Edward Thompson: Wartime C.M.E. Discussion

Not only that, but because it is rotating, the axle is subject to a cyclic reversing stress. The relatively large wheel diameter will also increase the angle of attack of the flange on the rail, and a larger bending moment compared with the relatively smaller wheels of other eight coupled locos, all leading to higher bending stresses in the axle.

 

A1SLT undertook extensive analysis and computer modelling of the P2 crankshaft during the redesign process, as well as an exhaustive Vampire study of the pony truck design. I find it difficult to believe that any forces that might come to bear on the crank axle were not identified during these processes.

 

A while back there was a discussion about the validity of evidence, potential for manipulation of data. I happened to be reading Brian Hollingsworth's 'The Pleasure of Railways' in it there is a discussion of timing of trains and of timers, and he points out how easy it is to fake a log. He gives an example of a faked log - I can't work out if it is a fake log he created or a fake log that was published in Railway Magazine. There isn't much way of verifying the veracity of logs ie two different logs - which one is wrong, two the same - accurate recording or collusion? and so on. It is pretty clear that Hollingsworth is not really a fan of timers, generally considering that they miss the point of railway travel (ie too busy counting mileposts to see the bigger picture). Really the only way to possibly deal with this (not knowing what the percentage of logs were badly recorded or faked) as with log books etc is to go for volume rather than the single or small n sample size. This should smooth out the exceptional good and bad outliers. (Although ironically, Allen and Nock specialised in the 'exceptional' run rather than the banal ordinary run).

 

Money makes the world go round

 

Surely management (Townend, Hardy etc) is interested in large n. Given that we have these locos available, how should we roster so as to get the best result out of the squad at our disposal? Data on performance must have informed the balance of work allocated to the different Pacific classes. Relating this to Thompson, an interesting question is how the A2s in their various sub-classes were actually deployed at different stages in their career--1947,1952,1957 etc. Where did they fit in in the hierarchy, what were their relative strengths in operation?

Enthusiasts are interested in small n for the same reason we are interested in who wins a football match on a given day, who wins the Olympic 100 metres final etc. I can remember reading Allen reporting in TI on very disappointing runs. Where's the interest in that? We wanted to read about 112 down Stoke Bank or 70 up.

 

Absolutely, and I think it illustrates how far apart the world of railway journalism is from the world of railway management. The problem is that most writers remain in the shadow of Allen and Nock. Even now you can read plenty of Nock tribute act writers.

IMO a month of unexceptional, efficient, economical and to time runs is probably of more value in assessing a class than a single run or series of runs recorded from the comfort of first class. However, I will also say that if you have no other data then a snapshot is useful as long as you accept that it is a snapshot. The flipside is a month of point to point timings will perhaps be less use than a 'thick description' run of cut offs, regulator settings, firing rate, water consumption, weather etc etc (all the different variables included). In an ideal world, you'd have 30 years of detailed records for class repeating the same duties as well as a variety of duties, I suspect that when it comes to logs of loco performance the best we have is a lot of data of variable quality for elite passenger services and not much else.

 

"Trust but verify." I can't see a downside to releasing the modelling report.

Noel

 

No reason to argues against that. The vertical axle loading on the crank axle is quoted as being a nominal 20 tons 10 cwt which works out at half that, nominally 10 tons 5cwt, on each axle box. However, with 21" bore cylinders on the original and 200 psi pressure applied to them, this would give a piston thrust of 30-odd tons, some three times more than the axle loading. Probably less than that in practice, but still considerably more than the vertical axle loading.

 

Those numbers aren't directly commensurable (they're apples and oranges). The piston thrust is converted to torque (around the rotational axis of the axle) by the web and crankpin; the magnitude of that torque is given by multiplying the piston's force by the distance between the rotational axis of the axle, and the center axis of the crankpin. So, if that's 1.5 feet (somewhat guessing, looking at a picture), that would be 45 foot-tons of torque. The weight of the loco also produces a bending moment, in the section of the axle between the axle box and the wheel; very roughly (I'll explain why accuracy is not so important, below) its magnitude can be given by multiplying the force of the weight by the horizontal distance between the middle of the wheel (like center-of-gravity, the middle is where the total force can be simplified to be acting), and the middle of the axle box bearing (I'm not even going to try and guess that distance, although it looks to be less than a foot).

The reason the exact number isn't important is that the stress patterns in the two cases are very different. Most importantly, the rotational torque in the axle is even, fed into it by the web all across the axle, and then transmitted down the axle evenly (i.e. the force in the material will be the same, roughly, at every point across the width of the axle). The bending force in the outer axle, however, will be concentrated at the inner edge of the wheel, and the outer edge of the bearing in the axle box. The actual pattern will be affected by the deformation of the material, under load; discovering the actual pattern, and the magnitude at the point of the peak stress, would likely require a finite-element model of the whole. (Such tools not being available before the advent of computers, no doubt the old-time engineers had a 'rule of thumb' to use to specify the axle diameter for a given loco weight.)

There's a potential further complication, but alas my knowledge of materials science is insufficient to say much about it [and my local structural engineer is busy this morning; she's not a materials scientist either, but she knows more about it than I do ;-]. That is that the direction is constantly changing for the weight's bending moment - as the axle rotates. The rotational torque will vary in magnitude over the piston's travel, but it will always be in the same direction. My guess is that that is more likely to be important for crack propagation; given the mostly isotropic nature of the axle material, it probably shouldn't be an issue during normal operation.

Given how complicated this all is, I'm amazed at the success rates of the old-time engineers (and draughtsmen, who were the ones who did a lot of the actual design). I gather their guesses were pretty good, and then actual usage showed them where they needed to beef things up (e.g. in the many cases of frame cracking on various loco designs).

Noel

 

I can find no mention of any forces other than piston and axle torsional ones in the P2 design study.

 

The P2s had a 26" piston stroke. Posted to just simplify your calculations.

 

You have to remember that the Thompson Pacifics comprised:

6 x A2/2
4 x A2/1
15 x A2/3

A total of 25 Thompson 6ft 2in Pacifics. There were 15 Peppercorn A2s by comparison. The largest group of Pacifics on the LNER were the A3s (78), then A1s (49), A4s (35). If you treated the Thompson 6ft 2in machines as one class of 25, then they were quite spread out as a class across the LNER, so the perceived "hierarchy" of locomotive classes doesn't quite work here. The Thompson 6ft 2in machines were specifically mixed traffic, and did fast fitted freights with great reliability and availability. Obviously they were all centred at specific depots: the general trend on the EMCL was to try and group similar classes together.

Once Thompson was out of the picture, it is obvious that his Pacifics would become second string due to a) small numbers and b) larger classes of newer locomotives (Peppercorn A1s) coming into service. One can cogently argue that the Thompson and Peppercorn A2s suffered a bit from being the "second string" - the Gresley A4s and Peppercorn A1s were the top link locomotives, but the Gresley A3s were more numerous, and therefore filled in the gaps more readily.

I think LNER, and then BR (Eastern) region history has been littered with writings that extol the exceptional, and don't think too much on the everyday, and the profit making. Which made more money: an A4 doing two round trips of expresses a day, or an A2/2 pulling two fast fitted freights? We actually have some numbers to compare on this (for another book!) and the answer isn't as clear cut as you may think.

 

Considering that Bulleid's Pacifics with the same size driving wheels as the A2s, although nominally mixed traffic, were mostly used on passenger trains, likewise the Britannias, and 71000 was intended as a replacement for an express Pacific, how much difference was there in practice between the smaller- and larger-wheeled LNER Pacifics, either in the duties that they were mainly rostered for or how they performed on those duties?

 

Moreover, were there any significant manufacturing or operational differences between conventional spoked wheels and either flavour of BFB wheel (cast or fabricated)?

 

Cox mentions that tests on LNER had shown no advantage of Peppercorn A1 over A2 below 90mph that was speed limit from civils.
Adhesive weigth was 66 tons for both.A1 104 and A2 101 tons loco in working order.A2 must have been a better deal all over.

 

We do actually have some numbers for those too. The significant thing is that these comparison trials were recorded and noted by some of the timekeepers in their books. I also have a greater number of reports thanks to my time at Kew. The ones which were significant largely put the water/coal consumption as better with the Gresley A4 normally tested against, with the larger grated A2/2 having higher coal consumption (but not by much) in the main, with the A2/1 (V2 boiler type) producing erratic results. The A1/1 tested better than the single chimney A4 but none of the tests conducted had a double chimney A4 rostered.

The reason I stated with such confidence regarding the mixed traffic work is that we have enough data and evidence to show the duties the locomotives were mainly rostered for, and can cross reference this to our availability statistics and engine cards data.

I don't know how you'd measure that with the information or metrics we currently have. Sorry.

No, because fundamentally the evidence we have shows the Peppercorn A2's mileages and availability stats were not as good - by a long way - as their larger wheeled brothers the Peppercorn A1s. The Pep A1s trump everything else on the LNER that's a Pacific, with only one locomotive coming close - 60113 Great Northern.

But arguably as a one off, you'd ignore her and focus on the next largest class that was running significant mileages and that next class in the pecking order was the Thompson A2/3 or a double chimney Gresley A4: for which there were only four in LNER days and the rest were converted to double chimney from 1956 and did better from then on.

 

Do we know if they did same kind of work and was used indiscriminately?
If the A1 was reserved for ligth ,fast and few stop jobs mileage goes up and wear and tear down.

 

I believe they are lighter for the same diameter, which was one reason Bulleid went down that route.

Tom

 

Looking at a shed allocation list for 1955, most of the LNER Pacifics (of all types) seem to be concentrated at the depots along the length of the ECML between London & Edinburgh. A few A3s were on the GC Section and a few A3s at Carlisle for the Waverley route. Peppercorn A2s were the only Pacifics then based north of Edinburgh, suggesting that they had a niche as top-link power on the Edinburgh to Aberdeen route. 11 of the 15-strong class were based at Edinburgh, Dundee & Aberdeen. 3 of the 4 Thompson A2/1s were also in Edinburgh.

 

Simon will say if this is wrong, but I would say the A1s tended to be rostered for the heaviest trains such as the White Rose and the other trains with Leeds and Bradford portions.

I've just read in 'Top Shed' that for a while in the fifties, the down Flying Scotsman was a 34E duty rostered for an A2, I presume as far as Grantham. So there was a bit of glamour,