Doug Hewson hairpin superheaters.

[davewoo]

I recently bought a early stage part built Springbok with a commercial boiler from SRS, with it came four of Doug Hewsons hairpin superheaters, two radiant that would go to the back of the firebox, and two that would stop at the end of the firetubes. I already have a stock of spearhead superheater tubes in both 6mm and 8mm that I made while waiting to be redundant at my last job, we had tons of suitable tube, used to feed SF6 gas to electrical switchgear, seemed a shame to let it go into the skip!

Looking at the hairpin superheaters they seem very small in diameter, and the total csa of four tubes comes out at just less than a 1/4" bore. I notice in Dougs design for his B1 that he employs eight of these superheaters. If I were to go down the route of trying the Hewson ones it seems to me the small bore of only four of these tubes would be a bit marginal and could lead to low steamchest pressure, I'd be interested to know whether I'm miles out in my assumption.

Doug's B1 boiler has eight superheater flues whereas my standard Martin Evans boiler only has four, I'm pretty sure I will use the spearhead superheaters that I have made, but the Hewson design of header and superheater looks interesting, and quite easy to clean the tubes, but I am concerned about the small bore of the superheater elements given their length, my thoughts are that this would cause significant resistance through the superheaters, am I completely wrong? It has been known!!!

I would be interested in peoples thoughts, Many Thanks.

Dave

[JonL]

I guess worries about resistance are why they are usually mounted in parallel. I shouldn't imagine it would be a problem, the actual flow rate through each superheater is probably relatively low. As usual, this is just guesswork from me.

[simon6200]

Doug sings their praises but it would be good to hear from a disinterested party who has them fitted to an engine that is used regularly, and for at least a couple of years. The tube must be light gauge so I wonder how long they last in service.

[kipford]

Dave
Velocity in the tube is a square law on dia, but proportional to area. Hence if you halve the area (8 down to 4 tubes) the tube velocity doubles, which doubles the dynamic loss (velocity head) through the tube, as it is a longer slender tube frictional losses also comes into play. Add in the spear head joint which is about the worst thing you can do to the losses in turning a fluid and you have the potential for problems. I would keep as a minimum the same tube area as per the original design.
Dave

[davewoo]

Thanks, that confirms my rather nebulous thinking on the whole thing, The Hewson superheaters are probably fine if the boiler and steam system is designed for them, but I've decided to stick to the spearhead design I have used before.The boiler was built by Steam Technology and I called in to see them to confirm a few details about the certification that came with the boiler, very helpful people to deal with. They gave me an extremely reasonable quote for a fully welded superheater system incuding headers, so I'm taking the boiler in on Wednesday so they can make me a set. For some reason I've gone from Mr make everything yourself to buying things in, decided life's just too short and I have too many projects. Moving house this year so time for a rationalisation and big clearout!
Thanks for the replies.

Dave

Edited to put the R in superheater!

[tenor]

I don't very often pass by this forum, but thermodynamic boiler design is one of my interests and I have written a fairly complex program to look at all the flows in a loco boiler - See Model Engineer 4584 April 2018 and following. I have been working on the subject (and the program) ever since.
I think the Davewoo has made the correct decision, but not for some of the reasons stated. Firstly, steam flow resistance through the superheaters on a typical Martin Evans design is usually fractions of a psi because the flow velocities are much lower than you would normally design for. That being so, a significant reduction in flow area is unlikely to create a big problem. Obviously, one can go too far down that road though.

Kipford says that halving the flow area doubles dynamic loss. Actually velocity head and flow loss vary as the square of velocity (in turbulent flow which is what this would be), so halving the flow area pushes up losses by the fourth power (roughly).

The real killer is that Doug Hewson's 5/32" elements are designed to go in 3/4" flues. I don't have the Evans drawings, but would bet on 1" flues. So Hewson elements in an Evans boiler would leave too much space. That would lead to the flue gas flowing down the easier route of the flues rather than the tubes, which in turn would mean the evaporation rate would drop away. It is possible to get things wrong the other way and end up with too much flow going through the tubes, leaving the flues starved of gas - which would give you plenty of steam but not much superheat.
The key to a good boiler is to get those flows nearly balanced and end up with the same flue gas exit temperature from both flues and tubes. That needs a bit of juggling with sizes to make it work well. My experience is that if you can do that, the available steam volume, net of condensation losses will be at a peak. So putting Hewson elements in an Evans boiler would not be a good way to achieve that.
I would really like to run both the Hewson and Evans designs through my computer program and see just how the superheats, steam volumes etc. compare. I have the Hewson design from the ME, but I don't have the Springbok design - can anybody help with a scan of the boiler and superheater element design for Springbok, please? If some kind soul can supply, I would be happy to summarise the results on here. I might even be able to come up with an optimum Hewson / Evans hybrid!!!
Martin

[davewoo]

Hi Martin
I can't scan the Springbok drawings, but I have some A3 size drawings that I think came free in M.E, I can copy and send those to you by snail mail if that helps,just pm me your postal address. I would be interested in what you find and it could help others in the same dilemma. I had a little feedback that there has been a 5" black 5 visited our club with only four hairpin superheaters in it and reportedly it goes very well. The Springbok flues are 3/4" bore and the hairpins are a nice fit in them, I took my boiler over to Steam Technology on Wednesday and so I'm now committed to spearhead elements. They are also going to fit two more bushes in the boiler for me so I can have twin water gauges (It's one of their TIG welded boilers) and carry out another shell test and issue updated certification for the boiler. Very helpful people and an interesting place to visit, suffered quite a bit of welding skills envy whilst there!
Dave

[chris vine]

Hi Martin, interesting to have your analytical skills on the forum!!
I once read that with superheaters, you cannot get heat transfer without pressure drop. While clearly there will be a pressure drop in any tube with flow through it (so that makes the statement true!), can you say whether there is a strong link between pressure drop and heat transfer?
I did google it but it is one of those difficult questions for google…

Thanks
Chris.

[simon6200]

Martin Evans specified 7/8 flues for Springbok.

[oldnorton]

Yes, thank you Tenor, some interesting thoughts.

It hadn't occurred to me that if you took out a pair of 'inefficient' large diameter concentric superheaters, and replaced those with one or more pairs of small tubes of 5/32" or whatever, then the now less-full flue tubes would have too much fire gas flow and upset the efficiency of the boiler.

We have been taught that it is the radiant part of the superheater in the firebox that picks up most of the heat. I would be interested to hear any thoughts about the energy that large diameter concentric superheaters might pick up, compared with multiples of smaller tubing.

I have also realised why my double-concentric superheated loco might be showing wet steam at the exhaust, when pulling slowly on a cool day, whereas another small-tube superheated loco has dry exhaust. A wet exhaust means that the steam has fully expanded in the cylinders whereas if the exhaust is dry that steam still holds unexpanded energy. The pressure resistance of the smaller superheater tubes might mean that the cylinders are being filled at a higher pressure? Of course, a wet exhaust might mean that the concentric superheaters are doing nothing useful at all. Any thoughts or comment would be of interest.

Norm

[tenor]

Thanks for the interest in a numerical look at the various options.

Davewoo, I have sent you a p.m. regarding drawings.

Chris V, Yes it is true that heat exchange goes largely hand in hand with pressure drop and that applies both on the flue gas and steam sides of a boiler surface. It is so, because both the heat exchange and pressure drop depend on the thickness of the boundary layer - that dead(ish) layer of fluid attached to the metal surfaces. If the boundary layer is thick (low velocity and low pressure drop) then you have a relatively thick layer of gas which insulates or prevents heat flow. The converse applies. Another way of looking at it is heat transfer in terms of Kjouls/m2/Deg C difference is a function of Reynolds number and the pressure drop is also a function of Reynolds number.
Then life gets tricky, because the obvious conclusion would be to make all the tubes smaller, increase the draught requirement and improve the heat flow. That doesn't work because you lose heat exchange surface and that makes things worse and the improved heat transfer will not offset all the loss. Next step, put a lot of small tubes in which will generally improve heat transfer but it pushes up draught requirement quite fast and is prone to blockage. That was the basic conundrum that led me to put some code together to run all the various numbers and see what was really going on.

Simon6200, Thanks for the flue diameter on Springbok.

OldNorton, Glad to have pointed out the importance of balance in the flows.
The radiant superheater thing is a myth. My program accounts for the radiant (where relevant) and convective heat transfers at all points along the superheater and the rate of heat transfer in the radiant section is not significantly different to that in the convective section. HOWEVER, radiant superheaters do get you a useful bit of extra heat exchange area and that makes a difference.
I have to admit I am not a convert to concentric superheaters and my program will not handle such a geometry (it's a pig to work out how much heat you are taking and then how much you are losing to the incoming steam). So in the absence of having worked the numbers and got to know how they stack up, I admit my opinion is pure bias at this stage.
A point not often acknowledged is that superheat temperature (for most loco type setups) varies with how hard the engine is worked. Superheat temperature is low at low firing rates rising to maximum as the loco is flogged - so showing visible steam pulling slowly on a cold day comes as no surprise. Put a few passengers on and knock off half a dozen laps and you will see a big difference as she warms through!
Broadly speaking, steam pressure drop through the superheaters is not an issue. The numbers I have worked for typical 5" gauge designs give pressure drops in fractions of 1 psi at high steam rates, so I doubt that would be having an effect.

I shall try and get stuck into looking at the two B1 boiler options.
Martin

[uuu]

In a concentric superheater, you are "losing" heat to the incoming steam? How is this heat escaping, please - where is it going, such that it is considered "lost"?

Wilf

[kipford]

Martin, tha is for correcting my school boy error. I know it is a square law on area having worked with fluid flow all my working life. I had a senior moment and convinced myself it was only proportional.
Dave

[oldnorton]

Jan 27, 2024 20:12:31 GMT uuu said:

In a concentric superheater, you are "losing" heat to the incoming steam? How is this heat escaping, please - where is it going, such that it is considered "lost"?

Wilf

Yes, you are right Wilf in that no heat is 'lost' to the outside, but the argument is that the superheated steam delivered to the cylinders will be at a lower temperature than when it started its return journey , because it 'lost' some energy heating the incoming stream (whether this was in the internal or external of the two tubes). Of importance is that the concentric system 'lost out' in the battle against a pair of parallel tubes.

[Cro]

Interesting discussion and for those that have read my A4 thread recently will have seen I have removed the original concentric type to fit spearheads.

I went with 2 spearheads per flue, 8mm OD in a 7/8” (from memory) flue. The original concentric type essentially was 1 8mm so I’m working on the basis I have increased the surface area being heated whilst maintaining the flow of gases through the tube.

Time will tell!

Adam

[tenor]

I have received some drawings from Davewoo today, so will get onto those shortly. I had a first go at the Hewson drawings last night. Provisional finding as yet, but I was getting 30 m/s steam velocity in the superheaters and 7 psi pressure drop in them. The velocity is not too bad, but the pressure loss is worrying.
I got better results with twice as many elements, but not of the Schmidt type. The challenge then is to work out a manifold with 16 wet and 16 dry steam connections that would fit in a smokebox!

More interesting than I was expecting.

Martin

[JonL]

I might be being daft, but will the pressure not drop due to the increased volume of the superheaters compared to the pipework anyway?

[chris vine]

Hi Jon,

Not daft at all, fluid and thermo dynamics must be one of the trickiest things.

The pressure drop comes from friction of the steam flowing along the circuit. If you think of a long train of wagons with the engine pushing from the back, the force at the buffers between loco and train will be large, but as you move to the front of the train, the force between the buffers will be less. Starting from the front of the train, each wagon is adding its bit of friction so there is a lot that the loco has to push against.

Another way to look at it might be to imagine the same steam circuit with superheater and think of only allowing a very small amount of steam to flow through it. Either by throttling at the regulator or at the cylinders. With a small flow speed there will be almost no friction so no pressure drop. Yes, the steam will expand on its way through the superheater, but that just means that less steam will come out of the boiler.

Maybe that doesn't help at all!!!!

Chris.

[uuu]

Jan 31, 2024 17:46:18 GMT JonL said:

I might be being daft, but will the pressure not drop due to the increased volume of the superheaters compared to the pipework anyway?

As noted, fluid flow can be tricky.

In a carburettor, or a steam injector, a constriction in the pipe work - the venturi, or the first cone - causes a drop in pressure. So increasing the cross-section of a pipe can cause an increase in the pressure. Of course, at the other end of the device, if you return to the same cross-section you had before, the effect will be reversed. As happens in the steam injector, where the final expanding cone delivers the water at higher than boiler pressure.

Wilf

[chris vine]

Hi Wilf,

Your example offers another proof that fluid flow is tricky: although there is a pressure drop in a venturi, it has nothing to do with friction. As you say, after the venturi, the pressure will rise again as the cross sectional area increases and the flow velocity slows.

The miracle of all this is that M. Giffard managed to propose and get his injector to work. Despite thermo and fluid dynamics being very poorly understood at the time, he seemed to have a very good feeling for what was going on.

Chris.