Author Topic: Talking Thermodynamics  (Read 121143 times)

Offline paul gough

  • Full Member
  • ****
  • Posts: 404
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #675 on: January 23, 2018, 06:10:58 AM »
Hi MJM, Had a little bit of time to myself and took the opportunity to quickly look at a couple of your cooling graphs. I for one am interested and pleased you are diligently delving into cooling rates, these may well prove to be valuable in working out a standard method to assess heat losses and insulation performance. This is an area that is often given scant attention as many model operators only care about the engine running to their satisfaction and often having the view that a boiler is a somewhat inconvenient accessory with the only requirement that it makes enough steam. It has been said to me that insulation is irrelevant in gauge 1, this statement by a very experienced builder/operator. I don't hold to this as an absolute, it depends on what you regard as important.

Now a question, looking at run 2 and 3 cool down graphs, reply 671, I note a change in slope of the line from about 60 degrees on both graphs despite one being much more of a straight line than the other. Do you think this is showing something or just an artefact from your method?

When ascertaining fuel or water consumption, are you taking into account the initial or residual wetting of the tank, piping and with fuel the wicks. What is used to wet things or what remains behind inside may be significant with such tiny volumes? To ascertain wetting volume just fill the equipment when absolutely dry with a known volume and then accurately measure the remaining fluid when emptied and subtract. I look forward to getting back in a couple of weeks to start catching up on all I have missed. Regards, Paul.
« Last Edit: January 23, 2018, 06:18:59 AM by paul gough »

Offline MJM460

  • Full Member
  • ****
  • Posts: 897
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #676 on: January 23, 2018, 10:56:25 AM »
Hi Paul, great to have you back, even briefly.  I think the cool down test is most interesting for boilers which have the ability to be insulated for at least part of the shell.  Willy's electric boiler is an obvious example, but any boiler mainly fired within a centre flue fire tube, can have insulation on the outer shell.  The usefulness of the test is more obscure for pot boilers with all or most of the shell exposed to the combustion gases. 

As I understand your gauge 1 locomotives, the boiler has an external firebox at the back end, but the water space is heated by the combustion gases passing through a centre flue.  In this case, the outside shell of the boiler is exposed to ambient air, which will obviously involve a heat loss, and could potentially benefit from insulation.  In addition to the shell, the firebox enclosure is another area for heat losses that could be reduced by insulation.  The cool down test might not help much with this, but it might be possible to see some effect in the heat up test, before steaming starts.

The heat loss to the atmosphere, is heat not available to raise steam.  The cooling test gives the opportunity to quantify the heat loss, so it can be compared with the heat release from the burner.  In fact, it should be compared with the heat absorbed in the steam, as it is not lost directly from the flue gas, it is lost after passing through the copper and the water.  If it is a very small proportion, it might be easier to turn up the burner than find space for insulation.  Obviously the limitations of scale appearance are also relevant.  And it is possible that having more water space might be considered more important.  But if you can fit a thin layer of timber or cork, is it worth while?  Cooling tests with and without insulation can quantify the difference.  Unfortunately there is diminishing returns in adding additional insulation.  For example, 2 mm of cork will not save twice as much heat as 1mm.  Cooling test data is probably the easiest way to answer this question also.

On the other hand, if you are trying to get efficiency, for an efficiency competition perhaps, then insulation is a good way to reduce losses, and hence increase efficiency.

I had given some thought to that initial wetting of the burner parts, I weigh the burner when it is clearly empty, and not been used for sometime, so any remaining fuel would have evaporated.  The quantity I use to fill the burner is selected to burn out before the boiler runs dry.  And I usually weigh again after the run to check that the fuel has all gone.  Possibly some of the fuel either evaporates from the hot burner, and certainly some is used in that final burn down between the steam temperature fall off and when the flame extinguishes.

For the water, unfortunately the whole steam plant including that MDF base is above the capacity of my scale.  So for the water, I weigh out the 130 g in a plastic jug and carefully pour it in with a funnel.  When the run is complete and the boiler has cooled down, I extract the remaining water with a tube on a syringe, and weigh the water extracted.  Obviously there may be a bit left that the syringe does not extract.  I think I should possibly pour in a measured amount to the really empty boiler, and immediately empty it with the syringe.  This would give an indication of how much might be left behind.  But in any case it would give me a more repeatable start point.

I have continued to think about those cooling tests on my little fired boiler, and doing some calculations.  I think I am getting something interesting if not entirely sure that it will be useful.  I will be travelling tomorrow, so no calculations, though I will almost certainly check in.  I also had a preliminary look at my second boiler.  So that is still on my list.

Thanks for following along,

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

Offline MJM460

  • Full Member
  • ****
  • Posts: 897
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #677 on: January 23, 2018, 11:05:05 AM »
Hi Paul, sorry, I missed answering your question.  Certainly a great observation.  I had noticed that the initial part of the cool down curve seemed to be showing something else happening.  Certainly in those tests that started at 100.  That slight inflection at about 60 degrees does seem to indicate a transition between different processes.  I am not really sure what it is about.  I am wondering if there is some water still evaporating, providing some heat to counterbalance the losses.  Because there is no regulator, perhaps something to do with air entry.  I need to pay more attention to the engine position during cool down to use the engine ports to better close off the boiler.

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

Offline MJM460

  • Full Member
  • ****
  • Posts: 897
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #678 on: January 25, 2018, 11:04:46 AM »
Analysis of the cooling curve-

The cooling curve has the obvious use in exploring the effectiveness of insulation on those boilers which include insulation on a significant area of the shell.  However the question remains, is there anything else it can tell us, particularly on a fired boiler where there is minimum shell area available for insulation?

If you remember back to when we started looking at heat transfer and heat transfer coefficients, calculations are based on the basic equation, Q = U x A x delta T.  You will also remember that the equation is not as easy to use as it might look, as the coefficient U in particular, is not a constant, and it's value cannot easily be determined from theory except in a few specific cases.

The interesting thing about the cooling experiment is that we know three out of the four terms, Q, A, and the temperature difference.  Thus we can rearrange the equation to read
U = Q/(A x delta T)
The cooling experiment covers a range of known temperature differences.  Using the equation in this form, we can can calculate a value of U over that range.  The delight of using a spreadsheet is that the facility to copy and paste formulae makes easy work of these repetitive calculations.

There are two different approaches we can take to the calculations.  First we can look at the heat loss over each small temperature interval, and calculate the heat transfer coefficient at each temperature level.  Strictly, the temperature difference is changing throughout the interval, so I should really use the log mean temperature difference.  But I took a short cut, and took the average temperature difference.  It's a minor simplification that introduces a very small error when the temperature change is small.  Obviously the relevant temperature difference is the difference between the inside of the boiler and the ambient temperature.

The first cooling test started at 82 degrees, and the ambient temperature was 22.  The next reading was at 81 degrees, and my spreadsheet already had the heat loss between the 82 and 81.  I calculated the temperature difference as (82+81)/2 - 22 or 59.5.  Using the equation above, the heat transfer coefficient was 77.2

It is always a good principal to make sure that the units are consistent.   I converted the heat loss figure to kJ/hr.  Then using delta T in degrees C and heat transfer area in square meters, you can see the units are all consistent.  So the answer was 77.2 kJ/hr.m^2.C for this first time interval.

Then the formulae were copied, and pasted down the column to calculate the result for each interval down to a boiler temperature of 30 where my recording of the temperature and time ceased.

The spreadsheet is then used to very easily plot a graph of the results.  This is in the first attachment.  The x axis is the temperature difference between the boiler and ambient, while the y axis is the heat transfer coefficient.

You can see the graph reflects, or even amplifies the little errors in the experimental results.  It is also possible that the coefficient flows a relationship to a temperature difference of 40 degrees, or a boiler temperature of 62 degrees.  Then it possibly follows a slightly different relationship, a change that Paul noticed in the heat loss curves.  I am not sure of the meaning of this, whether it is significant, or just coincidence within the accuracy of the experimental method.

The equations were just as easily copied into the relevant rows for the second and third cooling tests.  These are shown in the second attachment.  Again, you can see that quote distinct change in behaviour above about 60 degree temperature difference, or boiler temperature of 80 degrees.

It is clear in all three graphs, that the heat transfer coefficient is not constant, but generally increases a small amount with increasing temperature.  It is tempting to extend the trend to higher temperatures to see what else this could tell us.  It is always a danger to extend an interpretation beyond the limits of the experimental data, called extrapolation.  Of we were to extend this to a temperature difference of say 200 or 300 degrees.  It would probably still be below 100.  However, there is always the possibility that those changes that have already been noticed, might influence the actual direction as the temperature difference increases, or something else might come into the picture.  However I will come back to explore that a bit further next time.

You will have noticed the reference lines, labelled "overall".  I mentioned earlier that there were two approaches to analysing the data.  I have already looked at analysing the data incrementally, that is, looking at each increment.  We have seen earlier that his does tend to amplify those little experimental errors.  The other approach is to look at the overall picture presented by the data.

What if we calculate a overall coefficient based on the cumulative heat loss from 82 degrees down to 30 degrees.  The temperature difference is sufficient that we had better calculate a log mean temperature difference, or LMTD.  Remember the method, in words,
LMTD = (Starting temperature difference - end temperature difference)/ln(Starting temperature difference / end temperature difference)
The function "ln" is the natural logarithm, or log to base e, rather than the log to base 10 that you might be more familiar with.

The LMTD for the first test was 25.5.  This compares with the average of 42.9, or the median value of 45.5, reflecting the fact that the time at low temperature intervals heavily influences the total heat transfer.  Then using that LMTD, I calculated the overall heat transfer coefficient of 67.0 degrees.  This is the value for the overall reference line and you can see how this lies a little below the average.

The LMTD is the relevant temperature difference to use when the temperatures are not uniform, whether due to changes with time as in this cooling experiment, or changing with position, for example the cooling of flue gas along a centre flue in a marine boiler.

I thought it worth checking those overall heat transfer coefficients with my heat transfer text book.  I rushed in a bit, choosing kJ/hr as the heat transfer rate units, I should have used Watts, as strictly, the unit of time in the SI system is the second.  Then I would have had units of Watts/m^2.K for the heat transfer coefficient, the units in the brief table in the text book.  Not to worry, to convert we multiply kJ by 1000 to get Joules, and divide by 3600, the number of seconds in an hour.  (Watt is the name given to J/sec in SI.). So if we divide numbers around 70 by 3.6, we get 19, right in the range of 5 to 25 the text book suggests for natural convection in air.  The upper limit of that range, 25 W/m^2.K is equivalent to 90 kJ/hr.m^2.K on my graphs.

Ok, so that was pretty heavy going.  But I am getting an idea of where it might take us.  Please don't hesitate to ask a question if any of this is not very clear, you will almost certainly be speaking for many.  I will explore the idea a bit further tomorrow.

Thanks for looking in,

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

Offline kvom

  • Full Member
  • ****
  • Posts: 2017
Re: Talking Thermodynamics
« Reply #679 on: January 25, 2018, 05:03:21 PM »
I took thermodynamics in college 49 years ago.  The textbook was shorter than this thread.   :Lol: :mischief:

Offline steam guy willy

  • Full Member
  • ****
  • Posts: 2206
Re: Talking Thermodynamics
« Reply #680 on: January 26, 2018, 01:59:30 AM »
Hi MJM,  still following along but the recent posts are getting a bit difficult to comprehend with all the equations but i am getting the gist of things.....so keep the info coming ....thanks ....

Offline MJM460

  • Full Member
  • ****
  • Posts: 897
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #681 on: January 26, 2018, 10:29:51 AM »
Hi Kvom, great to have you checking in.  I am still a long way short of either of the two books I generally refer to, and the heat transfer book is much bigger again.  Any sort of conversation is bound to be less concise than a typical text book.  However, it is not the length that is important, but whether it is interesting, logical and understandable.  So I hope you are finding something of interest in amongst all the words.  I am always willing to have another go if I have not been sufficiently clear, and questions or comments are always welcome.

Hi Willy, we have all missed you the last few days.  Sorry to hear about your shed.  I hope you have it all back together without too much loss.  Good to see your freelance engine moving again.  I am enjoying following your progress.

I like to include the key formulae so that if people want to do their own calculations, they can see what I have done.  But it is also important that I sufficiently well explain the results and my interpretation.  But of I leave too much out, please prompt me to fill in the gaps.

Last time, I did some analysis of the cooling curves for my little Meths fired pot boiler to see what might turn up.

I found that the heat transfer coefficient is roughly constant as the temperature difference reduced from 60 deg to 10 deg.  With constant heat transfer coefficient, the heat transferred is proportional to the temperature difference, and in the range 20 Watts down to 3 Watts for those temperature differences.  When I compare this with a heat transfer rate of around 400 Watts during steam raising, this clearly indicates that those low temperature differences are not very useful for our steam raising.  I remember some time back, Paul asking about the minimum useful temperature difference.  I think this test gives a more definitive answer to that question. 

In a refrigeration system, temperature differences are quite limited, so a large area is required for sufficient heat transfer, and systems do operate with these very low temperature differences. In our small boilers, where the practical area is limited, much higher temperature differences are required.  Burning more fuel to increase the temperature difference is a good way to increase the steam production rather than a small increase in area.

Are there any other ideas of what else can be learned from the cooling tests on a fired boiler? Or have we exhausted that topic?

Thanks for following along,

MJM460

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

Offline steam guy willy

  • Full Member
  • ****
  • Posts: 2206
Re: Talking Thermodynamics
« Reply #682 on: January 26, 2018, 11:01:11 PM »
Hi MJM , The shed is back together again.!! Here is a practical thermodynamic event you might like to talk about....not mine but something i saw at an exhibition!!! 
    so ..Heat input /loss / horsepower ,,etc etc etc...

Offline MJM460

  • Full Member
  • ****
  • Posts: 897
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #683 on: January 27, 2018, 10:03:52 AM »
Hi Willy, glad the shed as together again.  That video clip is a very interesting pendulum demonstration.  It is hard to tell from a 15 second clip, but I assume the pendulum bob is a magnet, and the bolt in the block of steel allows a fine position adjustment.  I must admit that I had to look up the Curie effect, it describes the way magnetism changes with temperature.  As temperature increases, the magnetisation decreases, or inversely proportional.

I assume it is all carefully set up so that as the magnetic bob approaches the steel bolt, the flame increases the temperature enough to reduce the magnetisation, so the bob is allowed to swing back.  On the back swing it cools and the magnetisation increases.  I imagine it would take quite a bit of experiment to get it all just right. 

It is not an area that I am familiar with, however my thermodynamics text book says work is done when the magnetisation changes, so it is that work that is just enough to overcome the friction and keep the pendulum swinging.

The text book also contains a little paragraph about systems involving work other than pressure, enthalpy, velocity and elevation.  It includes electrical energy, surface tension, strain in a wire, and also magnetic systems, etc.  Each time we analyse a system using the first law of thermodynamics, we should in principle include all these forms of work in order to make a complete statement of the first law.   However, in most practical problems, many of the forms of energy do not change, so they can be ignored.  Changes in velocity and elevation are often negligible, for example.  Similarly surface tension and magnetisation.

In the Curie pendulum, we have a case where velocity and elevation are both changing, there is friction in the pivot and due to air resistance, and the little flame changes the magnetisation to put enough energy into the system to overcome the friction and air resistance.  A delightful reminder that we must include all the relevant forms of energy in our statement of the first law, or conservation of energy.

As regarding efficiency, I imagine the efficiency is incredibly low.  Most of the energy released by the burning fuel will just heat air, and only a very small portion will be turned to work, but please don't ask me to put numbers on it, but the second law says there will be losses.

But it is interesting that magnetic work is used in very low temperature refrigeration systems, systems attempting to reach as near as possible to absolute zero.  But I don't have any experience with those systems.  Nor do I really understand how they work.  But magnetic forces are able to do much more than just keep a pendulum swinging.

A great diversion from some repetitive calculations that will give me a chance to continue the calculations to the point where I can compare my two boilers, and let you know the results.  Might be a couple of days, had a few interruptions today.  It is important to have a life apart from thermodynamics.

Thanks for looking in,

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

Offline steam guy willy

  • Full Member
  • ****
  • Posts: 2206
Re: Talking Thermodynamics
« Reply #684 on: January 28, 2018, 03:37:36 AM »
Hi MJM, I did look it up on the web but as usual you have added more info to make it more interesting and relevant ...

Offline MJM460

  • Full Member
  • ****
  • Posts: 897
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #685 on: January 28, 2018, 11:21:28 AM »
Hi Willy,  keeping it interesting and relevant is the whole idea.  Glad I was able to add something, and that you enjoyed it.

Not much progress here today, another scorcher at 41 deg C.  I took my thermocouple outside at about 8:30 pm this evening, it was still reading 30.  Not expected to go down a lot more overnight.  Just as well we have an air conditioner in the bedroom.  Another hot one tomorrow with a cool change in the afternoon and guests for dinner, so may not have much to report tomorrow either.  However then we have a few cooler days coming.  Thank goodness.  I check in each day so I can start thinking on any questions, so any ideas on what next are welcome.

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

Offline steam guy willy

  • Full Member
  • ****
  • Posts: 2206
Re: Talking Thermodynamics
« Reply #686 on: January 29, 2018, 06:42:09 PM »
Hi MJM I saw this in the "Engineer" magazine of 1887 and it shows a refrigeration boat using charcoal to insulate it ....So Charcoal is a fuel like wood and it reminded me of one of my brick workshops that had a woodburning stove, anyway after a cold winter i realised that i could insulate it with the wooden pallets rather than burning them ,there by using them to keep the place warm for ever rather than for a few hours, so my wood burning would be more economical !! So is there a figure for the insulation coefficient of crushed charcoal ?   ...here is the accompany  article...
« Last Edit: January 29, 2018, 06:54:35 PM by steam guy willy »

Offline MJM460

  • Full Member
  • ****
  • Posts: 897
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #687 on: January 30, 2018, 12:04:44 PM »
Hi Willy, that sure is an interesting concept.  My text books don't include a figure for the thermal conductivity of charcoal, but has figures around 0.15 to 0.19 for various timbers varieties.  Cork is better at about 0.04, but it is in the range of other good insulation materials.  I had a bit of a search, but most of the information was about methods to increase the conductivity for use in adsorption refrigeration systems, not quite what we want in this case.

There was one reference from Finland, where it was mentioned that charcoal has a lower conductivity than timber, which they were using in a similar way to what you did.  But if there was a figure for charcoal, I missed it.  As I read of your experiments I was reminded of that saying about give a man a fish, similar to give him a piece of wood to burn and he is soon cold again, but show him....

I know that you asked about the insulating properties of charcoal, but the really interesting thing about that article was the refrigeration system, which used air as the refrigerant.  I have a special interest in the topic, as when I was a boy, I spent my holidays in my uncles orchard, picking apples for export to UK.  Lack of refrigerated ships was the limit to suitable varieties that would stand up to the trip.  I don't think apples would tolerate such low temperatures however, so that had to wait for more conventional refrigeration.

In that article, the air was pressurised to about 45 psig, and of course the hold and return pressure must be very close to atmospheric, so an absolute pressure ratio of about 4:1.  Did you notice that  after being cooled in the after cooler to something above sea temperature, the compressed air was cooled by expanding and doing work in the expansion cylinders mounted in tandem with the compression cylinders? 

A simple orifice, as used in a conventional refrigerator which condenses the refrigerant will not work, as expansion by throttling of air only gives a very small temperature drop, determined by the Joule Thompson coefficient.  But with an expander producing external work, the story is quite different.

With a pressure ratio of about 4:1 with air (the ratio of specific heats, Cp/C = 1.4) doing work by expansion gives a temperature ratio of 0.673, which has to be applied to the absolute temperature, and in principle, an ideal adiabatic expansion would give a temperature around -97 F, assuming the air could be cooled to about 27 degrees, but depending on the sea water temperature and how close the air was cooled.  So the claimed temperature, while probably a bit optimistic, was not totally out of range.  Obviously if the air is to be cooled to that temperature, the dew point must be a bit lower again, so the system does not block up with ice.  So the air system would have to be totally closed, and the makeup air (to compensate for any leakage) would have to go through a very effective dryer.

The expansion work would help with the work of compression, but as we now understand the second law of thermodynamics, we know it will never be enough to do all the compression. However it would reduce the power required to drive the compressor.  A bit of complexity, the expansion cylinder(s) and the necessary drier, and you have a very reliable safe system with free refrigerant, which does not require a separate evaporator with its attendant temperature loss.

The article seems to finish a sentence or two early, do you have the rest of the article?

Still doing the maths, with many interruptions, to get the results from the tests on my second boiler.  Should be able to compare results later in the week.

Thanks for dropping in.

MJM460

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

Offline steam guy willy

  • Full Member
  • ****
  • Posts: 2206
Re: Talking Thermodynamics
« Reply #688 on: January 30, 2018, 03:42:51 PM »
Hi, MJM,  here is the rest of the article  and i thought it might interest you !! It was in the "Engineer" 1887 .October 14th.... that you can look up in Graces Guide.  I am slowly reading through them all !!! quite slowly.............there may be some follow up articles later ....
« Last Edit: January 30, 2018, 04:01:11 PM by steam guy willy »

Offline MJM460

  • Full Member
  • ****
  • Posts: 897
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #689 on: January 31, 2018, 10:42:30 AM »
Hi Willy,  thank you for posting the rest of that article.  It is quite an interesting machine that would make a great model with its compound expansion engine, air compressor and air expander all on the one base.  An engine with a real load built in.  Would be a great compliment to someone's ice cream machine.  Unfortunately, quite a bit above my current skill level, though I would like to think thatI will eventually get there. 

The valve gear looks interesting.  There would be no need for reversing, but it seems to have two eccentrics to the engine, one of those two part valves perhaps.  Then rotary valves for the air expander, and obviously the after cooler in there somewhere.  The compressor also seems to have driven valves, of some kind.  I don't know why it has that arrangement, but it seems to have a further four eccentrics!

It also gives me new admiration for the research that you and others put into figuring out the design of your historical models.  Unless there are more detailed sketches or drawings available, there would be a huge amount of work to just figure out how it all works. 

The calculations are going slowly.  Not that they are so hard, but the spreadsheet is getting a bit big, and there were a couple of hiccups on the iPad.  I wondered if it is too big, and decided to move it over to the computer.  Unfortunately I have a Windows machine, and don't have Numbers on it.  The iPad numbers spreadsheet does not move so easily.  Probably should have started there in the first place, but the iPad is so conveniently portable.  I am having to recheck all the formula, and redraw all the graphs.  That, and the inevitable interruptions for the medical  appointments that seem to come with advancing wisdom, this week anyway.  Time consuming and exhausting but at least all good results so far.

MJM460

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