Exhaust Modifications for #17

Probably.....and given the small square firebox, getting good air mixing with the atomized waste oil for complete combustion is a challenge.

Nigel has had a lot of success with his swirl inducing firepan designs used with central burners in even smaller squarer fireboxes, and we know that vaproized fuel burns more completely than atomized. His superheating rhe atmoizing steam with a lop of pipe in the firebox probably also does a lot of good - having the steam cool the oil while it blows it out is counterproductive in a vaporizing setup.

So we have the situation where the secondary draft from the door damper bleeds fuel out to combust in the cab rather than adds air to complete combustion within the firebox. I think this does argue for a better front end. What we want is air sucked in to promote combustion where it will do some good, not to let fuel and heat out where we don't want it.

So.....now I'm wondering about how far we can push the combustion efficiency of waste oil towards that of a vaporizing burner by cleaning the fuel and better firepan / burner design.

Dave

Dave- I'd think that the other advantage of superheating the burner steam is that less steam will be required to do the job, just like superheating the steam to the cylinders.  This means there's less steam used (the boiler doesn't have to produce as much) and therefore fewer pounds of steam are blown into the firebox, so the exhaust system can draw more combustion gases to boil water.

Hugh

Agreed, of course. Steam not used doesn;t need to be replaced.

I forgot about headaches. Everybody has them when I'm there, but they recover soon after I leave. I don't think there's a correlation with the combustion problem.

Dave

Dave,
Imho the present order of attack of the problems is a) proper air quantity firebox
b) vacuum to draw that quantity through the boiler c) optimisation of the locomotive effort to arrive at that vacuum.
Oil quality/burners a.s.o. should be kept in "steady state". Unless we start using the Japanese theory on approach to experimentation we better stick to changing one item at the time. If not we will never really understand what is/was happening!
Kind regards
Jos

Thank you for keeping me on the steady course, Jos - I tend to diverge in ways that aren't entirely linear. Which I'm going to do now......

First, our friends from Monticello are very willling to assist us, and will be joining us as soon as hugh can get them set up. They are now where we hope to be 5 years along the line. I also hope we can assist them as they continue to work on tweaking their performance on Southern 401.

Second, I'm starting to write a grant for the purpose of working on optimizing the combustion of cleaned waste oil in steam locomotives. It's going to be a small amount I'm requesting, just enough to cover the cost of the centrifuge, pump, filtering system, miscellaneous piping, making a few burners and some steel and firebrick to build test chambers for whatever we make to compare. I'm hoping to involve engineering students at NC State and machining and fabrication students at Wake Tech.

I look forward to the Monticello guys comments on air supply and adequate vacuum levels.

Dave

The way I've always described firing 17 to people is like this. Imagine the fire we need to run 17 as being the size of a grapefruit, and imagine the firebox as being the size of a matchbox. We're trying to put one inside of the other, that's why the fire keeps shooting out of the box.

The increased draft has helped us a lot since we put the bar on. The stack is running noticeably lighter and isn't having the trouble keeping steam like it use to. I've been able to go to the pop climbing a hill, back off on the fire and inject water to seat it, then go back to the pop while still climbing. 

I should also mention that 17 has an oil heater in the tank and we use it before runs.

Of course decreasing the orifice is a remedy for the problem. However it is a solution that carries a price tag in the locomotive effort. The solution lies in increasing the size of the air holes. If I write down in a spreadsheet my line of reasoning about the calorific value of your fuel, the amount of air it needs to burn and the amount of steam that can be produced I can calculate the pressure needed to get the air through the air valves. It is a whopping 6 inches of water, please repeat the calculation yourself as I am extremely untidy in math! Added to that the resistance of the boiler tubes of which I have no inclination to calculate giving the uncertain temperatures and the composition of the mixture in the tubes you will need a very high smokebox vacuum, imho unnecessary.
Kind regards
Jos

I also wonder if it isn't just the reduction in total opening area, but also the division of a single into a multiple nozzle with an interesting bit of turbulence and broadening of the exhaust stream included.

Robert wrote about air preheaters a while back.....maybe this would be a good time to think about those in conjunction with increasing the intake opening area. I know I don't have a clear vision of them.

Dave

It would be interesting to see what kind of combustion air pre-heaters we could come up with. Something as simple as an automobile radiator fed with exhaust steam to the inlet and a steam trap on the outlet might do the trick.

Hugh

I quite agree with more sophistication, but please one item a the time! #17 should be made a good to perfect steamer in its present state.  Then and only then the next steps with equipment that isn't there right now.
kind regards
Jos

Fine, no argument, just thinking ahead.

So, when we hear from our newest members about their work on air inlet areas on 401, we'll have some basis for looking at the correct proportions for 17. After than, we can better figure out how to achieve the area we need.

Also agree about the detrimental effect of reducing nozzle area increasing back pressure. So, once we have draft areas ironed out, perhaps we can investigate the options for improving the front end arrangement for adequate vacuum.

But, I see no reason not to think further ahead while we're working on the practical matters at hand. I have no expectation that something we can hope to accomplish now would include things that will require a couple years of R&D.

Hugh - would there be any benefit to heating the incoming air with exhaust steam injection rather than through transferred heat? Would moister air be better for combustion?

Dave

No, I think you want to minimize the amount of steam injected into the firebox, which is another benefit of superheating the atomizing steam. Any steam injected into the firebox displaces air that could otherwise be there to support combustion.

Makes sense....just mentally connecting with overfire jets, which were more about forcing air then moistening it. The turbulence was apparently an advantage, but that can probably be optimized through the directing of the air entry ports.

Dave

Hugh, Dave,
Did anybody do the math on air preheating? I used to be a naval architect a long time ago and we discussed intercooling equipment to get as much oxygen in the cylinders as possible. What I mean is that air is heated at the rate of 1kJ/kg.degree K while the burning oil is releasing heat at the rate of 44 MJ/kg for which you would need something like 17.5 kg of air which you cannot get above 100 degree C/273K.
As such you a working on 1.75 MJ near 44 MJ released. Is my math correct?
Kind regards
Jos

Jos,

I've never done the math, but I know Wardale planned combustion air preheat for the 5AT. so it must be reasonably achievable. I'll have to convert all the units before I can check your math! I'll see if I can run some numbers tonight.

Hugh

And if you could please explain how he planned to do it, Hugh....I'm still out in the weeds about how it can be practically possible.

Dave

I searched the 5AT website and couldn't find any details on combustion air preheating.

Steam heat is still used in building heating (HVAC) although it's largely been superseded by hot water.  The basic principal is that low pressure steam is admitted to the top of an air heating coil.  The steam flows into the heat exchanger and condenses as it passes through.  The steam flow is controlled by a steam trap on the outlet.  A steam trap will let only water pass through, not steam.  Thus, it automatically controls the steam flow by limiting it to what can be condensed as it passes through the coil.  Colder air or more air flow condenses more steam, so the steam trap lets more condensate leave the drain in response, effectively matching the steam flow to the required heat.  

Condensing steam has a very high heat transfer rate- remember with no temperature change steam gives up a huge amount of heat condensing from steam to water.

We should be able to pick an off-the-shelf HVAC steam coil based on the desired airflow and heat rise characteristics that could easily be adapted to this application.  Something like this:



I'm thinking we'd provide a short transition duct fabricated from sheet metal that would duct the air from our coil into the combustion air inlet(s).  Providing a couple of smaller coils might be easier to fit than one large one.  Air would be drawn into the firebox through the coils.

Heat added to the combustion air should improve the vaporization and combustion of the fuel oil, and it adds more heat to the boiler as well.  I'll have to sit down and crunch some numbers to see how much benefit we might expect.  Suffice it to say practically all power plant boilers are equipped with combustion air preheating, regardless of fuel used.

Hugh

According to Snyder's patent (Hugh has a copy and might want to post it here; I don't have it but want to say 1935) you don't need the fragile fins or pins. He relies on turbulence between multiple rows of tubes to get the heat transfer (from exhaust or low-pressure steam) accomplished, and any mixing with unheated air accomplished by the time it comes to be drawn up as primary air. I believe some of his logic is described in the patent description.

Heat for the primary combustion air in an oil-burning system is very different from what Snyder was designing. One difficulty with spray burners that use compressed air rather than mechanical fuel atomization is ... exactly the same issue with compressed-air injection in the Diesel cycle: the expansion lowers the temperature so much that it impedes combustion. You would need to put CONSIDERABLE preheat on the compressed air before using it in the pressure burner. That might be done with concentric preheat coils around the burner flame, a bit like the burners used for hot-air balloons (there is sure to be a technical term for this kind of burner but I don't know it).

The primary air for something like a von Boden-Ingles would have exchangers of appropriate characteristics on the openings, with appropriate flow ducting at the 'edges' -- probably optimizing smooth flow into the exchanger bundle at all draft levels.

Secondary air preheat is a bit more tricky, as the pressure to get the secondary air where it is needed is done either with compressed air or steam, and in either case secondary-air heating will become a bit complicated.

Keep in mind that powerplant boilers have 600 degrees of preheating, which is easy to accomplish ... on a stationary boiler. You would need some sort of Franco-Crosti arrangement to get that kind of air temperature effectively on a locomotive, and I suspect you would need a much larger boiler length than on 17 to achieve the necessary gas path length to get effective transfer. Worth considering, I suppose... but there will be a lot happening under the boiler if the exhaust is to wind up back at the stack where people expect it to be ;-}

OK, let's say we're installing some form of heat exchanger coil across the air inlets going to the firepan, and running a portion of the exhaust steam through them......I'm going to assume here that the temperature of the exhaust steam isn't much higher than 212F and the efficiency of the transfer isn't real high either.....so maybe we get inlet air temps about 60F or so higher than if we didn't? If I'm underestimating the practical impact of the idea please set me straight.......

And if I'm not, then perhaps the exhaust steam can be better used for making vacuum or, if we can spare it, preheating the feedwater more efficiently instead?

Dave

The Snyder patent is #2,096,439, filed in 1934, issued October 9th, 1937. I believe Lamb has some discussion of the idea.

When I was first exposed to this design, I thought heat transfer from the size of pipe involved was, shall we say, a bit optimistic. On the other hand, I have heard from reasonably reliable sources that when this design was tested on C&O, its use produced nontrivial savings in fuel and, I believe, water.

A substantial advantage of this design, if it works as reputed, is that it is made of robust ordinary materials, with no fancy fins or need for modular construction, etc. I am sure that there are much more well-developed heat exchangers suitable for air preheat -- but I have to wonder what it costs to fabricate or buy them, and what it will cost to keep them maintained. (I have spent my time with the fin comb, especially after hailstorms, and would not wish that job on a dog...)

Robert- I sympathize about the fin comb!  You'll appreciate it that I ALWAYS require hail guards when I specify HVAC equipment.   

I had forgotten I had found that patent; here's a link for anyone that's interested: http://www.google.com/patents/US2096439

Not having fins would definitely make the things easier to construct and maintain.  I'm reading the patent again...

Hugh

Gents, I had a very interesting visit to the NHVRy last weekend, thanks to all concerned for their excellent hospitality.

As a new cab denizen of #17 I found the blowback of flame into the cab somewhat alarming, more so when there were at one point five of us aboard and escape from the risk of burning apparel below the knee was near impossible. That aside, everyone assured me that the loco was steaming better than it had prior to the addition of a crossbar to the blast nozzle and a new petticoat pipe.

I have been reviewing this #17 drafting thread and applying some new numbers to my calculator sheet, here is what I have.

If the as-built stack and petticoat is assumed to behave like a lempor mixing chamber/diffuser (I know it probably doesn't but take it as an input assumption), as-built it has an in/out area ratio of 2.64:1.

I took the steam consumption to be 1,355 lb/hr (calculated at 7.5 mph and 50% cut off) no special allowance was made for atomisation steam or air pump and blower usage.

The choke diameter I set at 10" to match the existing and then adjusted the requested vacuum to get a nozzle area that matches the existing one but *with two 1/2" crossbars* (effective area 8.84 sq") rather than just one. The sheet results look acceptable at 9.2 inches of water vacuum. This is playing silly-buggers with the calculator in a sort of semi-reverse engineering way but why not?

Since we have no existing smokebox vacuum numbers with which to compare this exercise means nothing. However, the consensus amongst the #17 crew is that things are much better after the addition of the first crossbar so my recommendations are:

1/ Measure the existing vacuum under load with a train.
2/ If 9.2" is seen as a potential improvement, add a second bar to make it a cross (but please use square bar not round, the corners create useful turbulence), leave everything else alone and measure the resulting vacuum.

An easy way to do this might be to laser-cut a plate, weld it to a pipe to fit over the blast nozzle and add three radial bolts (weld nuts to the tube) to clamp it into place.

Michael.

Thanks Michael, wish I could have been there. I'm sorry our volunteer didn't get the manometer installed as planned. You will just have to come back for more.

I hope Robert, who has been chief tinkerer with the front end, will document and post the extent of his work so far and continue to fill us in on what he does, why he decided to do it, and how it is working out.

Dave

These are constructive educational comments.

The petticoat here is not a mixing chamber as it dose not have the impingement angle of the exhaust steam on the wall.

Cross bars do nothing except increase back pressure. It's true that will increase slightly the velocity up to the terminal back pressure at around 12 psi. Once you active that you will not get more velocity till the nozzles are designed to go super sonic. Even though the back pressure will increase the vacuum will not. You will reduce the cylinder cycle efficiently and then increase the boiler steaming demand for a given rate. If you get to test this loco you will find it will produce about half an inch of vacuum for every pound of back pressure.

Turbulence even caused by round bars is massive and effects the co-efficient of discharge in a negative way. A square bar is even worse. Above describes the results of such bars with respect to back pressure. You want the minimum of turbulence there.

If you want a quick fix do the following. Make your selfs a splitter petticoat like in Chapelon's front ends. It will produce the boundary layer effects you need. You can also make simple delarvel single nozzles to find the optimum size.

I keep hoping Robert will pitch in with the details of his work.......

but in the meantime, the stack and petticoat sketches are already posted here - is there a more optimal petticoat we could stick on to the nozzle (with a second bridge or without) that might help us bring more of the fire away from the door?

The burner is below the throat - perhaps a central burner would be less likely to blow out the door vent.

Dave