Steam & Preservation

One of the shibboleths is that steam is of low thermal efficiency, especially at the boiler pressures used with steam locomotives.  I have been wondering about the theoretical limits to thermal efficiency and how serious are the design restrictions of the steam locomotive, that one is restricted to a once-through non-condensing cycle and to about 300 PSI boiler pressure.  Sure, there have been condensing locomotives and locomotives of higher boiler pressure, but they were either rarities, unsuccessful on account of high maintenance or other non-fuel expenses and so on.

I also got to wondering, why steam?  What does a steam cycle get you that an air cycle engine does not?  Why was steam developed first, before the air cycle engines (Brayton, Diesel, Otto, Stirling)?

Just about all of the engine cycles are approximations to the Carnot cycle, where engine efficiency depends on adding heat at high temperature, removing heat at low temperature, and changing the working fluid between the low temperature, high temperature, and low temperature states by what are known as isentropic (constant entropy) or also as reversible cycles.  A reversible means of going from low to high temperature is with a pump that increases the fluid pressure; a reversible means of going from high to low temperature is with an expander or engine that decreases the fluid pressure.

Thus, the pressure ratio between the low and high temperature parts of the complete engine cycle is important for an efficient engine.  Also, to go from low to high temperature, one needs a kind of pump that does this without appreciable friction, thermal, or fluid leakage losses.  One big advantage of the steam engine is that this pump only operates on the fluid in its dense liquid state and hence is much easier (the injector on a steam locomotive) than, say, the compressor of a gas turbine, that needs to be very carefully optimized that its losses don't eat up all of the "profits" of the energy you get out of the turbine.

But again, if pressure ratio is the thing, the ratio of your boiler pressure over your input pressure limits efficiency, and high pressure boilers never caught on with railroad steam.

But is the pressure ratio everything?  Is the steam locomotive boiler pressure that low from a thermodynamic standpoint?  For about 15 PSI atmospheric, 300 PSI boiler, the steam engine "pressure ratio" is 20:1, which is actually considered high for a gas turbine: people build usable gas turbines with that much pressure ratio and less.

One thing the gas turbine people do is "regeneration."  As the pressure ratio goes down, the exhaust of the gas turbine gets even hoter, and the idea is to take some of that exhaust heat and move it to the high pressure side of the cycle.  Jerry Pier was talking about plate heat exchangers for regeneration of the next-generation railroad gas turbine, but there are other forms of regeneration.  One of them is the steam-injected gas turbine.  What you do is put a water tube boiler in the turbine exhaust, and then you inject the steam into the high pressure side of the gas turbine -- this thing becomes a kind of hybrid gas turbine-steam turbine.  They also use distlled water as their feed so as to not scale up the water tube boiler or the turbine blades, and some systems attempt to recover the steam from the turbine exhaust with a condensor and others are once-through.

If regeneration does such a good job for low pressure-ratio gas turbines, could it do the same for low boiler pressure steam?  Now one of the things the gas turbine people can do is to go to high temperatures at the turbine inlet -- the steam equivalent to this is to use high superheat temperatures, in fact, insane levels of superheat.  The gas turbine people, however, have an edge because they are still "internal combustion" even though they are "continuous combustion."  They generally have the high temperature on only one side of the combustor wall or of the turbine blade, and they employ air cooling of the other side.  A steam superheater is still "external combustion" and has the superheat temperature on the inside of the superheater tube and an even higher combustion temperature on the outside.  On the other hand, I had seen where Porta suggested a superheat temperature as high as 550 C (1000 F), and similar superheat temperatures are used in steam power plants.

So I ran some numbers with my trusty book of steam tables.  Consider a 250 PSI boiler (common in the Superpower era), 265 PSI absolute pressure, 406 F boiling temperature.  Superheat to 1000 F and then expand with 25 percent cutoff, discharge into a 5 psi back pressure (20 PSI absolute).  I calculate the work output as 291 BTU (all figures per pound of steam), the heat input as 1486 BTU, starting with 70 F water in the tender, giving a theoretical efficiency as 291/1486 or 19.6%.

At 25 percent cutoff, the steam in the cylinder, just before you open the exhaust valve and blow it down into the blast pipe, is at 45 PSI and 520 F.  Because I am proposing such a high rate of superheat, the steam is still superheated after expansion.  But because the steam at this point is hotter than the 406 F of the boiler water at 265 PSI, theoretically, theoretically I say and don't everyone get all hot and bothered quite yet for proposing some kind of magic feedwater heater that is able to recover all of the heat in the spent steam without creating too much back pressure and ruining the drafting of the blast pipe, theoretically one could pump the feedwater up to boiler pressure and use the waste heat of the exhaust steam to heat it up to boiler temperature.  In that case, the heat input to the cycle becomes 1144 BTU, the work output remains 291 BTU, and the efficiency increases to 25.4%.

Now with 25.4% cycle efficiency, your net efficiency of HP to the wheels will be a lot less because of all of the usual suspects -- pressure drop in the superheater header, heat losses, piston friction, etc, etc, etc.  But 25.4% is a long way from the 5% efficiency that was common in the heydey of steam locomotives, and by careful design, perhaps one could get a 15% efficient locomotive, using 1/3 the coal and 1/3 the water, which would have had important logistical advantages of running a steam railroad -- all this using a simple-expansion locomotive with conventional valve gear at 250 PSI boiler pressure.

But my "magic" steam locomotive depends on having a really good feedwater heater.  So my question to all of you is this:  what kinds of systems were used for feedwater heating/exhaust steam heat recovery back in the day of steam, and how effective were they?  I understand one system was simply a plate heat exchanger in the smokebox, another perhaps ran off the combustion gas/spent steam mix in the stack?  Was there something called an "exhaust steam injector" where one used steam tapped from the blast pipe to drive the feedwater injector and got some kind of heat recovery that way?  What were the relative merits of these different systems.  Thanks.

I always thought that the efficiency of a steam locomotive was measured from coal to output. You measured from boiler to output. Isn't a lot of the combustion inefficiency due to the privative boiler draft system? A 15% efficiency from boiler to wheels and an 30% combustion and boiler heat transfer efficiency would get you to the 5% locomotive efficiency - coal to drawbar.

Yes, I was assuming a 100% efficient boiler in calculating the theoretical thermodynamic efficiency of the cycle.

30 percent is perhaps a little low for the combustion efficiency, but combustion efficiency as low as 50 percent when a locomotive is "working hard" may not have been unreasonable (and sending a lot of the thermal content of the coal skyward as black smoke containing carbon particles). 

There was some kind of statement that a Big Boy went through 40 tons coal/hour climbing this one hill.  The coal they were using was low-BTU coal (kind of like the PRB coal -- low BTU yes, but also low sulfer), but even allowing for that, I think I once did a rough calculation that the overall coal-to-drawbar efficiency was something like 2 percent under those conditions.

Paul Milenkovic:

There was some kind of statement that a Big Boy went through 40 tons coal/hour climbing this one hill. 

Tender capacity was 28 tons, so they emptied it in 42 minutes... wonder how far they got in that time. Coal docks must have been closely spaced.

Steam power preceded thermodynamics and metallurgy by about a hundred years. The first steam power was derived mostly by the vacuum of steam condensing to operate pumps to de-water mines. Notice they didn't use windmills as was done in Holland. The first mobile use of steam was on steamboats, notably in this country the "Clermont" on the Hudson River about 200 years ago. The British used coal from the start but we used bio-fuel (fire wood) up until after the Civil War era. The early steamboat power-plant was very bulky for the power not to mention that volume of fuel required for such inefficient engines. Most early steamboats were primarily for passenger service and express cargo. Fabricating a boiler with blacksmith's tool was a problem also. I believe the early boilers were copper and they operated on something like 20 psi pressure and even at that boiler explosions were fairly common. The development of railroads starting in the 1840's really provided the wherewithal to develop the steel industry that took a lot of investment to build a plant on a large scale to make steel cheap. At first the railroads were the only market for steel. I think that thermodynamics became a science late in the 19th century. Before that it was all cut-and-try by gifted tinkerers.

Paul Milenkovic:

But my "magic" steam locomotive depends on having a really good feedwater heater.  So my question to all of you is this:  what kinds of systems were used for feedwater heating/exhaust steam heat recovery back in the day of steam, and how effective were they?  I understand one system was simply a plate heat exchanger in the smokebox, another perhaps ran off the combustion gas/spent steam mix in the stack?  Was there something called an "exhaust steam injector" where one used steam tapped from the blast pipe to drive the feedwater injector and got some kind of heat recovery that way?  What were the relative merits of these different systems.  Thanks.

There are two basic kinds of feedwater heater 

The heat exchanger type... and

The mixing chamber type.

The better known type is the heat exchanger type, best represented by the "Elesco" cylindrical unit that sat transversely across the smokebox, either in front or recessed into the top of the smokebox. This had the feed water under pressure in a miniature boiler with the exhaust steam passing through  heating the water. Elesco is from "LS Co." the Locomotive Superheater Company.

The mixing type basically adds the exhaust steam to the cold water in a chamber and the combined mixture is compressed and fed into the boiler. This has the greater efficiency, since there are no heat losses once the steam hits the water. The disadvantage is that there will be some lubricating oil carried over into the feed water. This is less of a problem in a locomotive where the majority of water is used only once, compared to a power station where the water is recirculated. The Worthington SA is a mixing type heater. The exhaust steam injector is sometimes called a poor man's feed water heater, and is a mixing type since the jet of exhaust steam adds heat to the feed water, and efficiency is improved since the heat in the exhaust would otherwise be wasted, compared to boiler steam being used in a live steam injector.

The last Canadian National 4-8-2s and 4-8-4s had exhaust steam injectors while earlier units had Elesco heaters. I guess they were trading theoretical efficiency against simplicity and first cost.

In Germany, the standard locomotives dating from the mid 1920s had heat exchanger type feed water heaters (called oberflachenvorwarmer) but post WW II both East and West Germany adopted mixing feed water heaters (called mischvorwarmer) as standard.

So you really seem to be looking at either a mixing type feed heater for maximum efficiency or an exhaust steam injector for simplicity and low first cost.

M636C

I did some more research.

Especially when you are using any type of steam injector for water feed, be it live steam (boiler pressure) or exhaust steam (blast pipe pressure), you want cold water as the feed.  I don't yet understand how the injector works (but the story is the patent examiner who handled the original patent application couldn't figure out either and demanded a working model), but the feed water is supposed to condense the feed steam to get the necessary suction in the device to get the feed water to squirt through the injector into boiler pressure, and if the feed water is too warm, it won't go.

So, the heat exchanger type feedwater heater must as you say be a kind of "mini boiler", or at least a pressure vessel, that acts on the feedwater at boiler pressure.

What I have also figured out is that one mixing type "feedwater heater", the exhaust steam injector, not only worked off exhaust steam, it also worked off boiler steam, and it had springs and check valves so it would revert to the correct steam supply depending on how hard the locomotive was working (amount of steam pressure in the blast pipe).  The exhaust steam, of course, was obtained where it was still under pressure, with a tap from the blast pipe, rather than where it was not under pressure, in the smokebox or stack after leaving the blast pipe.

The exhaust steam injector was, as mentioned, often called a "poor man's feedwater heater" as it was substituted for an exchange-type feedwater heater to avoid the expense of same.  But there seems to be a lot of controversy surrounding this device -- there are some who claim that the thermal performance of the exhaust steam injector was at least as good as the exchange feedwater heater, others who claim it wasn't and that it was a complicated "kludge" with all of its internal springs, check valves, and multiple expansion cones.

I still have some questions.  I think it is pretty clear that the exhaust steam injector took steam from inside the blast pipe where it was still under cylinder exhaust pressure.  Where did the exchange-type feedwater heater get its steam supply?  Does it need to pass all of the exhaust steam through the heat exchanger or does it divert a portion of the supply?  Where does it exhaust steam downstream of the heat exchanger?

The second question relates to Porta's idea of using the tender as a massive heat exchanger and store of heated feedwater.  If the tender water is heated appreciably, how do you get an injector, exhaust steam or live steam, to work?  I know that there are alternative types of feed pumps, but I always thought that expecialy with a locomotive boiler, you wanted both belt and suspenders, you wanted multiple and alternative types of feedwater pumps into the boiler as backup systems to each other.

Finally, for the open-type feedwater heater that was not an injector, where did they take the steam, discharge the steam, and was the feedwater under smokebox pressure or boiler pressure?

What I read for all of these types of feedwater heaters, you could get the feed water up to about 190-200 F, that is just short of atmospheric boiling temperature, but from a thermodynamic standpoint, that is a long way from boiling temperature, of about 400 F in my example.  The surfing on the Web on "regenerative Rankine cycle" suggests that an ideal regenerative cycle would heat the feedwater as it was pumped up to boiler pressure in a counterflow with steam being expanded in the engine, suggesting that you would want to heat the feedwater in the cylinder jacket.  As an approximation to the regenerative Rankine cycle, steam power plants using compound expansion use steam from the intermediate pressure receiver between compound stages to run their boiler injector.

As I suggested, there is "money on the table" for running a thermodynamically effective feedwater heater, raising efficiency for a theoretical cycle from 19 to 24 percent, and the gains may even be bigger when the engine is "working harder" with a larger cutoff percent, where more heat is being dumped into the cylinder exhaust at higher pressure and temperature.

As to the question regarding whether a Big Boy goes through 40 tons coal/per hour, I believe I got that number from Perfecting the American Steam Locomotive.  I looked up How Steam Locomotives Really Work on Google Books, and they gave a figure of 10 tons of coal/hr and evaporating 10,000 gal water/hr.

The 40 tons coal/hr works out to about 2 percent efficiency on low-BTU coal at 6000 cylinder HP.  I believed that number at the time because if the rated HP is generated at large cutoff (hard working), the cycle efficiency could be around 4 percent, and then if you are generating those large billows of picturesque exhaust smoke, your boiler could be operating at 50 percent combustion efficiency.  For the 10 ton coal-10,000 gal water/hr claim, if the locomotive is generating 6000 cylinder HP, the cycle efficiency works out to about 14 percent, with the coal-to-HP efficiency at 7 percent and again, the boiler at 50 percent combustion efficiency.

Paul Milenkovic:

Yes, I was assuming a 100% efficient boiler in calculating the theoretical thermodynamic efficiency of the cycle.

30 percent is perhaps a little low for the combustion efficiency, but combustion efficiency as low as 50 percent when a locomotive is "working hard" may not have been unreasonable (and sending a lot of the thermal content of the coal skyward as black smoke containing carbon particles). 

There was some kind of statement that a Big Boy went through 40 tons coal/hour climbing this one hill.  The coal they were using was low-BTU coal (kind of like the PRB coal -- low BTU yes, but also low sulfer), but even allowing for that, I think I once did a rough calculation that the overall coal-to-drawbar efficiency was something like 2 percent under those conditions.

 

 

Dear Paul,

if you do not have Kratvilles Big Boy book, the test-results show a coal/water ratio from 4,5% to 4,7% efficiency, while the 3 runs reached an avarage of 20 mp/h. If you mean boiler efficiency of BB, I read somewhere they were round about 60%.

If you need exact numbers of this book, I will look up for you.

Kind Regards

lars

UP purchased coal for Wyoming Division from Rock Springs, Superior, Kemmerer, and Hanna mines, which was a mid-vol bituminous that varied from 10,500 to 12,000 BTU.  I think most people seem to think that UP purchased subbituminous.  What the specific BTU of the coal into the tender might have been is anyone's guess now (unless one can look at the records).  Often railways bought the worst possible coal out of the mine, full of bone and shale, whereas the "seam value" might be washed and graded coal.

RWM

Dear Railway Man,

at UPRR's homepage, a table of current operated Wyoming coal mines (here: http://www.uprr.com/customers/energy/coal/coalspec.shtml ) show a value of 9700- 9800 btu/lb.

They are not necessarily the mines that supplied fuel for steam-engines (Hanna / Rock Springs is not listed anymore), but give a course of how low the grade of coal in this area is. Anybody wonder now, why some steam engines smoked so badly?

Cheers,

lars

 

Paul,

if you read somewhere BB used 40t coal, that was the overall consumption from Ogden to Wasatch/Evanston.

BB used 10 tons coal per hour, with 28t in the tender this was ample load for a 1 hour call, then covering the 40miles to Echo within one hour and 40 min and a reserve for occasionally stops on sidings for more priority trains or traffic. At Echo, they were reloaded within 20min, then they worked hard up the grade, including 15miles or so on continious 1,14%  to Wasatch, covering 30 miles in 2 hours. Engine 4016 exerted a cont. TE of up to 115.000lbs at 14mp/h with almost 4000tons freight. The av. produced DBHP Ogden - Wasatch was ~3700, maybe 4100-4200 cyl. HP.

That was BB's run.

Cheers

lars

Lars Loco:

Dear Railway Man,

at UPRR's homepage, a table of current operated Wyoming coal mines (here: http://www.uprr.com/customers/energy/coal/coalspec.shtml ) show a value of 9700- 9800 btu/lbs.

They are not necessarily the mines that supplied fuel for steam-engines (Hanna / Rock Springs is not listed anymore), but give a course of how low the grade of coal in this area is. Anybody wonder now, why some steam engines smoked so badly?

Cheers,

lars

 

Paul,

if you read somewhere BB used 40t coal, that was the overall consumption from Ogden to Wasatch/Evanston.

BB used 10 tons coal per hour, with 28t in the tender this was ample load for a 1 hour call, then covering the 40miles to Echo within one hour and 40 min and a reserve for occasionally stops on sidings for more priority trains or traffic. At Echo, they were reloaded within 20min, then they worked hard up the grade, including 15miles or so on continious 1,14%  to Wasatch within 2hours, covering 30 miles. Engine 4016 exerted a cont. TE of up to 115.000lbs at 14mp/h with almost 4000tons freight.

That was BB's run.

Cheers

lars

So 115,000 lb tractive effort at 14 MPH works out to about 4300 HP -- somewhat less than the max 6000 HP, but to be expected at "grade climbing speed."

10 tons of 10,000 BTU/lb coal at 4300 HP works out to about 5.5% coal-to-tractive-effort efficiency -- about in the ballpark of those mentions of efficiency in the mid 4% range.

Evaporating 10,000 gal/hr of water to 300 PSI boiler pressure (315 PSI absolute) to 600 degree-F superheat temperature, about 1300 BTU/lb water according to my steam tables, assume 150 BTU/lb heat recovery from an exhaust steam boiler feed injector, works out to 11.9% thermodynamic efficiency.  This puts the boiler efficiency at 46%, low compared to the reported value of 60%, but perhaps reasonable under conditions of "hard working" where one probably is losing some fuel as carbon particles in the exhaust smoke.

For 315 PSI (absolute) boiler pressure, 600 deg-F superheat temperature, 50 percent cutoff, 20 PSI (absolute) exhaust back pressure, the theoretical thermodynamic efficiency computed from steam tables works out to 14.2 percent.

For the actual thermodynamic efficiency (steam-to-tractive-effort) to be within 80% of the theoretical thermodynamic efficiency seems very good.  For the boiler to be operating below 50 percent efficiency (below the 60 percent value quoted earlier but above Don's estimate of 30 percent for some steam engines) seems not so good, and perhaps the best place to work on improvement.

Dear Paul,

with big interest read your calcs. We also have to consider some variable factors (cold weather, slippery rails occured, non so feasible conditions). They also used 3 different throttle/cut-off settings during the tests, the lightest train at 3500tons showing better fuel consumption results. This fits nicely to the originally task of taking a 3600tons train unassisted to Wasatch and not to overload steam-locomotives.

We only may ask, what did they had in their mind to set the ratings at Wasatch up to 4450tons later? Any plausible explanation?

Cheers

lars 

 

 

Lars Loco:

Dear Railway Man,

at UPRR's homepage, a table of current operated Wyoming coal mines (here: http://www.uprr.com/customers/energy/coal/coalspec.shtml ) show a value of 9700- 9800 btu/lb.

They are not necessarily the mines that supplied fuel for steam-engines (Hanna / Rock Springs is not listed anymore), but give a course of how low the grade of coal in this area is. Anybody wonder now, why some steam engines smoked so badly?

Cheers,

lars

 

Dear Lars:  I am sorry, Wyoming is not a uniform state with uniform coal values.  Nor does any U.S. state produce uniform coal with uniform values.

The Wyoming mines you have cited above are in the Gillette Field of the Powder River Basin Coal Region, and those mines produce sub-bituminous B or C coal of 8,500 to 9,800 BTU -- mostly more in the 8,500 range.  These mines never produced coal for railway fuel for Union Pacific Railroad, and these mines produced none for any railway locomotive at all as none of these mines even existed until the 1970s. There was some mining in the Gilette field prior to the 1970s, particularly in the Glenrock area, but it was not a major source for railway fuel except for some branch lines and short lines in the area.  

Union Pacific Railroad purchased its railway fuel from mines extracting coal from the Rock Springs and Hanna Fields of the Green River Coal Region, and the Kemmerer Field of the Hams Fork Coal Region, which is an entirely separate, unrelated, geologic formation, located several hundred miles distant.  These mines typically produce a high-volatile bituminous C or mid-volatile bituminous coal with heating values in the 11,500 to 12,500 BTU range.  These mines produced from circa 1870 to present, though most of their eastbound utility market disappeared after 1985.

Any conclusions drawn about locomotive performance on UP using the characteristics of the coal you cite above would be incorrect conclusions, but I feel like I am tilting against railfan windmills here.  If you really want to know about railway coal use, and its relation to railway locomotive performance, read the book "Railway Fuel," by Eugene McAuliffe, who was the president of the Union Pacific Coal Company (many railways owned their own mines).

Note that Northern Pacific Railway did use Powder River Basin coal extracted from a company mine at Rosebud, Montana, which is the poor-coal end of the basin with values in the mid-8000 BTU range.  This is a sub-bituminous C coal but railfans often incorrectly call it lignite.  Lignite is a lower grade of coal yet, typically with BTU values in the 6,000 range.

RWM

Thank you RWM for correcting this.

You are right on, I did not looked up where these mines sit now, but was not aware of  those great regional differences of quality of coal in Wyoming/Utah.

I just cited those values, to have a start for Paul's calculations.

Yet, we still do not know what coal was used (maybe UP's coal companies sold the better coal for customers, leaving the worse coal for their engines?).

I do not think that would lead to so false conclusions at all:

Paul calculated 5,5% efficiency with 10.000Btu/lb., with ~12.000Btu/lb, those mentioned 4,5 - 4,7% efficiency would be still in the ballpark, wouldn't it?

Paul, have you interest to re-run your calcs with 11.500Btu/lb. and 12.500Btu/lb.?

 

Cheers

lars

OK.

10,000 gal water/hr evaporated times 8 lb/gal times (1314 - 150) BTU/lb = 105 million BTU/hr.

The number 1314-150 is the enthalpy of steam at 315 PSI absolute (300 PSI gauge), 600 deg-F superheat temperature, minus the enthalpy of water at about 180 degrees, which I am assuming is the temperature of the feedwater with whatever feedwater device they had.  Enthalpy is this fancy name for the energy in a system at constant pressure but you allow the volume to change -- when you boil water, you want to know the enthalpy change because boiling takes place at constant pressure but allows the volume to change.

Assuming 12,000 BTU/lb coal

10 (short) tons times 2000 lb/ton * 12,000 BTU/lb = 240 million BTU (per hour)

 

That means the combustion or boiler efficiency is 105/240 = 43.8 percent.

I remember some time in the late 1960's living in Boulder, Colorado and hanging out with the local model railroaders and railfans, and hearing something about a "steam ejector" that "could have saved steam."  Not to get the "should have kept steam longer" debate going again, but you have to remember that 1969 was just 10 years after Norfolk and Western dropped the fire on their last Y-class, and Colorado narrowgauge steam was a going concern until the mid 60's?   

Years later I think they were talking about Kylchap or Giesel or whatever kind of improved exhaust drafting nozzle, and I always wondered how what I considered to be minor modifications to the boiler draft could make that big a difference for what was inherently a low boiler pressure, non-condensing, inefficient steam cycle.

But if you are running your boiler at 44 percent combustion efficiency, there is a lot of money left on the table, or perhaps a lot of money simply going up the stack.  One school of thought is that if steam was to have modernized and kept up with the emerging Diesel competition, one would have had to go to what the electric utility people did: high boiler pressure, condensing, steam-turbine electric drive, and so on -- the closest thing was that GE condensing steam turbine that ran on the UP and other places in the 1940's.  The other school of thought is that even within the confines of an "inefficient" thermodynamic cycle, improvements in the boiler draft, combustion arrangement, boiler and cylinder insulation, steam circuit pressure drops, that the accumulation of such "little" changes could bring large improvements.  Chapelon and later Porta and Wardale were in that camp.  Simply bringing efficiency from 4% to 8% would cut coal and water use in half, with important logistical savings in supplying that much coal and water out on the railroad lines.