built much more expensively then hours

18:06:31 From Brian Duguid to Everyone:

Another one for any non-bridge engineers to look at on this topic is the notorious collapse of the FIU bridge in Florida, where the brand-new bridge was visibly cracked, and the cracks were growing, but the engineer was still insisting work should continue … because they had a model, and the model said it could not crack. There are some good reports on this online. A similar issue of faith being elevated above evidence, and hence again it was about psychology.

So that we don't think only bridges have problems we should also consider the Sleipner oil platform disaster. The model used for analysis was very coarse grained. T he problem could have been overcome if normal detailing rules had been followed but too much faith was placed in the analysis.

18:06:36 From APARR to Everyone:

Is there absence of this damage where there aren't the heavy goods trains e.g DLR in London (where they run over these Victorian arches).

I have wondered for some time where I might look. DLR might well be a good start.

18:07:21 From Marcus Beale to Everyone:

Yes. Romans were very good, practical, understood flexibility and precision but also the deeper aspects. See Vitruvius, always well worth a read.

indeed,

18:08:02 From Sapna Nundloll to Everyone:

@William York, (elec eng trained here, working as consultant now in ped movement, here out of curiosity/ new to Eng Club) feels like the Engineering Club's big strength is in providing "psychological safety" combined with high quality engineering knowledge, and don't underestimate how precious that combination is.

I certainly don’t!

18:08:43 From Sapna Nundloll to Everyone:

https://hbr.org/2017/08/high-performing-teams-need-psychological-safety-heres-how-to-create-it

Will look at that shortly. Thanks.

18:11:05 From Daniel to Everyone:

Many thanks for the presentation, I found it very interesting. Hopefully the behaviour will continue to be researched and we can gain confidence in arch assessments soon. Perhaps 3D non-linear FE modelling of individual bricks and mortar (or a simplification of this) may be a way forward.

Far too expensive to be useful in the short term. An RA recently spent a year building such a model but a critical element was wrongly dimensioned so the structure failed in the wrong place.

Models like this can tell you a lot. Took a day or so to build the data set and a few seconds to run.

And the equivalent wobbly bridge models.

18:12:36 From Nina Baker to Everyone:

Politically, the London Bridge viaduct is too big to fail so it will have to be replaced in the middle term.

They can be strengthened once we understand how. It’s a stiffness issue. We have to be able to reduce deflections that are only 1mm.

18:14:57 From Vanessa Collins to Everyone:

Are we unable to reform without catastrophic loss of life?

I certainly hope so.

18:15:38 From Aaron Moison to Everyone:

 Have you got a picture of that Crack Pattern @Bill

18:15:48 From Marcus Beale to Everyone:

Historically, no. Look at Grenfell.

18:16:23 From Gary  Willis to Everyone:

nothing from me

18:16:41 From o.diamond to Everyone:

 I think a lot of these Victorian arch bridges are also too beautiful to replace!

Absolutely.

18:16:49 From DAVID WIGGINS to Everyone:

'In order to repair something, you have to understand it' - Bill Harvey, Mic drop

Thank you!

18:17:30 From KC Leong to Everyone:

Is the Finite Discrete Element Modelling FDEM type analysis will take account for this multi-arch / viaduct behaviour?

Yes. But only to answer the questions it is set up to answer.

18:18:06 From Derek Taylor to Everyone:

What is the relevance to the cathedrals?

None really, looking at live load issues. In cathedrals, the issue is creep.

18:20:17 From Nina Baker to Everyone:

I assume that for bridges the same applies to brick lined tunnels?

 The problem I am discussing is about multiple spans carrying live loads. Tunnel issues are different.

18:20:17 From Steve Williams to Everyone:

Bill, given the need to " do less" with the carbon agenda we will need to sweat assets more and more. Listening to your talk this evening it appears there is a case for further analytical research into arch behaviour and effective repairs of faults. IF this could be funded it would need a remit. What would the remit say?

Yes, we need a lot of analytical work, but the sort of simple stuff most academics don’t want to engage with. But we also need physical measurement to test the validity of the models.

18:21:14 From Aaron Moison to Everyone:

@Nina Baker - you'd tend to have much thicker backing/ cover on a tunnel that the bridges, so the loads should be dissipated 

18:21:44 From CAFA Team to Everyone:

Can you address APARR's question regarding whether there is an absence of significant damage on structures which only carry passenger trains.

I don’t really know. In any case they may have carried heavy freight previously. It really needs someone observant and thoughtful to travel the network and look.

18:22:07 From Aaron Moison to Everyone:

@Nina, but yes, I THINK it would apply to an extent as they are a similar type of structure

18:23:30 From Nina Baker to Everyone:

@Aaron I was thinking of some gunnited tunnels.

Scream were heard from the speaker.

18:23:46 From Derek Taylor to Everyone:

How do Gustavino/Catalan vaulting compare?

For buildings. No serious live loads. I think there are some on a USA Metro under highways but surely none carrying railway loads?

18:25:27 From Aaron Moison to Everyone:

 @Nina, I imagine that the gunniting would prevent any lime mortars ability to breathe as it needs to, so... probably? The Scottish Lime Centre might be a good source to check.

They would scream like me about gunniting brickwork.

18:27:22 From Marcus Beale to Everyone:

What do you think of Lutyens’ bridge at Hampton Court ? He said this was the best in London, is this something to do with span ratio?

Beautiful, but not relevant to this discussion, I think.

18:30:20 From Daniel to Everyone:

Perhaps a series of complex FE models covering various spans/arch shapes/backing fill levels could be used to calibrate a simplified model.

Yes but very many simple, understandable models each run in a few seconds.