Author Topic: Talking Thermodynamics  (Read 122055 times)

Offline MJM460

  • Full Member
  • ****
  • Posts: 902
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #750 on: March 06, 2018, 10:06:44 AM »
Hi Paul, glad to hear that my descriptions are making things clearer for you.  Please let me know if you need any gaps filled in.

The formula for heat transfer q=U x A x (T1 - T2), in which U is the heat transfer coefficient.  I have quoted this formula before, but it is very informative.   For conduction through steel or copper, U = k/L, where k is the thermal conductivity and L is the thickness of the material (in meters).   The thermal conductivity of copper is 399 W/m.K, while for steel it is 36.  Stainless steel is only 14.    For the same resistance to heat transfer, copper can be 11 times thicker than steel and 28 times the thickness of stainless steel.

The other factor of importance to heat transfer is the specific heat of the material.  Copper is 383 J/kg.K while for steel it is 486 and stainless steel 499.  This means more heat is stored in a kg of steel or stainless than steel, but the smaller thickness means there is actually less stored heat in steel, especially as copper has slightly higher density.  For your small locos, this means a slower heat up time, but once you are at steam temperature, it becomes irrelevant.  Essentially it is important during temperature changes, but not during steady steaming conditions.

The thermal conductivity, specific heat and density are all included in the parameter, thermal diffusivity which is equal to k/d.c where d is the density, c the specific heat, and is included in material properties in heat transfer texts.  For copper the value is 116.6 x 10^-6.  For steel, 0.97 x 10^-6, and for SS, 0.387 x 10^-6.  It is used in unsteady heat transfer equations, and a higher value means that temperature "spreads" more quickly if one surface is heated.  But it is not important for steady state problems.

But an even more important consideration is corrosion.  Corrosion of copper is not much of a problem at the conditions in our boilers.  However for steel it is a a significant consideration.  Full size boilers made from steel require careful feed water treatment to minimise corrosion, and operating procedures to minimise oxygen entry.  You might think of stainless steel, however, SS is prone to an insidious form of corrosion called chloride stress corrosion cracking.  Full size pressure vessel failures have been attributed to minute concentrations of chloride in potable water used for hydro-testing.  You need to use a Duplex stainless, but that introduces real fabrication difficulties for even experienced welders.  It's diabolical stuff to weld.

You mentioned the strength of steel, which allows it to achieve the necessary strength without too much thickness.  This is in addition to its lower cost per kg.  Copper would be quite impractical for a high pressure due to both strength and cost.  But all around, I think copper is still by far the best material for our model boilers.  And it is also the easiest to fabricate.

Hi Admiral DK, I was hoping you would come in on that one, thank you.  You deal with the issue much more continuously than I do.  I have worked outdoors in temperatures ranging from -33 to +55, even experienced -45, though only for a walk in the park.  I definitely prefer the lower temperatures.  Never was in Antarctica, though that would be an appealing adventure, with adequate clothing.  I think the T shirt in the freezer indicates your speaker has pretty good circulation, or perhaps is well acclimatised and active in the freezer, but I agree with him on humid air around zero.  In my experience,  the worst temperature range for comfort is that range from -5 to +5.  It is nearly always moist, and often windy.  The coat that kept me cosy in -33 was just plain inadequate at a windy +5.  And yes, I have worn it in both.  Below about -5, the air is usually quite dry, the sky often sunny and blue, so long as there is no wind, and you are wearing good clothing, it can be quite pleasant.  Obviously, blizzards are another matter.

Wind, of course, increases the convection heat loss, so increases the demands on the clothing.  You definitely need a wind proof outer layer.

I will attempt to get back to forced convection tomorrow,

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

Offline MJM460

  • Full Member
  • ****
  • Posts: 902
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #751 on: March 07, 2018, 01:10:25 PM »
Well I indicated that I would come back to that forced convection.  I have gone over the calculations many times and I can't see any sensible, simple way to present them so I will stick with the key points in the process and the end result.

I have already mentioned that the treatment of forced convection starts, as with most energy problems, with writing out the equations for conservation of energy, momentum and mass, then attempts to solve these for some simple cases.   For a flat plate, like the side of Paul's firebox moving through air, the first step is to obtain the equations for the velocity of the air flow close to the plate.  The air flow meets the front of the plate at the full velocity.  The air immediately against the plate is considered stationary with respect to the plate, and as the air moves along the plate a velocity profile develops between zero at the plate and the free stream velocity some distance away.  Remember, it does not matter if the plate is stationary in an air stream, or the plate moves through still air.

Solving the equations with some very clever maths, and a few judicious assumptions, gives a mathematical description of the velocity field.  The distance from the plate at which the velocity is 99% of the free stream velocity is defined as the edge of this boundary layer.

In this case, the plate is hot and is being cooled by the air.  The air also varies in temperature from equal to the plate temperature to equal the free stream air temperature some distance away.  The temperature normally reaches the free stream temperature at a thickness a bit less than the thickness of the velocity boundary layer.

I have attached a copy of the sketch the text book provides to illustrate this velocity boundary layer, and which also shows the thermal boundary.  It might help make sense of the words.  The scale in the y direction is very different from the x direction.  The boundary is relatively thin.

Our old friend, dimensional analysis is used to come up with some non-dimensional numbers that might be relevant.  However, the dimensional analysis cannot predict the precise relationship.  When experimental work is carried out, and the results graphed in terms of those non-dimensional numbers, some clear relationships appear.  In particular, one such number is called the Prandtl number, after a pioneer in this field.  It is a combination of some properties of the fluid that varies only with temperature.  It appears commonly in these and other calculations so is directly listed in tables of properties of gases and even some liquids, particularly when those properties are included in a heat transfer book.  It turns out that there is a relationship between the thermal boundary layer thickness and the velocity boundary layer thickness that depends only on the Prandtl number.

Further dimensional analysis shows another dimensionless number called the Nusselt number, which is a combination of the heat transfer coeficient, the thermal conductivity and the length of the plate.  And a bit more maths shows there could be a relationship between this Nusselt number, the Prandtl number, and the well known Reynolds number.  Experimental work then reveals the relationship between the three.

Nu = 0.664 x Pr^1/3 x Re^1/2

Now that looks very mysterious until you remember that the Nusselt number contains the heat transfer coefficient along with two other known quantities, the length of the plate and the air conductivity, while the Reynolds number contains the air velocity, length of the plate and the air viscosity.  Hence that one equation relates the heat transfer coefficient to the air velocity.

If you remember Paul's locomotive was moving at about 1 m/s, and the side of the firebox is 35 mm long and 25 mm high.  If we assume the firebox outside surface is 127 deg C (an odd looking choice that gives 400 K as the absolute temperature, so avoids interpolation of the gas property tables) we can use that equation above to calculate the heat transfer coefficient.

I put all the numbers into a spreadsheet and found the heat transfer coefficient was 21 W/m^2.K.

To give you an idea of the sensitivity to velocity, I put 2 m/s into the spreadsheet and the heat transfer coefficient became 29 W/m^2.K, higher but not directly proportional.

Now the area of the firebox side and the temperature difference can be inserted with the heat transfer coefficient Into the normal equation for the heat loss.  Because the area of the firebox side  is very small, the heat loss is only about 0.2 milliwatts.  Small enough to be neglegible.  It is likely that radiated heat exceeds this, and is thus the predominant mode of heat transfer.  It looks like I will have to look at radiation heat transfer after all.

I am not sure it is useful to say too much more about it.  I find the maths quite intimidating, so I doubt that many will be interested in more detail.  However one interesting thing that did came out of the calculations.  I found the heat transfer text book branched out into fluid mechanics.  Because that forced convection analysis was based on the basic equations of energy and motion, the maths also yielded the drag force on the plate.  And by solving the equations together, another no dimensional group appeared that related the heat transfer coefficient to the drag force on the plate.  So it was concluded that if you measured the drag force, you could predict the heat transfer coefficient without having to do heat transfer experiments.  Remember, even those results I have presented require a measurement of the wall temperature, for which I made an assumption.  Unfortunately, it is much easier to measure that wall temperature than it is to measure the drag force to a satisfactory accuracy.

I hope that is of interest, please let me know if you want any more information about the procedure.

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: 404
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #752 on: March 07, 2018, 06:53:27 PM »
Thanks for the analysis thus far, and look forward to some guidance with radiated heat losses. I have to say I am surprised at the minuscule result. I had in mind something amounting to a few watts, not fractions of a milliwatt!! Thus several orders of magnitude in error. Just shows what difference there can be between ones ideas and the real situation. My only comment is that the surface temp. is probably higher than the one used as the metal has discoloured, but assume this would not make any significant difference to the result. Again, thanks, I find these investigations intriguing. It's 5 am. so off to a swamp to see whats happening, heaps of rain. Regards, Paul Gough.

Offline MJM460

  • Full Member
  • ****
  • Posts: 902
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #753 on: March 08, 2018, 09:56:56 AM »
Whoops!  Paul, you are right in feeling those heat losses were a bit low, your intuition was not too bad after all.  I must admit I also thought they looked a bit low, but the heat transfer coefficient looked about right, and the heat loss formula is so simple once you have the coefficient.  I checked that the formula referred to all the right cells, and that I hadn't dropped a line, even did a quick order of magnitude estimate, a la slide rule, and got the right figures, but of course, not the power of 10.  I assumed that it was probably due to the tiny area.  However it was after midnight, fortunately not rising at 5 am, so I posted it and made a mental note to do a bit more checking this morning.  I finally found it, a divide sign instead of multiply.  That will do it every time, especially when it applies to the temperature difference, a figure very close to 100.  Eyesight must be dimming more than I thought.

Thanks for a better estimate of the wall temperature.  It should obviously be higher than the value for my insulated firebox, and that colouring will be a good clue.  I did the calculations again for 200 deg C wall temperature.

The corrected value for the heat loss for 127 degrees wall temperature, 1 m/s is 1.9 watts, I am sure you will agree a more reasonable answer.  When I increased the velocity to 2 m/s, it was 2.7 watts.

When I increased the wall temperature to 200 degrees C, the heat loss increased to 3.3 watts for 1 m/s, and to 4.6 watts for 2 m/s.

The heat transfer coefficient calculation was not affected by this error, and did not change significantly for the higher temperature.  All the change in heat transfer coefficient was when the velocity was changed.

Remember that even the text books only claim this analysis to predict heat loss within 25%.  Also, it is carried out for one surface in the air stream much like the side of a tank loco.  Apart from the other side, top and back surfaces, I suspect the firebox sides are partly behind the boiler, and may even be inside frames.  Some cross sections and a plan view, even just simple sketches, would help understanding the actual arrangement. 

If you look closely at that velocity profile and temperature profile in the boundary layer illustration that I posted yesterday, you will notice that right at the front of the plate, the boundary layer is very thin, so the temperature gradient between the wall and free stream temperature is very high.  A high temperature gradient means a very high heat transfer rate.  Further along the plate, the boundary layers for velocity and temperature are thicker, so the temperature gradient is less steep.  This is reflected in the calculations which, when looked at in detail, show that the average heat transfer coefficient over the whole plate is actually double the value at the back end of the plate.  Such is the effect of that higher temperature gradient where the boundary layer is thinner.  It is this average for the whole plate that I used. 

Essentially a higher velocity reduces the boundary layer thickness and so increases the heat transfer rate, as the calculations showed.

In those calculations, the properties of air in the boundary layer are taken at the temperature mid way between the wall temperature and the free stream temperature.  The text book simply used the nearest temperature which was tabulated in the tables. When I increased the wall temperature in the calculations, the rounded out mid temperature was the same in each case.  Just to better understand the influence of temperature, I tried using the air properties for the next 50 degree temperature interval in the tables.  It only affected the fourth significant figure in the answer, so the text book simplification of using the nearest temperature for which the values are tabulated, very desirable in terms of avoiding the work necessary to interpolate all the values, seems fully justified for a calculation that is probably only within 25%.

Redoing all those calculations on a slide rule, or even with a calculator, would have been quite tedious.  However, with a spreadsheet, I first corrected the equation with the error, then copied the whole block and pasted it along side, to get a second true copy.  Then it was a simple matter of changing the figures in the cells for velocity or temperature, and the alternative calculation was complete in a flash.  All the work is in the first run through of the calculation, and exploring variations is a very simple exercise on a spreadsheet.

I will be checking in, but not making longer posts tomorrow and over the weekend, as our daughter and her family will be visiting, 6 in a very small space will make for a busy and distracted weekend.

I hope the frogs turned out on cue so there were plenty to count,

Thanks to all for looking in

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

Offline paul gough

  • Full Member
  • ****
  • Posts: 404
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #754 on: March 09, 2018, 01:04:38 AM »
Hi MJM, Glad you discovered the error. I am assuming you are saying one side of the firebox is about 3 W. loss. So the total area would be of the order of 10 W. I'll be very interested to see an estimate of the radiant losses. Finally did a reasonably accurate sketch of the firebox, trust it is of use.
 
When I get to rebuilding the loco I will renew the insulation inside the firebox, this may get the wall temp. down to something akin to the lower figure you used. By then I might even have a thermocouple to get a reading. Awful morning at the swamp, endless heavy rain and one of those "buggar all" trips in terms of frogs, but plenty of waterfowl so we spent a couple of hours counting/identifying reptiles with feathers. Regards, Paul Gough.
« Last Edit: March 09, 2018, 01:12:51 AM by paul gough »

Offline steam guy willy

  • Full Member
  • ****
  • Posts: 2227
Re: Talking Thermodynamics
« Reply #755 on: March 09, 2018, 02:54:04 AM »
Hi MJM, we talk about the heating up of the boiler and can calculate the amount of neat needed to produce steam  but can we also calculate the amount of extra heat needed with the insulation that we may use when we construct a boiler...and do they do this in industry ?

willy

Offline MJM460

  • Full Member
  • ****
  • Posts: 902
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #756 on: March 09, 2018, 11:17:16 AM »
Hi Paul, it is nice to have found the problem.  Always unsatisfactory to feel that a result is not quite right, but be unable to track it down.  The answer now seems to be the right order of magnitude.  So the next thing is to understand how the air flow past the firebox actually looks.  Is that plate in a clear airflow?  I can see it is tucked in under the boiler.  I assume the wheels and engine are also under or just outside the boiler in plan view, and these will "drag" the air along, and also add extra turbulence.  Also the engine cylinders must be sitting in there somewhere.  The heat loss from those cylinders is even preheating the air, some of which will enter the firebox.  I think I mentioned these factors a few days back, but you were travelling, and may have missed some of those posts.  All these factors reduce the overall accuracy of any prediction. Turbulence makes the heat transfer better, but preheating due to cylinder and boiler losses reduce the loss from the firebox by reducing the temperature difference to the air.  And for the front of the firebox, the air is probably travelling close to loco speed, so nearly stationary relative to the front face.  Back to natural convection.

Insulation on the inside would reduce the loss significantly, but if that firebox is soldered to the boiler, it is a bit like a heated fin, so again there are pluses and minuses whether the insulation is inside or outside. 

At the end of the day, it all has to be kept in proportion.  So now we need to know how much fuel is being burned in a given time to see how the burner heat release compares with that initial estimate of the losses.  Similarly, if the water consumption, approximately equal to steam production, is measured, along with the time of actual steam production, we can calculate the heat input to the steam, if we know the pressure, so with the three main factors we can easily see their relative importance.  I usually measure the quantity of water used to fill the boiler, then extract the remainder with a syringe at the end of the run to determine the quantity of steam produced.

Unfortunate about the early morning weather, but I guess expected at this time of year.  However, the feathered reptiles also have to be counted, so not all is lost.  I hope you had a good rain coat!

Hi Willy, insulation is put around the furnace casing to reduce the heat losses so the boiler actually requires less heat, both for heating up and for steam production.  Of course when the boiler is started from cold, some heat is stored in the insulation which has a temperature profile due to the heat conducted through it, but that heat is already lost from the furnace, so I believe it would come from the lost heat, it would not require additional heat from the burner.  However, even insulation has a density, thermal conductivity and specific heat.  So as long as the material is known, and the temperatures of the inside and outside walls calculated or measured, it is possible to calculate how much heat is stored in the insulation.  It is only temporarily stored in the insulation, then finally lost when the boiler cools down.  Initially, when the boiler fires up from cold, the furnace heat transfer is determined by the flue gas temperature, and the initial cold temperature of the furnace side of the insulation.  As the insulation warms, that inside temperature rises, thus reducing the temperature difference from the flue gas, so reduces the rate of loss to the steady state value.  In an industrial boiler, the heat up time is usually a very small part of the total operating time, so not considered important.  At least in the idustry I was in.  However for other industries where boiler operation is more intermittant, heat up time could be a consideration.

In an industrial boiler, I am sure the heat loss through the insulation would be directly or indirectly calculated, in order to determine the most economical thickness.  Obviously this requires data on fuel costs etc.  By indirectly, I mean that the designer may have specifications for a given size and operating conditions, which specify the thickness to use. However, the original writer of the specification would normally do some sort of heat loss and cost calculations to determine the thickness and type of insulation to specify.

We have a very full house here tonight, so just a short post.  I hope that answers the questions.  Back tomorrow.

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

Offline paul gough

  • Full Member
  • ****
  • Posts: 404
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #757 on: March 10, 2018, 03:57:40 AM »
As you say there is probably some shielding of the firebox by the cylinders, wheels and framing etc. and some heat floating around underneath to make precise analysis virtually impossible. Nevertheless a credible approximation for the open side of the firebox is nice to have and given a loco operating outside can be subject to 'chilling' low temperature wind from any direction, often blowing at a greater speed than that of the forward motion of the engine, I am inclined just to extend the result, say 4 W. by  X3 or X4 to give a 'worst case' projection that might be worth trying to mitigate, especially if you are able to contribute any further loss due to radiation to the final figure. These little gauge 1 locos have very little in reserve, anywhere, whether it be heat production, volumes of water or fuel, and of course losses are myriad. Add even a mildly adverse operating circumstance of cold windy conditions and the impact can lead to unsuccessful or problematic running. So, every avenue to an advantage or mitigating a disadvantage, if practicable, is likely to be worthwhile. As far as I'm aware, your analysis and quantifications are the only ones I have seen, and advance our knowledge in what might be actually going on in these tiny model circumstances. Investigations such as these may throw light on something that has not been noticed or ignored or is even counterintuitive to that applying on the larger scale, who knows, no numbers no knowledge. So I look forward to more. Regards, Paul Gough. 

Offline Zephyrin

  • Full Member
  • ****
  • Posts: 394
  • near Paris, France
Re: Talking Thermodynamics
« Reply #758 on: March 10, 2018, 07:51:45 AM »
Hi Paul,
A point I notice in the loco boiler drawing that you show is the dry firebox...not surrounded by water.
As you are rebuilding the loco, why not change that; I agree that this most probably requires a new boiler...
I also struggle with small Gauge 1 loco boilers to make them to produce a maximum of steam while saving alcohol at the same time, and not being to hot to handle at the end of their half hour run; a difficult equation to solve !

Offline MJM460

  • Full Member
  • ****
  • Posts: 902
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #759 on: March 10, 2018, 10:36:59 AM »
Hi Paul, it is always important to understand the assumptions on which any calculations are based.  You will remember from the sketch of the boundary layer development, that the air stream is shown  as the full free stream velocity at the front edge of the plate.  The very thin boundary layer at the front leads to a higher local heat transfer coefficient close to the front of the plate, which reduces along the plate as the boundary layer develops.  This leads to a higher average heat transfer coefficient for the whole plate.  Turbulence and heating prior to the front of the plate almost certainly increases the thickness of the layer, as well as reducing the temperature difference,  so reduces the overall average heat transfer coefficient.  Hence the calculations probably do, as you suggest, provide an upper estimate.  Especially for the front and back plates which are almost certainly nearer natural convection.  But on the other hand, a cold wind in almost any direction almost certainly increases the loss.  It is all about picturing the air flow.

Thank you for you encouraging words about the calculations though, I am glad they are providing useful insights.  I will try looking at radiation more closely when I am back at my bigger desk where I can spread things out a bit.

Hi Zephyrin, thanks for coming in on this one.  I am glad that you are still following along.  Water cooled fire boxes in the style of full size and larger models would help by both increasing the area for heat transfer for steam raising, and also limiting the outside wall temperature to steam temperature.  With a little insulation on the outside it would help enormously to reduce losses.  It would be interesting to know how you approach the issues of doing this in such a small size, especially staying the flat surfaces in metal that is presumably relatively thin, and how you manage to put it together.

Well, the little family had a great day and have moved on to their next gathering.  We went on a ten mile boat trip then after exploring the local market and buying a lamb roast from the farmer's stall, which we immediately put on the bar-b-que, and had it with cake and chocolate from the other speciality stalls as son-in-laws birthday celebration, before sailing back on a perfect sunny day with light winds.  Now we are exhausted, and the peace is wonderful.

Thanks for following along,

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

Offline paul gough

  • Full Member
  • ****
  • Posts: 404
  • Tropical Queensland, Australia
Re: Talking Thermodynamics
« Reply #760 on: March 10, 2018, 02:20:40 PM »
Hi Zephyrin, Many times in my idle moments I have thought about a true loco type boiler in gauge 1, but unless it is a large model or a freelance design accomodating the firebox within the framing for a narrow firebox type could be difficult. So for practical reasons I tend to think it is not viable, (though not impossible), unless one went for a modern engine design with a wide firebox type boiler where it was more or less above the top of the frames as in many of the modern U.S. locomotives and a sufficient number of wicks could be installed. There would be a larger heat output required for a wet firebox loco, two layers of metal and the water to heat up, so one would have to look at an existing model, say one of the successful larger Aster models of U.S. locos, e.g. the Great Northern, S2, 4-8-4 and guesstimate the increased heat demand and whether more wicks could be installed, I think there are 5 in the S2, so getting more in might be difficult if you tried to stick exactly to modelling the prototype. If you are into non main line locos it might also be practical to look at a narrow gauge type, such as the 600mm Fowler engines used in the sugar cane industry. The Fowlers had relatively 'fat' boilers and when built to 7/8ths inch scale give fair sized boilers, but again getting a wet firebox between the frames might be problematic. I have never thought about what might be possible with the newer types of gas firing with permeable ceramic grates where the fire is above it, it might be easier than trying to use methylated spirits and wicks. The other possibility is some adaptation of a Briggs type boiler which has a wet crown sheet but dry firebox sides but with a rectangular coil pipe inside, these are now very common in Australia in larger model gauges and very successful. However, it would be difficult to produce the relatively sharp bends in the coil without getting crush. There have been some models running without the coil and just well insulated dry sides. There is much room for experimentation in G1, but boilers are an area that holds little interest for most builders and there is now another layer of compromise, in that if the boiler is to be operated in anything other than your own shed with nobody else around the legal boiler codes can make innovative designs difficult or impossible to get certified. I have mentioned mostly the difficulties, but hopefully this might save someone some time. For myself, I have now reached the limits imposed by advancing maturity and the impact of disabilities so my construction capacities are mostly over, but I hope curiosity and the challenge spurs someone with more years available to try some innovative designing.  Regards, Paul Gough.

Offline steam guy willy

  • Full Member
  • ****
  • Posts: 2227
Re: Talking Thermodynamics
« Reply #761 on: March 10, 2018, 02:51:48 PM »
Hi MJM, Just some observations and practical questions about steam engines..................with an early Newcommen engine , all the heat used to create the steam (3/4 Lbs) was lost in the condensing process ....so could this be a termed a really efficient engine ?? rather than a modern loco where the Heated steam just whizzes through the engine very quickly and just Exhausts ??  !! These Newcommen engines were sometimes used for up to 100 years !!! Also thinking about Mr Farienhiehts thermometer and 0 Degrees being ice and salt ,what is the actual reason for this being very cold ??

Offline Zephyrin

  • Full Member
  • ****
  • Posts: 394
  • near Paris, France
Re: Talking Thermodynamics
« Reply #762 on: March 11, 2018, 11:26:09 AM »
Hi Paul,
The Aster models of club members I have had in hands for fixing parts doesn't have a particularly efficient boiler, simple pot or smithies boiler, in a casing lined with asbestos cardboard; and leaving little room for the hot burning meths from largely dimensioned wicks. It is the intense draught by the exhaust which allows this configuration to run properly.  The  alcohol consumption being large to huge.
As wagon pullers, these locos are generally good, as long as MJM460 is not looking at their efficiency...

As a model builder, I try to build more "efficient" locos, at large, according to common sense; this involves boilers internally fired, with a semi-wet firebox, smoke tubes and small wicks; it is also much more satisfying than doing a plain pot boiler.   
While reading this thread, I understand I have to do some measures in addition to ethanol consumption only to demonstrate that they are more efficient !

Quote
For myself, I have now reached the limits imposed by advancing maturity and the impact of disabilities so my construction capacities are mostly over, but I hope curiosity and the challenge spurs someone with more years available to try some innovative designing.
I got this point, I'm not a young man anymore...

Offline MJM460

  • Full Member
  • ****
  • Posts: 902
  • Melbourne, Australia
Re: Talking Thermodynamics
« Reply #763 on: March 11, 2018, 11:55:39 AM »
Hi Paul, I was also wondering about the practicality of a water surrounded firebox, but like you was thinking about the design and level of skill required to build it, and not confident of my own skill.  I will be very interested to see where this conversation leads.  I suspect with the boiler equivalent of watch making, it can be done, and due to the small areas which have to be supported, "thick" plates could be used to simplify staying arrangements that are not practical in a much larger model.  For example a firebox top machined from solid as an insert to the boiler to direct the combustion gases into the fire tube, so that more of the wick flame is enclosed by water.

Hi Zephyrn, I am most interested that you already have such a design and would like to learn a little more about what you have done, and what a water cooled firebox would look like in this size.  The main reason I have placed some emphasis on efficiency is that if we have a reasonable estimate of efficiency, and measure how much fuel we burn, we can use the two to get a reasonable estimate of the power output.  Though of course, generally in a model, unlike industry, we don't really know how much power we actually need.  So in the end, it is just about better understanding the science that enables our little models to turn heat to mechanical work.  Reducing losses does increase efficiency, but in this size range, I assume that we are not looking for efficiency for fuel savings sake, but more because it helps us get more power out of limited heating capacity, so helps tip the balance in favour of getting a working model.

And I suspect that many of us are that age where experience exceeds strength, and eyesight and coordination are not what they were.

Hi Willy, I suggest quite the opposite, those Newcomen engines would not be very efficient at all.  Let's quickly review the basics.  Efficiency is the energy out as useful work divided by the energy in from burning fuel.

Let's say the boiler is fed from the river at a balmy 15 deg C, it is heated to boil 3/4 psig, around 5 kPag.  I wish you had chosen a figure closer to 3 psig, which would have been directly in the steam tables, but, just for you, I did the interpolations, and found the heat required to raise that river water to about 102 C and boil it to produce dry steam is 2615 kJ/kg.

Now those engines are referred to as atmospheric engines, as the steam pressure only raises the piston against gravity, and then the water is condensed by spraying in more of that river water.  Atmospheric pressure, aided by gravity, then does work on the piston by pushing it down against the reduced pressure on the condensing side.  (That might take more than one reading!)

More water is pumped from the river into the cylinder, involving more losses in the work required to pump that water), the steam is condensed, the latent heat warms the river water to let's say 80 degrees, which determines the level of vacuum attained, about half an atmosphere or say 7 psia, and the warm water at 80 is discharged back to the river taking the residual heat of 80 degree water with it.

Now, let's compare that with an ideal adiabatic engine.  Steam at 3/4 psig is expanded to 7 psia which condenses at 80 degree C.  Using the second law, entropy is constant through an ideal expansion, otherwise it increases, the ideal engine would do work equal to 127 kJ/kg of steam.  A real engine say 75 % of that so 95.25 kJ/kJ, though more likely those early engines were much worse, is there any chance that you have a figure for steam per horses power?  Or even lb of coal per horse power.  Anyway 95.25 divided by the heat in, 2615 means 3.5% efficiency as an estimate of the best that could be achieved from those steam conditions and not achieved by many of our models.  The only way that will set the world on fire is if spilt fuel sets fire to the factory.  A modern ordinary industrial steam plant will be in the range of 30%.  Perhaps they were considered more efficient compared with non condensing engines of the time which received no input from atmospheric pressure to increase the minuscule output.

Well, if the energy does not get converted to work, where does it go?  Basically, it goes to the river as heat in that 80 degree water.

You might say that surely the vacuum will be lower, but I suspect I have been optimistic.  You see, it is limited by the pressure determined by the final temperature of the water and steam, now condensed, in the cylinder.  To heat river water from 15 to 80 C requires only 271 kJ/kg of water, while to condense steam at 80C after expansion from 102 C requires removal of 2590 kJ/kg of steam, so the spray pump must inject 9.6 kg of water for each kg of steam it condenses.   Well, perhaps river water is cheap, and they use 20 or 30 kg of water for each kg of steam, that would lower the pressure a bit,  but definitely diminishing returns.  And any air leaking in lowers the condensing temperature, but does not lower the pressure any further, unless you have a very good air pump.

Such engines were the foundation of the industrial revolution, but they worked by brute force, not by efficiency.  They could do the work of many men in a relatively small space, so the men were instead employed cutting down trees, or digging coal, and feeding it to the furnace.

I have tried to avoid those equations, which have all been presented before, and just given you the results.  But it is just data from the steam tables as before. Efficiency is about useful work out, compared with the heat input.  I hope this clarifies a little how those engines worked, and even a little more about the efficiency of modern engines.

As for salt in water, I am not sure of the mechanism, only that salt does indeed lower the freezing point of the mixture.  Note that I don't think adding salt actually cools the mixture, just lowers the temperature at which the mixture freezes when it is cooled.  I was always told that 0 F was the lowest temperature he could find (remembering he did not have any refrigeration, and also probably never travelled to a really cold climate, so he assumed, incorrectly as we now know, that this was the coldest possible temperature, therefor could reasonably be termed an absolute zero.  Only 460 deg out!

Thanks for looking in,

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

Offline Admiral_dk

  • Full Member
  • ****
  • Posts: 1436
  • S°ften - Denmark
Re: Talking Thermodynamics
« Reply #764 on: March 11, 2018, 03:23:31 PM »
With the risk of side tracking this thread, I will ask a question I've been thinking about just about every time I see a steam cylinder here. I this case it was Ricks Conway build that triggered it.

A lot of our model steam cylinders has a steam passage going from the steam chest to the same end of the cylinder made of two or more parallel holes of small diameter. As this give a much higher flow resistance than a single hole of bigger diameter or an oval one + a much bigger cooling area ...!!!... it really should give a much reduced efficiency and reduced power of the engine .... or is there something that works differently here than in other engine types ?