Author Topic: Talking Thermodynamics  (Read 194421 times)

Offline paul gough

  • Full Member
  • ****
  • Posts: 457
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #720 on: February 17, 2018, 10:16:43 AM »
Hi MJM, Interested to see the heat loss diagram from the firebox wall. The ambients are more or less the probable operating range for most G1 locos in OZ so would likely be indicative of conditions for a stationary loco shielded from the wind, thus a baseline or starting point for an investigation. Can you quantify, (even approximately), 1) the distance at which temp. line goes horizontal, 2) the heat losses for the two ambient temps. even as a proportion or % increase in loss.

I think your incremental increase in fuel supply test should provide a reasonable guide, I can't think of anything better. If I ever get my engine back together I will probably pursue some adaptation of your experimental method, and so am pondering the degree of accuracy achievable and necessary to get repeatable and consistent results. I don't want to fork out about $500 for a very accurate scale if I can utilise a syringe to measure fuel and water volumetrically and then convert to grams. I recognise that we are dealing with minute quantities and volumes, thus accuracy is important, but would like your opinion on what % or range of inaccuracy would be acceptable at this scale for all the parameters. Regards, Paul. 

Offline steam guy willy

  • Full Member
  • *****
  • Posts: 3238
Re: Talking Thermodynamics
« Reply #721 on: February 17, 2018, 01:26:57 PM »
Hi MJM , just put the photo up again !!! Glad to see a slide rule ...with its curser still attached !! Also saw this in a Model Engineer  15th Jan 1901 edition of a chap fromNew Zealand doing practical experiments with boilers !!!  He says raising steam took 3 1/2 mins !!
« Last Edit: February 17, 2018, 02:55:31 PM by steam guy willy »

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #722 on: February 18, 2018, 11:50:31 AM »
Hi Paul, not so easy to answer that lot, but I will see if I can convey the general idea.

I really should have shown that the profile on the inside of the firebox which has a similar curve near the wall, in that the hot gas is cooled in the vicinity of the wall, so the profile curves down as it gets close to a minimum temperature at the wall.  Heat transfer to the wall on the inside is also by convection.  It is not so easy to give a figure for the thickness of the thermal or velocity boundary layers, the velocity starts at zero at the bottom of the wall, while the temperature starts at the wall temperature, and the profile develops as the air rises up the wall.  My sketch represents the profile at some height above the bottom.

Rather than try and define the profile totally, which involves some pretty heavy differential equations, it is usual to talk about an overall coefficient, a number that gives the same total heat transfer at the sum of each little bit of the wall.

I looked at some of the worked examples in my textbook to find some figures that might be typical of our case.  It looks like the boundary layer thickness essentially becomes constant some where in the range of 2 to 12 mm from the wall.  I hope that is close enough for your purpose.

The steel internal box makes very little difference to the profile as it is a good conductor compared with the insulation or the air.  The insulation and the air boundary layers resist heat flow like electrical resistors in series, but we are talking about conductivity, the reciprocal of resistance, so the conductivities involve the reciprocal of the sum of reciprocals, similar to the formula for resistors in parallel.

I assumed the conductivity of the insulation is about 0.05, and it was 3 mm thick.  Typical convection coefficients for similar cases look like something in the range of 2 to 5 J/m^2.K.  I assumed a typical flue gas part way through the boiler as 225, giving a temperature difference of 200 at 25 degrees ambient.  At 5 degrees ambient, the difference is 220 degrees.  On this basis, there is about 12% more heat loss at the lower temperature.

I think when I get through some more analysis of the cooling runs, I might be able to calculate an external heat transfer coefficient for my boiler, we will see later in the week.

Of course all this has been about natural convection.  When your locomotive starts moving, the velocity of the locomotive becomes an air velocity imposed on the system.  This is called forced convection.  The boundary layer is thinner, and the heat transfer higher due to a higher effective temperature difference across the air film.

Thanks for thinking about the incremental fuel experiment, I will give it a try.

My small boiler uses about 25 g of Meths, so 1 g would make an error of about 4%.  However, it uses 130 g of water, so the same 1 g is only about 0.7%.  Generally these percentage errors all simply add together through the process.  However, square or square root functions don't carry through so easily.   When the calculations are done in a spreadsheet, I would just change the fuel quantity or whatever, by the one g, and check the change in the answer, just to be sure.  It only takes a few seconds.  I would also balk at $500 for a scale.  My daughter in law has one from Jaycar that she uses for her hobbies as well as cooking, 2 kg, with a resolution of 0.1 g, that was around $150, but still sizeable cost.  To give you an idea, the scale included a wind shield to assist in getting steady readings.  A syringe might be a suitable alternative, that is much more accurate than the marks on a jug, or even a medicine glass.

Hi Willy, I thought might like the slide rule.  No use at all without the cursor.  I bought that little one for my pocket for when the more usual 10" size was inconvenient to carry.  When I told the younger engineers at work that I did not even have a calculator for the first eight years of my career, the look on their faces said they thought I must have grown up in the Stone Age.

I don't suppose many of the original subscribers to that 1901 magazine are reading these posts.  They are quite interesting.  He was quite the master of understatement!  If I have understood correctly, in one sentence he said he built a locomotive style boiler with about 6" barrel and five 1" tubes.  That must be almost the shortest build log in history.  You can imagine the reaction if that was all someone wrote here.  But then he said he tried larger (than the 1") and smaller, but found the 1" best!  So he built three locomotive style boilers in two sentences?

They are no small boilers either, barely classed as miniature.  As for getting those boats to the pond, you have to admire him.  However he makes a lot of sense in his opinions.  Lots of surface area per unit volume will give quick steaming if you have enough fire.  But the designs are complex, and certainly not shown in all their detail.  He did mention stays being required on one of them, another understatement.  And even admitted that one was not easy to build.  Only one?  But his figure 5 is an interesting concept, giving plenty of area and also enough volume.  But despite having designed a few full size vessels, I would have to give that one a lot of thought.  You have to admire those early guys, no codes to guide them, did not know what would not hold pressure, so they just built them, and presumably built a stronger one if the first effort failed.  And fail they must have done.  That is why we have the codes today.  Interesting articles, thanks for posting them.

Thanks everyone for looking in,

MJM460

The more I learn, the more I find that I still have to learn!

Offline paul gough

  • Full Member
  • ****
  • Posts: 457
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #723 on: February 18, 2018, 10:36:52 PM »
HI MJM, Thanks for the quantities, it helps me considerably when constructing a mental picture  during my day dreaming sessions. Following this investigation is teasing out some clearer thinking from the neuronal fog and eventually I'll be more certain about a line of tests suitable for a loco.

Regarding heat loss due to air movement, A reasonable guiding figure for velocity of G1 locos would be to use 1m/sec, (e.g.  Indoor or dead still outside conditions and assumes no contribution from wind). Am wondering if you could have a stab at predicting what further % heat loss there might be using this figure over that already mentioned.

I am inclined to think I should pay more attention to insulating the firebox on my loco!

A shame you have to sweat it out with Melbourne's elevated ambients, do you think it of any consequence to obtain or convert test results to standard temperature i.e. 20C, so making comparisons easy for others or future experiments.

Thanks for the tip, I will have to check 'Jaycar' more thoroughly, a satisfactory scale at $120 is attractive, also useful for biological specimens. I'm on the lookout for a measuring cylinder and a sizeable pipette to check my syringes accuracy, should be able to get some used scientific glassware somewhere at a reasonable price. Regards, Paul Gough.

Offline paul gough

  • Full Member
  • ****
  • Posts: 457
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #724 on: February 19, 2018, 08:32:48 AM »
Hi again MJM, I was pondering thermocouple placement for exhaust gas temperature and then realised this could be problematic. Most stack temperature measurements that I have had experience with on boilers were really to provide a guide to combustion conditions and an indicator that things were going well, or not, particularly when operators are somewhat remote from the boiler. It is the variance from the observed normal for any given load that alerts you to problems. As the actual temperature could be widely variable depending on placement of the probe in the stack  the obvious question arises, where is the best place for our purposes to place a probe/thermocouple? Is it necessary to compute the gas flow and heat content to determine the average gas temperature one should be getting, then go chase a spot where to drill the hole in the stack to insert the probe? Hope I'm not confounding everyone with unnecessary complications, but I really would like to know how to best place a probe and why. Regards, Paul.

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #725 on: February 19, 2018, 12:42:57 PM »
Hi Paul, I will have a look at the convection equations in my text book and see if I can get an idea of the difference with the moving loco.  However that will have to wait until tomorrow.  Drove 300 km today, slow with roadworks and other stops we had to make, and then a meeting this evening.  I don't think the equations will make any sense tonight.

Insulation of the firebox would help reduce the heat losses, if you can fit it within the locomotive outline.  In that size, it is not worth making it ugly to save a tiny amount of fuel.

Melbourne temperatures seem a bit more tiring than in my younger days, though we have a couple of cool spots in the house these days.  At least we do not get the humidity you have to put up with, though some tolerate it better than others.  Do you have the buildup before the wet like Darwin?

I suggest a good way to standardise the results is to calculate the overall heat transfer coefficient, as I suspect it is much less sensitive to the actual ambient temperature.  Then you can use that coefficient with the temperature difference for the conditions you want to compare.  I have not done enough tests over a wide range of conditions to see how important ambient is to a boiler.  Ambient temperature is much more significant to us, as we operate at about 37 so the difference between 35 and 25 is huge, but for a boiler operating at 150 - 300 deg?  However there are a few inevitable errors in the work we can realistically do, so in general, I think comparative results are more important than actual figures.  Melbourne is a bit problematic in that regard though, as we can have 35 one day, and 25 the next, or sometimes later the same day.  But most places probably don't change quite so rapidly, so tests one day can probably be reasonably compared with those done on another, at least in the same season.

I used to get laboratory glassware from Selbys.  They were still around to sell me a chemistry set for one of my kids. It was the real thing, not just salt and sugar like most you can buy these days.  Not sure if they are still around, but your local chemist can probably get some in for you, they should know, or have access to the current suppliers.

The measured stack temperature is probably quite sensitive to thermocouple position.  Partly due to heat loss from the stack, and partly due to the temperature profile across the stack.  The heat loss can be addressed by insulating the stack, so eliminating height as a big factor, but the profile across the stack is a flat topped parabola, due to convection loss from the walls, and due to development of the profile with height.  A very tall stack would have a full parabolic profile, but most model stacks are not tall enough for the full profile to develop.  The velocity is highest in the centre, and the flow is slowed down by the shear forces where it is nearer to the walls to zero at the wall.  That slower layer then slows the next layer in and so on.

My approach is to make a simple clip to hold the thermocouple in a fixed position, so readings can be compared, and to place the sheath (or thermowell if I was using a bare thermocouple) so it extends about 2/3 to 3/4 of the diameter of the stack.  This way, the thermocouple passes through, and hence is influenced by most of the profile.  Then it is reasonable to assume the reading is an average across the profile at that location.  I am assuming the error between the reading made this way and the true average bulk temperature is small enough compared with all the other errors in these simple experiments.   I have not yet drilled a hole for a more permanent mounting, though that will be the best in the long run.  In the mean time, a simple bent wire clip is holding the thermocouple in a fixed position.  I try my best to eliminate obvious errors where practical, then rely on comparative results to be close enough to not be misleading.  I know I am exploring the boiler and engine performance with some pretty deep theory, but I also have to keep in mind the practicalities.  Things have to be kept somewhere near balance.  I am trying to understand the principles so I can improve the steam plants, but have to remember that I am not writing a paper for the national physics laboratory.  I hope that helps with understanding how far to pursue absolute accuracy.

Definitely time for bed, it's a new day tomorrow.

Thanks for looking in,

MJM460
The more I learn, the more I find that I still have to learn!

Offline paul gough

  • Full Member
  • ****
  • Posts: 457
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #726 on: February 20, 2018, 04:58:21 AM »
Hi MJM, Thankyou for the reply to my enquiries, you must have a very resilient brain. If I drove 300 klm's I don't think the contents of my cranium would be capable of serious discourse or considered written replies at the end of the day.

The high humidity can be a little tiresome for some people when temps. are over 30, but it is when the herps., (reptiles), are most likely to be active, so I am quite happy to live with it. There is a pre wet 'build up' in Nov./Dec. but the wets are now paltry compared to what was. We have not had a real wet for years, so bad are wet season failures that central western Queensland has had six consecutive failures of wet season rains and are now in dire circumstances, hardly any cattle left in those areas. Regards, Paul Gough.

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #727 on: February 20, 2018, 10:44:06 AM »
Hi Paul, so long as my post made sense.  I actually bought a calibrated eye dropper from our local chemist last week.  Only 1 ml, so a bit small for fuel, but great for filling the displacement lubricator which tends to trap air bubbles. 

I have had a bit more of a look at the chapters on convection today.  The maths is quite intimidating, but when I skim the working out parts (basically mathematical gymnastics that are easier to follow than develop), and look for the conclusions, I am finding some equations that I can actually use.  I will put them in a spreadsheet to allow exploration of some of the variables, and easier checking.  It looks like it might be possible to estimate boundary layer thickness, heat transfer coefficient and heat loss.  Order of magnitude anyway, and more importantly, give helpful guidance on the difference when single variables such as ambient temperature, or velocity are changed to answer your questions.

You mentioned about 1 m/s.  What about size of that firebox?  I have assumed about 50 mm long, but better to use a more realistic figure.  Also the height.  Just thinking about the size of the firebox, which I assume is shaped like a conventional loco boiler, but in that size, dry sides and top?

I learned a few things.  Probably obvious in hindsight.  The analysis assumes only natural convection, forced convection or radiation, but in reality, usually all three are in operation.  In your loco, natural convection leads to rising air flow, while the motion imposes the forced convection with horizontal flow.  And of course your fingers held nearby tell you there is radiation.  All at the same time.  However there are criteria for the relative importance of natural and forced convection.  It's all tied up with Reynolds number, Prandtl number and Nusselt number.  Sounds complex, but when you put in the actual figures, it is not as complex as it first looks.  It looks like in this case, the thermal boundary layer is a little thicker than the velocity boundary layer, but most of the maths is about comparing three potential velocity profiles with experimental results, that are in reality all close enough for our purposes.   I will see if I can make a bit more progress tomorrow.

Also had a good day in the shop today, well, good for me anyway.  One of those stops along the way yesterday was at the bolt shop.   A long way from home, but right by the highway we were travelling.  Only a matter of stopping, whereas the ones closer to home involve 30 min travel each way.  So I now have enough M3 nuts, and the engine has four studs.  I won't have to hide anything next time I take a photo.  Found a bit of stainless steel shim, so am pondering whether I can use this as a spring to lightly hold the valve against the port face.  I can see big advantages in arrangements which have the port face on top, though they need radial valve gear, or extra linkages.  I think I also need a tiny condensate drain for the steam chest.  Along with that steam shut off valve.  All projects to follow, but I can make progress without them in the mean time

I also made up the missing hand wheel for that displacement lubricator.  Could have designed something more compact, but I used some small off cuts of brass and plastic that I had.  And in the process, I reminded myself how to centre on a shaft, drill and tap an M3 hole for the grub screw.  Won't need the pliers for that lubricator now, much better.

Also one of my thermocouples was playing up.  Looked terrific on the table beside the others with all reading the same for room temperature.  But it did not respond when I put it in the thermowell in the steam line.  Seemed to be something in the plug at the meter end of the wire.  Not the plug that has three screws holding it together, that would have been too easy, but one of the ones heat welded.  Unfortunately two of mine are that type.  Attacked it gently (if that is the appropriate word) with the hobby knife, it opened quite easily, but not much to see.  I tightened up the crimps on the terminals, but no consistent result.  Turned out that the inner insulation had been trimmed back to far.  The bare wires were just twisted together and the cover pulled down to conceal it all.  The point those wires crossed became the measuring junction, and it was just measuring the temperature in the plug.  Grrrrr!  I don't know how it worked well for so long.  Now, when I pinch the welded junction (the intended junction) between my fingers, the meter quickly responds, climbing to about 30 C.  You don't get much more when measuring body temperature that way, but it is an easy quick check that the thermocouple is working properly.  The previous calibration checks are not affected by any of this, so now it properly responds, I can use it with confidence again.  Might actually get a full run tomorrow, the first with the boiler turned around, burner raised stack temperature measurement and insulation on the furnace.  It has been a bit frustrating, and a long process, but progress is made one step at a time and let's keep fingers crossed that it is all right this time.  Tentatively tomorrow, but my 11 year old grandson has an electronics project for school, and discovered he left his tools in Darwin.  It is a complex life they lead.  That and possibly helping a friend who recently had a small stroke get to a medical appointment will take precedence.  And replenishing the fridge.

Stay tuned,

MJM460
The more I learn, the more I find that I still have to learn!

Offline paul gough

  • Full Member
  • ****
  • Posts: 457
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #728 on: February 21, 2018, 06:44:34 AM »
Hi MJM, Your posts are making sense and I am getting an ever clearer picture of what is going on and an appreciation of the subtle and complex goings on with respect to heat and its investigation. I would not be competent to do a lot of the maths nor in fact have the understanding of when and where to apply it appropriately, nevertheless, by apprehending some salient propositions and absorbing the conclusions advanced from your tests, along with your hints and suggestions, I think I can come up with a testing methodology and enhancements to my boiler/loco, some simple, such as modifications, some less so and needing re-manufacture of significant components to gain better performance. Measurement is now so easy with the cheap, (relatively), electronic tools available from the likes of 'Jaycar' it is important to develop skill at using them and interpreting the information. This includes things like siting of probes and understanding just what exactly one is measuring and the implications of it. Stack temperature an example, and for me, the extension of it to the smokebox, a somewhat more complex environment into which to stick a probe looking for something meaningful. Therefore I want to thank you for the considerable time and effort you put in to assisting me with your replies.

My firebox is approximately 35 long x 20 wide mm. and about 25mm high to the bottom of the boiler barrel and has a 12mm overhang past the rear of the barrel where a 20mm. extension upwards delivers gases to the tubes. The sum of all the firebox surfaces comes very close to 35 square centimetres, then there is the 7 sq. cm. of the open base, if that is to be included/added as heat loss area, (??). These surfaces have been insulated on the inside with the thin ceramic fibre sheeting (probably 3mm originally but now about 1 mm.), this internal insulation is very deteriorated and needs replacement. I am of a mind to remove it, run a test without it, replace it and then, lastly, add the insulation to the outside and re-run the test a third time to see if a second layer externally has any value. I'm also of a mind to stick a thermocouple probe into the volume of the firebox behind the rear tube plate as this might be more even and stable than the firebox proper, my thinking is it might help interpret firebox/combustion conditions as well as be indicative of conditions in the tubes. My only concern are does the 'cheap' sheathed K type thermocouples from Jaycar able to take the 1000 C.(??), or more, temperatures.  Also would it be problematic to leave the probe in such temperatures for say, half an hour, or more?

I am happy you state, "I learned a few things." I felt a bit guilty in loading you with hefty enquiries so hope this in some way makes it worthwhile. If you are after some small parts or components for you boiler etc. you might find, <argyleloco.com.au> has something useful. They are the main supplier for all things Gauge 1 in the OZ region, most of the bits are listed under "Accessories" and then choose one of the brandnames, not everything is listed, eg fuel/water tubing, ceramic sheeting, wick material etc., so a phone call might be worthwhile. Michael Ragg is the owner and is located in Olinda up in the Dandenongs. I'll shut up now. Regards Paul.

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #729 on: February 21, 2018, 11:07:11 AM »
Hi Paul, so long as you are gaining understanding, I am quite happy to continue.  The number of reads is still increasing, so I am assuming many others are reading too, but not saying anything.  I hope I am not too intimidating, it would be nice to hear occasionally what they think, and I am sure there are many other questions on people's minds.

As you say good temperature instruments are quite cheap these days, and the function is built in to most digital voltmeters.  The Jaycar catalogue and the online data gives the temperature limits, and many go to 1000 C.  The bare welded junction I showed in some earlier posts is probably not up to being inserted directly into flue gas, but in a suitable thermowell I would expect it to be fine.

I have one of the sheathed type, similar to the one in some of Willy's pictures.  I use it in the stack.  I keep meaning to do the finger test at various points along it to see how it would go in a thermowell as it seems a bit more rugged than the others. 

I will keep those firebox dimensions in mind as the calculations develop.  Not much progress on them today due to diversions I mentioned yesterday.  I like the idea of insulation inside the firebox.  Performance without detracting from appearance.  Most industrial furnaces are insulated on the inside.  Keeps the metal cooler, so stronger when you are trying to support a tall stack with attendant wind and potential earthquake loads.  An extra layer on the outside raises the metal temperature unnecessarily.  Inside the firebox is pretty rough on insulation.  Gas velocities fatigue fibres, and even erode solid reinforced concrete insulation.  I would suggest replacing what you have with the original 3 mm, and if you want to experiment, try two layers, but both on the inside.

A thermocouple (in a thermowell) in that chamber at the front of the boiler would be interesting to try, so long as it can be put so it is in a gas stream, but not in direct view of the flames, where radiant heat would upset the reading.  But more important, one at the entry to the stack so it gets a roughly uniform outlet gas temperature, as the flame temperature cancels out in the air flow calculations, but the second equation allows calculation of the gas temperature, well enough for our purposes.  I will look up argyle, they must be near miniature steam, who were shifting up there last time I visited.

I mentioned the trials of working on a troublesome thermocouple yesterday.  Did a run today.  Apart from being unusually slow to start, it went quite well and behaved normally once it reached about 30, until just before I quit taking the cooling readings.  I wanted just one more at 45, having a good series down to 50 when all of a sudden the boiler temperature jumped to about 70.  I rattled things around, took the thermocouple out and it immediately started falling to ambient.  Put it back in the boiler filler plug thermowell, same thing.  When I pulled it out again it responded well to the finger pinch test, so I swapped it with another that had been reading engine inlet during the run.  It also read the higher temperature.  It seems to be a valid reading, on two different thermocouples, yet not consistent with the stack temperature which was around 45 and infra red readings all around the outside of the casing.  Can't explain that one at all.  Could not see or feel anything untoward.

More analysis in the next few days, but the superheater does not seem to work well with the small burner in the large boiler, whether the burner is raised or not.  I think I need to go back to the bigger burner.  And I will also try running the 50 mm burner in the open to see if I can get an idea of how uniform the flame is over time.  I suspect it is a bit slow to warm up, then cools off a bit towards the end.  If it burns well outside the boiler, I will experiment with enlarging some of the vapour holes.

On the surface, a perfectly normal run, but amazing how much you see when you sook very closely with a few simple instruments.  Perhaps I am looking too closely.  I need to take a step back and assess where I am up to.

Thanks to everyone looking in, don't be shy about comments or questions.

MJM460
The more I learn, the more I find that I still have to learn!

Offline steam guy willy

  • Full Member
  • *****
  • Posts: 3238
Re: Talking Thermodynamics
« Reply #730 on: February 21, 2018, 11:45:00 PM »
HI MJM, still looking in......Does your meths burner give you the most amount of heat possible as i have seen drawings where the meths is heated and evaporated ,then ignited under pressure into the boiler ?? these are ideas from 1901 though !! also is the short letter about aluminium true and if so when you heat the AL and Antimony to the lowish temperature  to melt ,how do the get to the higher temp ?? or do the two metals freeze on contact ?

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #731 on: February 22, 2018, 11:49:09 AM »
Hi Willy, I think the amount of heat per kg of fuel is pretty much fixed by the chemistry.  However different burner designs may burn more fuel per second in the given size of furnace, and hence raise more steam.  More excess air lowers the maximum temperature attained, but more gas at lower temperature is still the same amount of heat.  Pressurised fuel with a vaporiser would force fuel into the burner at a higher rate, so releasing more heat per second, but only the same amount per kg of fuel.  That stove burner I showed some time back vapourises the fuel in the pipe that passes through the burner and boils a kettle quite quickly with no sign of a wick.  The pressure available is only that due to the height of Meths in the fuel tank, perhaps two inches depending on whether it is nearly empty or just filled.  The biggest problem is that if the flame does go out, the fuel continues to flow.  Not good if there are still hot surfaces.  I can increase the quantity of fuel for a longer run (if I can supply enough water) using a chicken feeder arrangement, but I am not sure what arrangement the pressurised ones use, unless it is a variation of the ones that were used in times past for kerosene or petrol.  I keep seeing references to the "silent type", that appear to have some vaporising provision, but I have never seen a design.  Do you know what they look like? 

I still need to experiment more with the size of vapour holes in my burner, and need to try enlargening them further, even to the point of making them too big and having to make a new burner.  But I would like to try one of those silent ones.  They look a bit like those poker burners that Chris mentioned, but with the vaporising tube along the top for liquid fuel.

That is an interesting article about the alloy melting temperature.  Generally in a two phase mixture, even when the phases are liquid and solid, the mixture properties are somewhere between the properties of the components, proportional to the composition.  But some mixtures form a different "compound like" structure, that has properties higher or lower than either of the components.  Such as in this case.  It has implications in some metal refining processes that involve progressively melting a section along a bar so some of the impurities float out to the end.  But that intermediate composition would limit the purity that could be obtained.  It would act a bit like a mixture of that alloy and one of the components, depending on which component was most prevalent in the initial mixture.  I suspect it is similar to distillation of ethanol and water, you can't get past that 95% ethanol mixture, which ever side you start. 

But your article is particularly pertinent for me at the moment.  I recently bought a book of recent science writings, in which one essay was about "Impossible Alloys".  A researcher, thinking about how alloys behave, like the small amounts of carbon in iron to make steel.  Generally the alloys are added in small amounts, and too much makes the material brittle, like a very high carbon steel.  He wondered if you go further and add very high amounts of the alloy material, would you get past this point and further improve the material.  The idea was that more chaotic structure is lower entropy, and hence more stable.  Hence they are called low entropy alloys.  He did some experiments and found alloys that did not follow any of the conventional rules so was stuck with trying random combinations, of which there are too many millions to exhaustively try.  But following the line of research is leading to high melting point stronger alloys for gas turbines and similar exotic materials.  Others have found that the copper in brass and bronze can be replaced by a lower cost mixture of aluminium, zinc, nickel and manganese, to make a stronger more wear resistant alloy which is potentially lower cost.  Unfortunately they did not mention machinability, is it too rude to say writers do not think of such things?  But keep a look out for "Low Entropy Alloys".  Possibly from Ampco Metal in Switzerland.  You might be able to update your engine with low entropy metal for the cross head slides and bearings!

I am continuing the analysis of my recent boiler runs in the background.  Trying to get some interesting results to talk about, without having to post too many equations.  It is going a bit slow, as we have the carpet layers coming Monday.  Carpet through most of the house is as bad as painting, in terms of the preparation and restoring of order after that is required.  Everything off the floor!  In the whole house.

Paul, I had a look at my Jaycar catalogue today, it is time I got a new one.  But it confirmed 1000 - 1200 degrees for the welded junction and its insulation, though I would suggest installing it in a thermowell to isolate it from flame or steam.  However the plastic handle at the end of the steel sheathed model that I used for the stack temperature limits that type to only around 250 deg C.  I was a bit suspicious of that handle, and if you look back to the picture, I fixed it so the plastic was to the side of the stack.  I don't think I exceeded 250 anyway, or at least not by much, and the handle still looks ok.  The industrial ones used in the hydrocarbon and other industries do have a more rugged arrangement with a metal sheath, but then they are used at all temperatures, 24/7, 365 days a year.  And inside they are just similar type k thermocouples.

Thanks for looking in,

MJM460

The more I learn, the more I find that I still have to learn!

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #732 on: February 22, 2018, 11:57:23 AM »
Hi Willy, apologies, I missed your question -  I imagine that if you have the two metals, both molten at a temperature somewhere between the 650 of the components and the 1000+ of the alloy, and mixed them, that you would get a solid phase forming, presumably floating or perhaps finely dispersed, a bit like a hot version of a slushy drink.  Just as if you melted the alloy and then cooled the liquid, you would get the solid phase forming, either floating or finely dispersed.  I am not a metallurgist, so have not seen it done, but I can't imagine how else it would happen, but I don't have any reason to doubt the basic facts of the letter.  Perhaps some of those who do their own casting can comment.

MJM460
The more I learn, the more I find that I still have to learn!

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #733 on: February 23, 2018, 12:09:28 PM »
Still continuing to do the analysis, but slow progress with all the moving books and all other small objects.  My wife said if she ever feels like replacing the carpet again, I should put her to bed with an Asprin until the feeling passes. 

However I did put together the equations for combustion of ethanol in air, and calculated the flue gas composition for exact theoretical air, two times theoretical and ten times theoretical air.  Next step is to estimate the specific heat for these mixtures.  It is interesting to see that while the mass of the fuel plus air equals the mass of flue gas, the volume of flue gas is greater than the volume of fuel (as vapour) plus combustion air if compared at the same temperature.  It is only about 7% for ethanol as fuel! though it is much more for longer chain liquid hydrocarbon fuels.  As the excess air just travels through absorbing heat, so it dilutes the flue gases, so reduces the change in volume.  And of course because it's temperature of the flue gas is higher, the volume of flue gas is even greater.  Worth considering when thinking about the diameter of the stack, but also any other restrictions to the gas flow.

As my first estimate of the air flow from the test run used the specific heat at ambient temperature, I will calculate a better value.  Equations for the variation of specific heat with temperature are available, and I found the table for all the flue gas components is included in an appendix to the thermodynamics text, so this gives the opportunity to calculate a better value, which will improve the accuracy of the calculated temperature and air flow.  Then I can go on to calculate those heat transfer coefficients. 

I also managed a few minutes in the shed, and tried that "pinch test" on the metal sheathed thermocouple.  It responded quite quickly to the last 20 mm being pinched between the fingers.  It also responded when gripped in the next 20 mm, if slightly slower to respond.  When pinched in the third 20 mm, the response was still obvious, but much slower than when the fingers were closer to the end.  This suggests to me that the heat is conducted along the sheath, and so affects the reading.  To minimise this effect, I suggest it will be worth slipping some heat resistant insulation over the portion of the sheath which is outside the temperature you want to measure.  The electronics shop sells a woven glass fibre tube that will work quite well, so I need to pick up a length of that and put it on the part of the sheath which sits outside the stack.  Probably better to use the bare thermocouple inside a thermowell when practical.

Not much more to report.  I hope to make some more progress on the calculations tomorrow.

MJM460
The more I learn, the more I find that I still have to learn!

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #734 on: February 24, 2018, 11:17:29 AM »
There is always a danger of loosing direction when you get too caught up in theoretical calculations.  This is not helped as the spreadsheet gets bigger, but you can only see a small part at a time.  It is worth stepping back for a few minutes and taking stock of progress and checking that the direction is still appropriate.  After all the purpose of all this investigation is to use theory to improve our boiler and engine operation.  Despite appearances, I am not into a purely theoretical exercise, I want practical results.

The initial reason for testing a model boiler is to see how much of the energy released by combustion of the fuel actually ends up in the steam.  In Willy's electric boiler, it is easy, as the electrical energy input either ends up in the steam or it is lost through the outer surface of the boiler.  The cooling tests provide a good method of measuring the heat lost.  A measurement of the water lost from the boiler in steam and the time over which actual steam production occurs tells us the rate at which steam is produced, from which we can make some estimate of the engine that the boiler would drive, and a check estimate of the heat losses.  We can reduce the heat loss by adding insulation.

In a fired boiler, it is a little more complicated.  First, heat is necessarily lost in the flue gas, and in addition there is heat lost through the furnace wall.  In my simple boilers the efficiency for steam production is quite low.  It seems worth trying to increase this efficiency before I try and make a bigger burner.  So it is worth trying to understand where the heat goes.

The flue gas obviously carries heat to the atmosphere.  There are two main contributors to the amount of heat in the flue gas.  First, we can't practically cool the flue gas below the steam temperature, and in fact, the outlet temperature has to be enough above the steam temperature to provide a reasonable temperature difference to transfer heat across the Copper to the water. 

We can measure the stack temperature, but we really need to know the gas flow rate and gas specific heat if we are to determine how much heat is in the gas.  So far I have progressed the calculations to the point where I found that we can estimate the air flow of we measure fuel consumption, ambient temperature and stack temperature.  This was a pleasant surprise, it is not always easy to anticipate what can be calculated until you actually try.

Calculating the air flow then revealed a second factor in how much heat is lost in the flue gas.  Burning fuel requires a certain amount of air to provide the necessary oxygen.  And the air comes with nitrogen which simply absorbs some of the heat.  But in practice, to get complete combustion, we actually need excess air, otherwise the dynamics of the combustion reaction tend to mean some of the carbon ends up as carbon monoxide.  This is undesirable, firstly because carbon monoxide is extremely toxic, even in quite small concentrations.  But secondly, burning the carbon monoxide in air to get carbon dioxide releases more heat.  The heat lost in incomplete combustion is a greater penalty than the loss to more air flow through the boiler.  In industry somewhere in the range of 15 to 20% excess air is found necessary in a well designed burner.  I found that I had more than ten times the theoretical air.  Never mind the accuracy of the measurements, that is potentially quite a penalty.

Why is excess air a problem?  Well, when the fuel is burned the heat released is all absorbed by the gaseous products of combustion, and also by the extra unnecessary air.  The temperature reached depends on the mass of air flow.  More excess air does not reduce the heat released in combustion, but it does reduce the maximum temperature reached.  As the flue gas progresses through the boiler, it is the temperature of the gas stream that determines the rate of heat transfer to the boiler.  Lower temperature means less heat transfer, even though the gas stream contains the same amount of heat.  And more mass going up the stack is also more heat loss.

If I can reduce the excess air, while still leaving enough for complete combustion, I will increase the maximum temperature in the firebox, and get more heat transfer to the boiler, so more steam.  Early quick tests by just closing off some of the air intake were inconclusive in terms of actually seeing a difference, however, with the more detailed test, I am hoping to see an increase in maximum temperature reached.  That may even result in a higher stack temperature, (I am not sure if it will actually be higher or lower) but with a smaller flow it will be less heat lost up the stack. So should also be reflected in more steam production.

So that is the first line of research/experimentation.

 I would also like to get an idea of how much heat is lost from the furnace walls.  I have made a crude attempt to reduce this heat loss by applying a layer of suitable insulation.  This gets the wall temperature to a level where I can use a more effective insulation like cork or fibreglass, even rockwool, which will be suitable but might melt against the hot steel casing if I did not have that first layer.  Insulation suitable for the higher temperature is generally less effective as insulation.

But the issue is whether I can estimate how much heat is lost from the furnace wall?  Is it important compared with the heat into the steam or the flue gas?  It's only worth reducing it so far.  So the convection calculations are aimed at estimating the casing heat loss.

These calculations get pretty heavy.  Not sure how far I will get, made some definite progress but I am wondering how much more effort is worth putting into it.  I am certainly clarifying my understanding of convection and will in due course will describe it in a bit more detail.  Perhaps I can see a way to make a very approximate estimate as a start.  I am also starting to think about whether I can do a simple test to make a measurement. 

The cooling test worked very well on the electric boiler which was completely insulated, and yielded very interesting results.  But the fired boiler has to have flue gas.  So I can't easily separate the flue gas losses from the wall losses while firing, but I am wondering if, when the burner is extinguished, I could block the stack with perhaps a cork in the top of the stack, the cooling rate for the boiler would then be essentially due to the heat loss from the walls.  That cooling test might give an answer, accurate enough for the purpose, and a lot quicker.  Besides its more fun to play with fire and steam than sit in front of even the most fascinating spreadsheet!

Food for thought while I shift more furniture tomorrow, then Monday is the big day.  Though then everything has to be put back.

Thanks for looking in,

MJM460
The more I learn, the more I find that I still have to learn!

 

SimplePortal 2.3.5 © 2008-2012, SimplePortal