Author Topic: Talking Thermodynamics  (Read 123550 times)

Offline steam guy willy

  • Full Member
  • ****
  • Posts: 2245
Re: Talking Thermodynamics
« Reply #660 on: January 14, 2018, 05:02:47 PM »
Hi MJM, thanks for more useful info and answering those questions, and  what happens to ice at absolute zero ? will it contract and split into small pieces? And why does your hand 'stick' to frozen steel pipes ??

Offline MJM460

  • Full Member
  • ****
  • Posts: 923
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #661 on: January 15, 2018, 12:12:45 PM »
Hi Willy, I don't know how much you and other forum members use the formulas available in the typical spreadsheet, so it is hard to know if it is helpful to describe the process, or if everyone knows all that stuff.  I guess there will be some at every stage along the way, so I hope that I can help some progress a little further.  But I am happy to continue answering any questions that are posted, with the proviso that I will say so if I don't know the answer.

It is interesting that with all the space in my text book on the behaviour of water, it does not clearly answer your question about cooling the ice, but I will explain the situation as I understand it.  While water expands when it changes from liquid to solid, I expect that the solid ice contracts like everything else if you further cool it.  I base this on the fact that molecules are always in motion, even in solids, and vibrate about a central point.  As the substance looses heat, the vibration slows, and with it the amplitude of that vibration.  The overall effect is that the molecules effectively take a bit less space and the solid contracts.  There is always plenty of space between the molecules, (relative to the size of the molecules) so they can always get a bit closer together.  I have read that even at absolute zero, this vibration never really becomes zero, though I am not sure how that works.  However the nearer you get to zero, the harder it is to get any temperature gradient, so there are minimal temperature differences to cause thermal stresses that might split the ice into pieces, so I suggest that it just gets cold with nothing exciting to look at.

When your finger touches very cold steel, it rapidly looses heat to the steel until the temperature at the surface of your skin is 0 C or below.  Any moisture on your skin would freeze, again my assumption, I assume there is normally enough moisture to freeze and stick the skin to the steel.  Ice on warm steel is slippery due to the contact area being lubricated by a thin film of liquid or even vapour, but I don't know much about the mechanism for sticking, but my observation is that ice sticks to the metal or plastic tray in the freezer in the kitchen, so I assume that your finger sticks in a similar manner.

A bit more analysis of my test run.  Yesterday I showed how I use a VLOOKUP function to find steam properties in a table.  Those figures are generally on a 'per kg' basis, so that has to be multiplied by the mass of water in the boiler to get the amount of heat absorbed.  The rest of the calculations are mainly simple multiplication or addition, so I will not go into detail unless anyone needs it.  I have recorded the measured water mass in a row above the main calculations.  As long as I use an absolute reference, i.e.use those $ signs, then any formula can pick up that value from anywhere in the spreadsheet.  I also include data for the mass and specific heat of copper in the boiler, and an assumed temperature difference between the copper and the water to enable heat transfer.  This is a significant difference from an electric boiler, where there is minimum heat transfer through the insulated shell so it is reasonable to assume no temperature difference.

I generally find that many of the values in the first row are initial conditions, like ambient temperature, and it is the second row where the calculations begin.  So on the spreadsheet I attached yesterday, the second row has the calculations for changes in the various properties through to the heat stored in each temperature interval.  Once this row is completed, the whole row can be copied, and pasted down the rest of the table and the whole calculation is complete.  Finally, the spreadsheet is able to plot a graph of any selected properties.  The chart type should normally be a scatter graph in spreadsheet terminology, plotting a different variable on each axis seems a foreign concept to spreadsheet developers, who seem to favour pie charts or columns for comparison.

I have attached a picture of the graphical output today.  Ignore the cooling curve for the moment.  The heat up curve has a few interesting features, and it is worth considering what these features might mean.  While the temperature-time graph appears to make a very smooth line, the energy calculations seem to amplify those little time variations between the temperature steps.  On the temperature -time curve, the points seem to vary very slightly each side of the smooth line much like they do on a hand plot.  But the energy curve seems to amplify these variations.  I have carefully checked the calculations, and there may yet be an error I have not found, but it does seem to agree with the recorded time steps and if anything, it is the smooth line in the temperature-time graph that is the anomaly.

Looking at the stored energy graph for the heat up period, my current interpretation is that the first time interval might be in error if I did not start the timer quick enough after I lit the burner, less than ten seconds longer, hardy visible on the temperature-time curve, would make the first step fall into the same line as the next few.  I will get a better idea when I repeat the test a few times.

Then at the 4 minute mark, the rate of heating seems to increase.  This is about the point where those blue tongues of flame from the burner become apparent.  I am thinking that might explain a higher heating rate.  Again, a few repeat runs will tell if this continues to happen.

Finally, the heat up rate falls right off.  This seems to start at about 100 C, when steaming would start in an open boiler, and sure enough this is also where those steam leaks I mentioned appeared.  Obviously heat is lost with that steam that I have not accounted for.  I fixed the steam leaks, fingers crossed, so that should line out on future tests.

One more very interesting point that turned up.  I calculated the heat rate from the burner assuming that the fuel consumption was uniform with time.  I don't think this is totally true, but I don't yet have the means to measure the rate at intermediate times.  So I calculated  the average heat release per second for the whole run based on Meths LCV 26,000 KJ/kg, and the measured 25 g of fuel poured into the burner, and 17 minutes 41 seconds from light up to flame out, as 615 watts.  I used this to calculate the heat input during each time interval compared with the heat stored in the copper and water.  Again the marvels of using a spreadsheet.  I would never have spent the time to do that by hand.  It looks like about 40% of the available heat is taken up by the copper and the water.  The rest is lost in the hot combustion gases leaving the boiler.

That is enough to think about for one day, I will talk about the steaming period next time.

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: 2245
Re: Talking Thermodynamics
« Reply #662 on: January 15, 2018, 02:44:57 PM »
Hi MJM , Interesting stuff going on with the ice... I am not trying to test your knowledge but asking for a fuller practical knowledge of what is actually happening !!... To get an idea of the meths consumption you could have the meths burner sitting on some sort of high resolution weight scales !! (just to add to the confusion)  On my graph where the heat rate falls off ,The red curve, this was also due to the steam escaping from leaking glands, and after tightening them up the line then returned to being linear again (The green line).   good to see where these experiments are heading.....Also I don't quite understand the   Heat up -energy stored graph  ??
Willy.
« Last Edit: January 15, 2018, 02:53:31 PM by steam guy willy »

Offline MJM460

  • Full Member
  • ****
  • Posts: 923
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #663 on: January 16, 2018, 09:57:16 AM »
Hi Willy, so long as my explanations make sense.  The important thing is to understand the physics behind the explanation, not just take a grab bag of miscellaneous facts to memorise.  Please let me know any time I need to have another try.  That atomic theory of matter explains a multitude of otherwise puzzling observations.

With the Meths burner, I have been thinking of boring holes through the burner channel of the base so the burner could be supported on columns off a digital scale underneath.  First I need a scale with 0.1 g resolution.  I don't think the current 1 g resolution is good enough.

Your graph with the effect of steam leakage so clear, is what prompted me to record multiple readings through the heat up.  There is so much more information that can be teased out with more data.

The heat up energy stored graph is a bit tricky, and as always it is difficult to know just what any graph of this sort is showing.  Is it highlighting experimental errors or detail of events?  The x axis is time and the y axis, stored energy, is the sum of the energy taken up by the water, and the energy taken up by the copper as it heats up.  That sum is divided by the time interval to give the heat rate in Joules/s or Watts.  That division seems to amplify the effect of variations in the time taken for each 10 degree interval.

I have carefully checked the calculations, but an error is always possible.  However, if I assume the data and calculations are correct, I then tend to assume a single reading out of line with the others could be the effect of a timing error or similar, while three in a row are probably showing something. 

On the graph I posted yesterday, I suspect the first point is a timing error, or some other startup effect, then comes two groups of three in a row that seem to be regular, and the change in gradient at about 4 minutes might be significant.  The burner seems to work like a conventional wick until about this time, then the tongues of blue flame appear at the holes on the two side rows.  This may reflect a change in the rate of heat release.  Finally the tail off seems consistent with the observed steam leakage, so is most likely due to heat not accounted for in the water or copper.

This is the basis of my explanations yesterday.  It remains to be seen whether these effects are seen in further runs.  I believe I fixed the steam leaks, so that much at least will be interesting to see.

Based on the total fuel burned, and the LCV, the average heat release is 612.1 Watts.  Quite a bit lower than your electric element, but probably appropriate for such a small boiler.  It certainly runs my little single and double acting oscillating engines (wobblers) quite well.

I did two extra runs in the period before my holiday, but on one, I got carried away with the technology, and cleared the iPad stop watch data in preparation for the cool down.  Unfortunately, I had not written down the times.  Whoops!  It was not even necessary, the iPad had plenty of capacity to record all the additional times I wanted without clearing the data between heat up and cool down.  But I got a track of the cool down anyway.  When I get through the analysis of this one, I will get out my notes and see how they compare.

When the boiler got to 110 deg C, I drained all the condensate, and the engine started as soon as I moved it off top dead centre.  It ran for 10 min 21 secs at a speed which varied a bit around 980 rpm.  I extracted 30 g of water from the boiler after it cooled down and as I started with 139 g of water in the boiler I assume that 109 g were evaporated as steam.  Averaged over 621 seconds, that is 0.176 g/s.  The boiler temperature stayed pretty constant at 110, while the superheater outlet/engine inlet varied a few degrees each side of 135, so I took that as the superheater outlet temperature.

Now the saturation pressure for 110 deg is 143 kPa ( so 43 kPa gauge or 6 psig).  The steam tables tell us that hfg, the heat necessary to evaporate water from 110 C to dry steam at 110, is 2230.2 KJ/kg.  It takes a little interpolation of the superheat tables to get the enthalpy change in the superheater, but I calculated it to be 51 KJ/kg.  Add those together and multiply by the steam flow rate and we get 400.4 J/s absorbed making steam. 

Now, remember the burner average heat release was 612 J/s, so 65% of the heat from the fuel was absorbed by the steam. 

Again, remember the heat up period, the heat absorbed was only 40% of the heat released.  Again it is interesting to contemplate the reasons for this difference.  Part of it will almost certainly be explained by the slow burner startup.  If less than average fuel was burned during those first 7 minutes, then more than average fuel was burned during the ten minutes steaming time.  Those differences would bring the percentage absorbed in each phase closer together.  I don't know if it is enough to explain the difference.

As a first step, if we calculate the total heat release from the fuel burned, we get 0.025 x 26,000 = 650 kj.  The total heat absorbed by the steam is 248.7 kJ, and the total by water and copper during heat up is 76.6 kJ.  248.7 + 76.6 = 325.3 kJ which is right on 50 % of the heat released in the burner.

Without a specific measurement to tell the proportion of the heat release during heat up compared with during steaming, it is perhaps best to take this overall figure  of 50% as the best estimate of the efficiency of this little boiler.

That seems like a good place to stop for the day.  I hope it has been of interest.  Tomorrow I will look at the cooling test.

Thanks for looking in,

MJM460
« Last Edit: January 16, 2018, 10:06:12 AM by MJM460 »
The more I learn, the more I find that I still have to learn!

Offline steam guy willy

  • Full Member
  • ****
  • Posts: 2245
Re: Talking Thermodynamics
« Reply #664 on: January 17, 2018, 01:28:50 AM »
Hi MJM, more good info here and this is me with the thermometer reading my alleged temp ....also lots of info on the net about efficiency of the Lancashire boilers,  however some say 65 plus %  and others up  to 85%.. To test the efficiency do you have to let it actually power an engine under load for a period of time?

Offline MJM460

  • Full Member
  • ****
  • Posts: 923
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #665 on: January 17, 2018, 11:25:37 AM »
Hi Willy, that thermometer certainly looks suspect, as you obviously look healthy enough.  I assume you had it under your tongue for the usual time to let it reach equilibrium.  How did it go for zero in a well mixed ice-water mixture?  And again for 100 in boiling water?  The best way to try for boiling point might be to use your nicely insulated electric boiler without the plug, so boiling is determined by atmospheric pressure (don't forget to read the barometer).  You could try the thermocouple in the filler opening for a few minutes, and also compare it with your test position under the insulation.

It seems a bit anomalous that to test a boiler efficiency, it is helpful to run an engine.  The engine tends to meter out the steam production and maintain a steady pressure.  If you do another run with a load on the engine, you will get a higher pressure. 

Obviously, the steam production is the period we are most interested in.  In addition, during steaming, the temperatures are all constant, compared with heat up when the metal and water temperatures are continually changing.  More subtly, the changes to water temperature changes the heat transfer coefficient a little, but there is a major change of heat transfer coefficient on the water side when boiling starts.  Remember the earlier discussion on heat transfer, there is a factor of about 10:1 in heat transfer coefficient when boiling begins.  This means the copper temperature is much closer to the water temperature, which in principal increases the temperature difference to flue gas.  I am not sure how important this is, particularly at the burner end of flue tubes, where the temperature difference is quite high.  However at the stack end, where the temperature difference is expected to be much lower, it may significant, though not so important in this little horizontal pot boiler where the flue gas flows across the boiler.  In addition, the burner output would be expected to be more uniform at a long term rate than during the initial heat up.  Clearly you would not expect the efficiency of this boiler to match those Lancaster boilers with their multiple flue tubes in addition to the outside of the shell.

I know that I said I would look at the cooling curve today, but when I thought a bit more I decided that it would make more sense to put the heat up results from my third run into the spreadsheet, so they can be compared with the first.  Please don't mention the second!

The spreadsheet really comes to the fore now.  No need to look up any more steam table values, the table already there is still easily accessed by exactly the same formula as for the first run.  A small adjustment to the formula to access the location of the basic data for the fuel and water fill and water removed at the end, and the main job is to simply insert the time and temperature data in the same columns of the rows below the previous data.  Then just copy the first row of formulae and paste them down for enough rows.  Then you can go direct to the graph function.

I have moved round the graphs so the ones for the first and third runs are one above the other with the same scales for easy comparison. 

Once again the temperature rise with time seems quite as expected.  Still a little bit of a slow down after 100 degrees, but no where as much as in the first run.  And interestingly, the graph of heat stored per second has much the same form as the first run.  The high rate in the first minute shows clearly again.  By chance, I had two quick readings in that first minute, and they show a level of consistency that indicates that feature means something.  I took all the readings about two weeks ago, and am only now getting to analyse them, so the only difference between the runs is practice, as I did not do any analysis to identify anything I should look for between the runs.

I am starting to feel that initial high reading, this time two readings, are telling us something.  I wonder if the burner initially flares a bit, then settles to its warm up.  I will have to conduct more runs and watch that burner more closely.  Perhaps I can set up the camera to take a photo every 10 seconds or so, to better understand its behaviour.

Then again we have two sections with a slightly different gradient, the gradient change at about 4 1/2 minutes compared with 4 minutes the first time.  Again the appearance of a similar behaviour tends to confirm that this is a significant feature.

Finally, a bit of a tail off after the temperature reaches 100.  Not as much as the first time, but still there.  My observation was that the steam leaks at the pipe fittings had been fixed, however the only shut off valve I am using is the port face of the engine, which I set at top dead centre.  It would not be surprising if that leaked slightly.  Also, at some stage around that point, I blew a bit of steam through that drain bypass on the engine, to clear the initial condensate.  Pity I did not note just when I did it.  But I think you can see, with the small boiler, heat up was pretty quick, and with all those readings it all seemed to be happening at once.  I tend to be a bit slow in getting it all recorded in those circumstances.  But clearly less tail off than the first run, and the temperature of 110 deg C was reached a little sooner than the first run.

Then the engine started.  It ran for 11 min 29 seconds before the flame visibly died, and was extinguished about 10 seconds later.  I will talk about that part tomorrow.

I hope the pictures are clear enough, a hand held iPhone does not give the same sharpness as a scanner, but all I have at the moment.  But the form of the lines is easy to see and compare

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: 2245
Re: Talking Thermodynamics
« Reply #666 on: January 18, 2018, 01:31:09 AM »
Hi MJM, i suppose with a naked flame  lots of other things can affect it ,like fido waging his tail....mother in law leaving the front door open....the cat finding a warmer place to  snuggle up to ...!! lots of things that don't ever appear in the formulas. I suppose laboratory situations can nullify extraneous effects like these, costly and quite time consuming. I don't have a fridge at home so may have to do the zero centigrade in the local cafe.. but could do the boiling water at home !! Good progress with the graphs btw....


Offline MJM460

  • Full Member
  • ****
  • Posts: 923
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #667 on: January 18, 2018, 11:25:31 AM »
Hi Willy, I think you are on the right track, thinking about wafting air currents.  Not the mother in law, dog or cat, but the front and back door were open, allowing beautifully gentle breezes to waft through.  I had not thought of it as affecting the burner and heat up, but wait until you see the cooling curves.  I suspect the effect is small compared with the heat release of the burner, but it played havoc with the cooling, where the heat rates are much smaller, so a huge percentage change.

Sorry you don't have a fridge, but if the cafe will provide a glass of icy water and ice blocks with your coffee, that will do.  But every engineer needs a fridge for calibrating thermometers.  It even comes in handy for keeping milk, butter and meat if you can't think of anything else.

Short post tonight.  My calculations for the steam production and overall efficiency did not stand some basic checking, I am skipping a line somewhere on this tiny screen.  The heat up curves are fine, it is the overall efficiency and the steam production energy balance that has the problem.  Oh for a real computer with a keyboard instead of losing half the screen.  Time ran out on me to find the problem and correct it, so tomorrow is checking, and I hope back on track. 

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: 2245
Re: Talking Thermodynamics
« Reply #668 on: January 18, 2018, 11:49:16 PM »
hi MJM ,  ok cool.....here is melting ice and boiling water on my temp gauge , the ambient being 17 degrees .... Also ,about  Meths....if you leave it to evaporate what is left in the container ??....

Offline MJM460

  • Full Member
  • ****
  • Posts: 923
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #669 on: January 19, 2018, 11:55:36 AM »
Hi Willy, there definitely seems something strange about that thermometer.  It looks like about 3 deg low at 100, but 2 deg low at 37.  Then 3 deg high at zero.  You could draw a calibration curve through those three, but I am not sure that it would be right. 

I have searched through some old notes and found my calibration notes from two separate occasions.  I had two different thermocouples and meters, and compared the two both times.  I now have three more thermocouples, and two meters, one of which has inputs for two thermocouples, so it's time I tried again.  However, from the notes I have, the first time I read 99 degrees at boiling, and 1 degree at the ice point.  I should have quit at that point, but I kept wondering if I could eliminate those 1 degree differences.  The next time, I managed to get readings of 1 deg C at the ice point, but unfortunately I could not get so close to boiling the second time and have recordings of 95 and 97 the second time.  I have also noted that I tried switching the thermocouples but got even worse results.  I noted that it took 3 to 5 minutes of careful maintaining the temperature and stirring to get stable readings from an instrument initially at room temperature.  I seem to remember giving up in disgust, with the intention of trying again another day.  I did use an insulated cup for the ice point and improvised a lid with a folded tea towel, in the attempt to minimise any temperature gradients in the ice-water mixture.  I suspect I mainly proved that the experimental technique is important, and I seem to have done better the first time.  Did not think of a photo, as that all happened before I found this forum. 

However, all the attempts help you understand a little better just what is required to get a good reading, and the time it takes to come to a steady reading.  It also gives some credence to those specifications that often come with the meter.  I expect, (or is it hope?) that differences between readings on a specific meter might be more accurate than the absolute values displayed.  And it is also a warning not to place too much faith in the accuracy.

One other point, did you try comparing the thermocouple with that good glass thermometer you have?

Meths is a mixture of ethanol with 5% water.  When it evaporates, the vapour is mostly ethanol but with a small amount of water, and in principal there is nothing left in the container at the end.  However I think you mentioned that your Meths is dyed blue, so there may be traces of the dye staining the container at the end.  I am not sure what happens to the flavouring they put in to discourage drinking, but the quantity may be too small to easily see the residue if it remains.  Our Meths is clear, and I don't notice anything remaining when it evaporates.

It was 42 here today, the second 40+ in a row, and yes we are a metric country, so not great progress today.  Must have been diabolical at the tennis.

However I did manage to check the calculations for heat up and steaming phases of my boiler tests, and corrected the problem that was worrying me yesterday, so I am happy that they are now correct within the limitations of the data accuracy, an important point considering our discussion on calibration. 

The third run, I found slightly higher efficiency for heat transfer during heat up and lower efficiency during steaming than the first run.  As I have mentioned before, I don't have data on how the burner heat release varies throughout the run.  So a figure for the overall test, heat absorbed during heat up and steaming is probably the most useful one.  The first run gave an efficiency of 49%, while the third gave only 45%.  Interesting that the third run for which I had fixed the steam leaks, gave lower efficiency, even though I had the impression that I had much less loss in terms of steam leaks.  An awful shame that I mucked up the second run, a third would be quite useful at this stage.  But it gives some idea of the repeatability of these tests.

I will try and produce the graphs for the cooling tests tomorrow.  The weather forecast is for below 30 tomorrow, so should be more comfortable.

Thanks for looking in,

MJM460

« Last Edit: January 19, 2018, 12:01:22 PM by MJM460 »
The more I learn, the more I find that I still have to learn!

Offline MJM460

  • Full Member
  • ****
  • Posts: 923
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #670 on: January 20, 2018, 10:39:33 AM »
Quite a big day for us here.  Fortunately not as hot as yesterday, but a big day for us, our fiftieth anniversary.  Celebrated with a lovely meal at a local restaurant.  Not easy for us with dietary requirements.  With a low fat for one, and Fodmap friendly for the other, most places do an abysmal job.  But tonight's meal was just perfect for us both, and really excellent.  Just shows what a real chef can do.

I know there are others who have reached this milestone recently or are fast approaching it, so it's not a unique event in this group, more a matter of joining the club.  Another great thing about being a member of this forum.

In a quiet break, I even managed to get those cooling curves sorted.  You might remember that I did three test firings in my small Meths fired pot boiler, recording the temperatures during heating and cooling in the excellent manner Willy demonstrated earlier. 

With a fired boiler, we don't have that highly predictable heat input of the electric element in an electrically heated boiler.  Instead the boiler heat up is highly influenced by the rate of fuel burning in the burner.  So the heat up and steaming measurements are as much a test of the burner as a test of the boiler.

The cooling test is a little different.  On Willy's electric boiler, the cooling test was a very good demonstration of the effectiveness of the boiler insulation.  This was done by comparing the cooling curves with and without insulation.  However on a fired boiler, we can't insulate the outer shell.  Well, no more than the boiler ends, which are outside the firebox anyway.    And the difference in cooling times between this little uninsulated boiler and Willy's nicely insulated electric boiler shows how much difference the insulation makes. 

Really, I am not yet sure what the cooling tests on a fired boiler actually show, I am still thinking about that.  However, the times and temperatures were recorded, and the calculations done, and I have attached the graphs below.

There are four graphs.  The time temperature graph is exactly as expected and follows the form predicted by Newton's law of cooling.  A very similar form is seen if I construct a graph of the remaining stored heat as cooling proceeds, as you would expect.  Note that between heat up and cooling, the boiler steamed for a period so the stored heat at the start of the cool down is only due to the copper and the remaining water.  I weighed the water I extracted with a syringe after the cooling was complete to determine the remaining water after steam production.  Subtracting this from the initial fill amount also tells the quantity of steam produced.

However, when I calculated that heat loss as Joule per second, and plotted that value, the variations are amplified and the graph is quite erratic.  The fourth graph is a simplified cooling in Joules per second, plotting only one point every five degrees instead of every degree.  This certainly smoothed the curve by averaging things out. But based on every degree, when I put in the connecting lines they look like the stitches from a very poorly adjusted zig-zag sewing machine.

I leave you with those tonight and will look at them more closely tomorrow.

Thanks for looking in,

MJM460

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

Offline MJM460

  • Full Member
  • ****
  • Posts: 923
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #671 on: January 21, 2018, 12:03:46 PM »
While I was at it, I put in the time and temperature data for the second and third run that I had recorded when I did those heat up runs.  Copied the formulae from the completed ongoing curve for run 1, and pasted them in to the rows for the additional data.

Then all that remains is to construct the additional graphs, and the three cooling runs can be compared side by side.  So today's picture has runs 2 and 3 to sit beside the ones for run 1 in yesterday's post.  At least, I hope it is attached.  Every time I send one of these pictures to myself, the iPhone offers me a different three file sizes.  The smallest ones appear a bit fuzzy so I have been selecting the largest one that fits in the forum limit.  Unfortunately, having offered me a specific size, when I open the received mail, it tells me a different size.  Very frustrating.

The runs are not quite the same.  It turns out that I only recorded the times at 5 degree intervals.  I think I found recording every one degree a bit too intense.  At that stage, I had not done any of the analysis, so did not know about the variation in the heat rate on a per second basis.

The other difference, for runs 2 and 3, is that I was organised enough to start recording at 100 deg.  Seemed like a good idea at the time, however, as you can see, it seems to have introduced more irregularities, ones that were missed in run 1 when I started at 80 degrees.  More to try and understand.

I have had some time to look at the calculations and the actual numbers.  The burner heat release rate for the two heat up tests were 612 Watts and 590 Watts.  Interesting variation in successive runs for a single burner, but probably within tolerance for 25 g of fuel weighed on a scale with a resolution of whole grams.  Like other tests, it is obvious that repeated runs are desirable to get an idea of the variability of the results.  The curves which rely on these figures seem quite uniform and predictable.

For the cooling tests, the cooling rate starts off at about 20 Watts, and reduce with reducing temperature to about 3 Watts at 30 degrees.  The ambient temperature is 22 deg.  This difference in the absolute size of the heat transfer rate is probably the main reason the cooling curves seem so erratic.

I mentioned earlier, that I had the boiler set up in the doorway to the outdoors, where I was getting very welcome light breezes making the day very pleasant.  I did not think of these same breezes affecting my test results.  But it makes sense that with the boiler at 70-80 degrees, the breezes might affect that cooling rate by a few watts, a big percentage of the nominal 15-20 Watts.  While when the boiler has cooled down, the cooling effect of the breezes might be less significant.  Clearly the next set of tests need to be conducted is still air.  However recording 5 degree intervals does average things out a bit.

In the later two cooling tests, starting at 100 clearly introduced something new.  The expected result is a more rapid cooling between 100 and 95 that we expect between 80 and 75. Yet this was not the result.  Clearly some more thinking is required.  And almost certainly more testing required.

I am still thinking about the practical implications of these cooling curves.  There is not much point in thinking too much about them if there is no practical outcome.  The heat up and steaming tests however, potentially give us a method to predict the steam output of a larger or smaller similar boiler, and also a method to compare the efficiency of different boiler designs.  Clearly a useful outcome.  So tomorrow, I will have a go at calculating a performance parameter.

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: 405
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #672 on: January 22, 2018, 08:47:06 AM »
HI MJM, Back early, sent you a P.M. Regards, Paul Gough.

Offline MJM460

  • Full Member
  • ****
  • Posts: 923
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #673 on: January 22, 2018, 11:15:37 AM »
Hi Paul, good to have you back.  I have sent you an email. 

Some time back, when thinking about comparative boiler ratings, I had a look at the parameter K. N. Harris used, cu. in./min/100 sq.in.  I suggested that a logical metric equivalent might be kg/hr.m^2.  So what would be the value of that parameter for the little pot boiler I have been testing? 

The boiler shell is 38 mm o.d. and 150 mm long.  However, only 130 mm is in the firebox exposed to the combustion gases.  The remaining 10 mm of the shell each end and the boiler ends are not exposed to the flames.  In addition, it is usual to assume only half of the shell area, recognising that that is about the average portion of the shell with boiling water on the inside, where the film coefficient is very high compared with the steam space, where the inside coefficient is very small.  On this basis the boiler heating surface area is 0.00776 m^2.

The steam raised was 0.153 g/s.  If we divide by 1000 to get kg, and multiply by 3600 we get 0.55 kg/hr.  Now divide by the heating area in m^2, and our parameter is 71 kg/hr.m^2.

This is quite a convenient figure, neither too large or too small, for comparison with other boilers. 

Of course, you can immediately see the deficiency of this parameter, the steam rate was determined as much by the heat release of the burner as by the boiler design.  Now I call that burner my 50 mm size, as the burner is 50 mm long.  It has a capacity of 25 g of Meths with the square tank, and the heat test runs so far indicate a heat rate of approximately 600 watts.  The test result calculations I have been describing the last few posts indicate that about 50% of this heat can be accounted for in the heat required to heat the boiler and its contents and the heat absorbed by the steam production.   Now I do have a second burner, a 90 mm size, with obviously higher heat output.  At the first opportunity, I will do some runs with that one. 

You can see that this simple performance parameter probably mainly applies to coal fired boilers, where the grate size is reasonably defined in relation to the boiler type and size, and it is probably reasonable to assume a predictable relationship between grate area and heat release from the coal burned.

The next step is to do similar tests on a different boiler.  I have a second boiler, fitted up with a slide valve engine.  It was a busy few days between Christmas and New Year, and I have my readings from that one as well, ready for analysis.  I am getting lots of practice at copying and pasting those formulae in the spreadsheet.  Saves an awful lot of calculations.

The further I go with looking at this little boiler, the more I find myself in Paul's position with those little locos, such a simple plant, yet so much scope for experiment.  But a second boiler will be interesting to compare.

Still thinking about what can be learned from those cooling curves.  Stay tuned.

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: 2245
Re: Talking Thermodynamics
« Reply #674 on: January 23, 2018, 12:44:51 AM »
Hi MJM, I don't quite understand the graphs on the right .? could you explain them as they don't seem to follow the cool down at all...Thanks