In Part Two, I set forth the criteria for an airship engine, briefly discussed some of the options, and explained the advantages and disadvantages of the internal combustion engine. Now we’ll examine the principal alternative, steam propulsion.
Small steam engines are built in the 1632 universe as early as January 1632. (Bergstralh, “Tool or Die,” Grantville Gazette 9). As far as I can tell, the first steam locomotive is in operation in late 1632. (Flint and Zeek, “The Suhl Incident,” 1634: The Ram Rebellion). Steam propulsion never caught on for aircraft, because the powerplants had a low power-to-weight ratio. Nonetheless, the Danish Royal Anne is reportedly equipped with “mono-tube steam generators, condensers, and uniflow steam engines. ” (Evans, “No Ship for Tranquebar, Part Two,” Grantville Gazette 28).
There is no doubt that airships can employ steam propulsion; the 1852 Giffard airship stands as proof. Rather, the question is whether steam engines will be competitive with internal combustion engines. A large airship is likely to cost considerably more than even a naval fleet flagship. Consequently, investors (private or public) will need to be persuaded that the proposed propulsion system will provide the desired performance.
In the competition for funds, steam proponents will be at a disadvantage. While steam locomotives enjoyed more than a century of success, the same cannot be said of steam airplanes or airships. Hence, one cannot make the argument, “it made lots of money in the old time line.”
The three great potential weaknesses of steam propulsion for airships are low thermal efficiency, low power-weight ratio (relative to gasoline but not diesel engines), and high maintenance costs. (I assume that steam engines for airships will be of the condensing type, i.e., will recycle their water, and thus water supply will not be an issue except as the condenser affects weight and maintenance). In this article, we will explore just how serious these weaknesses are, and whether they can be alleviated.
Steam propulsion also has its strengths, notably low initial costs, an ability to use a greater variety of fuels (including coal, wood, and heavy oil), and insensitivity of power delivered to altitude. (Odom). However, the fuel diversity advantage comes with the caveat that use of fuel of low energy density means that to maintain range one must carry a greater weight of fuel (and thus less payload).
There are railfans and steamheads in Grantville, and their personal libraries may well provide useful technical information about twentieth-century steam locomotive and car designs. However, this is something of a double-edged sword, as it will reveal that the steam locomotive was eclipsed by diesel-electric and straight electric locomotives because of higher fuel, watering and maintenance costs, and that steam cars were likewise defeated in the marketplace by gas-powered cars. (The steamheads may argue that advanced steam technology would have overcome these problems, but the financiers could still reject their proposals as too speculative.)
Thermal Efficiency of Steam Propulsion
Stationary steam powerplants can achieve very high efficiencies, perhaps 50% (Semmings 162), but they also use much higher pressures and boiler temperatures than any vehicular plant, and are equipped with subatmospheric condensers and other refinements that add weight and cost.
The steam locomotive evolved in a period in which fuel (wood, coal or oil) was cheap and therefore little concern was given to thermal efficiency. As Ennis (354) aptly states, “the aim in locomotive design is not the greatest economy of steam, but the installation of the greatest possible power-producing capacity in a definitely limited space.”
In The American Diesel Locomotive, Solomon explained, “During the nineteenth century, the best steam engines operated at about 6 percent thermal efficiency, a figure that climbed to 10 to 12 percent by the end of the steam era.” (13). Even that figure was tarnished by his The American Steam Locomotive, which said, “At optimum performance, the modern steam locomotive can theoretically produce a maximum of 12 percent thermal efficiency, but 6 to 8 percent is the maximum achievable in actual operation.” (120).
We need to be very precise as to what these numbers mean. Usually, steam locomotive efficiency is quoted as overall drawbar efficiency. That’s the product of (1)–(6) below.
Table S1: Efficiency Definitions
1) Boiler Combustion
(a) heat in combustion gas in firebox / heat content of fuel fired
2) Boiler Absorption
(b) heat in steam leaving boilers / (a)
(c) heat in steam to cylinders / (b)
(d) piston work / (c)
(e) wheel rim work / (d)
Efficiency may also be quoted as the overall wheel rim efficiency, which is the product of just (1)–(5). I have overall wheel rim efficiencies for the South African Railway steam locomotives 19D (4.0%), 24 (3.9%), 25NC (3.2%), GMA/M (2.4%) and 26 (4.8%). In contrast, the SAR diesel locomotives scored much higher: 31 (23.1%), 32-000 (22.2%), 35-000 (23.0%). (Wardale 44, 290).
Since we would be hooking the steam powerplant up to a propeller shaft, we need to find the overall cylinder efficiency, the product of (1)–(4).
In the mid-twentieth century, a typical steam locomotive had an overall boiler efficiency of 72%, a cylinder efficiency of 14%, and transmission efficiency of 90%; that would imply an overall cylinder efficiency of at best 9%. (Cox 178).
It is interesting to see efficiency breakdowns for some real (standard and second generation) and hypothetical (second and third generation) locomotives (Table S2). Steam proponents are quick to urge that the actual locomotive numbers can be bettered. Porta envisioned “second generation steam” (SGS) as featuring 290–362 psi pressure, 450oC steam, double expansion, a gas producer combustion system, feedwater treatment to minimize scaling, corrosion, etc., feedwater and combustion air preheating, and an advanced (Lempor) exhaust system; Porta prophesied that this could achieve 14% efficiency. His concept of third generation steam (TGS) added higher pressure (870 psi) and temperature (550oC), triple expansion, and multistage feedwater and combustion air heating, hopefully achieving 21% efficiency, or even 27% if a condenser were provided. (Rhodes). But this author is not inclined to put much faith in efficiency figures for hypothetical engines, and Wardale’s claims for his own hypothetical SGS modifications (grey columns in table S2) were more conservative.
|Table S2: Actual Efficiency Breakdown for Selected Steam Locomotives and Estimates for Proposed Modified Locomotives|
|C&O4-8-4(1946) J-3 (3)||SAR4-8-219D (5)||SAR25NC (6)||China QJ 2-10-2(1964–1988)(1)||Wardale (2)|
|Boiler Pressure (kPa)||1724||1380||1550||1550||1471||1600|
|Steam Temperature oC||405||331||366||450||400||460|
|1) Boiler Combustion||?||86||47||81||78%||87||95|
|2) Boiler Absorption||?||86||81||80||78.2||80||90|
|Overall Boiler (1)*(2)||45||74||38||65||61||67||85.5|
|Overall Cylinderproduct (1)–(4)||4.3||?||4.1||9.0||9.3||12.4||18.1|
|Overall Wheel Rimproduct (1)–(5)||?||7.4||?||?||8.3||11.6||17.0|
|Overall Drawbarproduct (1)–(6)||3.65||<6||3.3||7.8||7.8||11.0||16.3|
|(1) Wardale Tables 71, 78. (2) Wardale 501: 1978 estimate of maximum values to be expected from conventional, non-condensing, coal-fired steam locomotive (3) Wardale 397(5) Wardale 51,typical steaming conditions (6) Wardale 151, 184|
The numbers above present only part of the picture, as we have no idea of the circumstances under which efficiency was measured. It can vary depending on the feedwater temperature, the fuel, the firing rate, the actual (vs. design) boiler pressure, and the moment in the piston stroke at which the inlet valve is closed (cutoff). (Hudson).
The lowest efficiency in table S2 is the cylinder efficiency. Let’s look at what theory tells us is the highest possible value, to put this in better perspective.
Steam engines are true heat engines; their working fluid (water) cyclically receives heat from a high temperature source (boiler), converts some of the heat energy to work (steam-driven piston movement), and reject the remaining waste heat to a heat sink (condenser or outside air). The theoretical (Carnot cyle) maximum efficiency is expressed by the equation
with temperatures in oK.
All reversible heat engines have the same thermal efficiency when operating between the same two temperatures; all real-life heat engines have lower efficiencies than the corresponding reversible heat engines.
For a variety of reasons, real-life heat engines don’t even come close to approximating a Carnot cycle. The Rankine cycle is less efficient, but provides a better model of what happens in a real steam engine. A hidden limitation on the efficiency of the water-steam Rankine cycle is that some heat must be used to change water into steam (unless boiler pressure is supercritical) and this heat doesn’t do any useful work.
Calculating the efficiency of a Rankine cycle is much more complicated than doing so for a Carnot cycle, as you must consider the enthalpy and entropy of the steam at various temperatures and pressures, and steam under the typical conditions does not behave like an “ideal gas.” I have no doubt that the Grantville power plant (coal-fired steam turbine) engineers can make the admittedly complex thermodynamic efficiency calculations. All you need are steam tables (these exist in CRC and Marks’ Handbook) and knowledge of the thermodynamic relationships. I wrote my own spreadsheet and tested it against multiple thermodynamic textbook examples, so I know it can be done.
As shown in table S3, for a non-condensing locomotive with a mere 50 psi boiler pressure, even the theoretical efficiency is only a little over 6%. The theoretical efficiency of the Rankine cycle may be increased by a variety of means.
Three expedients—increasing boiler pressure, superheating the steam, and heating the feedwater to the boiler—were used reasonably frequently as the steam locomotive evolved.
Boiler pressure. To safely increase the boiler pressure, you need a higher tensile strength structural metal (steel is superior to wrought iron), a greater thickness of metal (thus, a heavier boiler), or a smaller diameter boiler. Looking at locomotive data available in Grantville, I see pressures of 50 psi (1830), 130 (1860), 180 (1880), 190 (1900). (Alexander). EB11 Table XII lists locomotives up to 235 psi, and Halberstadt, Working Steam, up to 300. “In practice, for locomotives there is an optimum in the range 200–300 psig.” (Semmens 152ff).
Steam locomotives were occasionally constructed with pressures above 350 psi, but none were particularly successful. Scale and corrosion problems increased, and distilled water had to be used in a closed circuit as in the Schmidt system. The increased acquisition and maintenance costs were found not to be justified by the efficiency gain. (HPSLT). However, historical steam car engines ran at higher pressures, typically 400–1200 psi. (Crank).
Superheating. This is heating the water beyond its boiling point, which not only increases efficiency, but also decreases moisture content, rendering the steam less corrosive. However, higher temperatures can weaken or even melt the steam containment structure. For carbon steel, the highest allowable temperature is 950oF. (Ganapathy). Superheating may also dictate use of high-temperature lubricants.
Using exhaust steam or combustion gases in the flue, you can increase temperatures perhaps 24–40oF, which is sufficient to dry the steam. For a greater effect, the superheater must be exposed to hotter temperatures. In an integral superheater, the superheater is inside the firebox, preferably in a zone with temperatures of at least 1000oF (the higher the temperature and the greater the superheater surface, the greater the heating effect). Or the superheater may be separately fired; it was found that the superheater could use a cheaper grade of coal. In early twentieth-century stationary plants, integral superheaters could achieve up to 300oF and separately fired ones at least 500oF (but the efficiency of the separate superheater was only 25%). (Ennis 428, Jude 252).
According to Marks, Mechanical Engineers’ Handbook 1216 (1922), typically up to 250oF (139oC) superheat was used on locomotives.
Feedwater heating. Exhaust steam may be used to preheat the feedwater for the boiler. Preheating to 150oF theoretically increases cycle efficiency by 0.7% in a 105 psi boiler and 1% in a 285 psi boiler. (Lamb 38). Exhaust steam at best increases the feedwater temperature to the atmospheric boiling point (212oF) (Ennis 429); that is, it causes a non-condensing engine (open cycle) to have the efficiency of an atmospheric condensing (closed cycle) one.
A second option, used in the “economizer,” employs the waste gas from the furnace. This can achieve a feedwater temperature of 300oF or more. (Ennis).
Steam may be deliberately bled off at an intermediate expansion state to heat the feedwater (regenerative heating) instead of being used to do work. While less work is done, the heat is used more efficiently. Stationary powerplants may have a series of feedwater heaters at different bleed points.
In an open feedwater heater, the steam is mixed directly with the feedwater (or condenser water), and in a closed one, heat transfer is permitted without physical mixing. For the former, the feedwater temperature achievable is limited by the saturation temperature corresponding to the pressure of the hot water pump, whereas the latter requires scaling up the heat exchanger tubes. (Wardale 157). Also, the former captures lubrication oil from the cylinder and mixes it into the boiler feedwater, whereas the additional piping of the latter is prone to leaks.
Both types were used on locomotives; the “Worthington” was open and the Elesco and Coffin were closed. (Barris). The weight of a feedwater heater was perhaps one ton for a 200 ton locomotive, and it increased sustained boiler capacity by 15% (closed) or 17% (open). (Wardale 156). While the feedwater heater was invented earlier, only three units were sold in the USA before 1919. “By 1936 heaters were in use on perhaps a fourth of all steam locomotives in services, and were built into all new steam locomotives.” (Hultgren 224).
Table S3NC shows the effect of different boiler pressures and degrees of superheat, and in one instance feedwater heating, on theoretical cycle (cylinder) efficiency and steam quality for several actual locomotives (the first two are in Grantville Literature), as well as some hypothetical variations.
Table S3NC: Theoretical Rankine Cycle Efficiency and Steam Quality in Cylinders for Various Operational Cases (Non-Condensing)
|Super-heatoCoF||fwoCoF||Qual%||Theor Eff%(actual in italics)|
|“Best Friend of Charleston,” 0-4-0,1830, Alexander plate 2|
EN: “Emma Nevada” (formerly “Sidney Dillon,” 2-6-0, 1881, Alexander plate 74
Schenectady No. 3 (Goss)
Pa. RR E2A, #5266, 4-4-2, 1904
#5266+250F superheat+OFWH (520 kPa, extract 200oC steam)
QJ without feedwater heater
QJ with feedwater heater, extraction assumed at 700 kPa
hypothetical locomotive, BP 300
(Exhaust is at 170.3 kPa, 10 psig. (Lamb 36). QJ locomotive had a feedwater heater but the extraction pressure is not reported and so efficiency calculation ignores it; the other locomotives didn’t have one.)
Several other expedients were rarely or never used in locomotive practice, but could be applied to an airship engine (with varying degrees of practicality).
Condenser pressure. Generally speaking, locomotive steam engines weren’t equipped with condensers and thus operated on an open Rankine cycle. The few exceptions were those operating in arid regions or close to a front line (where the concern was that the exhaust plumes could be spotted by enemy aircraft).
The 1917 Stanley Steamer had a primitive condenser (closed Rankine cycle) that operated at normal pressure thus the heat sink temperature was 212oF.
Reducing condenser pressure further (by cooling) lowers the “rejection” temperature and thus improves efficiency. It unfortunately increases moisture content and thus both drag force (Vosough) and corrosion. The condenser pressure cannot be lower than “the saturation pressure corresponding to the saturation temperature of the cooling medium” and because of heat flow considerations will need to be higher. For example, a stationary plant using water cooling might condense to a pressure for which the saturation temperature is 10oC higher than water temperature. (Cengel 522). Ennis (332) says that “condensing water is generally not available at temperatures below 60o or 70oF.”
Bear in mind that the lower the pressure, the more susceptible the system is to leakage; in the early 1900s, the minimum condenser pressure was perhaps 1 psia (Ennis 318) and that’s still a practical limit (Vosough). The initial cost also increases nonlinearly; in 1906 a 27″ vacuum (0.147 psia) costs 20% more, and a 28″ (0.98), 57% more, than a 26″ (1.96) (Jude 246).
While condensers are common in stationary plants, “it is much harder to accommodate an adequate condenser on a locomotive which has to be mobile and whose size is constrained by the loading-gauge.” (Semmens 155). Also note that “there is an element of danger involved in returning condensed steam from reciprocating engines to the boilers, on account of the cylinder oil it contains.” (Ennis 431).
Unlike condensers for stationary or marine plants, which dump heat into rivers or the ocean, those for land or air vehicles cannot rely on water as the ultimate cooling medium; the water remains on board, and would get hotter and hotter, eventually turning to steam itself. Rather, the water is used just as a heat exchange medium, transporting the heat to a radiator which dumps it into the atmosphere. The airship condenser and radiator, in combination, must—in sustained operation—continue to remove enough heat so as to restore steam to its liquid state, whereas a car radiator only must get rid of enough heat so that the engine doesn’t overheat.
But air, at sea level temperature and pressure, has only about 1/4 the heat capacity per unit weight and 1/3,500th the heat capacity per unit volume. And the heat conductivity of air is less that 1/10th that of water, which reduces the efficiency of heat exchange. (Valentine). The faster the engine is running, the greater the steam flow rate and therefore the greater the cooling rate needed.
Normally, a locomotive will use the exhaust steam, funneled into blast pipes, to create a strong draft and thereby achieve a high rate of combustion. With a condensing plant, you need an alternative source of forced draft, such as (on the SAR 25C) a turbine driven by the exhaust steam before it was condensed. (Smith).
Reheating. At an intermediate stage of expansion, the steam can be reheated by passing it in fire tubes back through the boiler, thus restoring it to the inlet temperature. This dries out the steam, improving its quality, and also increases efficiency slightly. Reheat was first used in a stationary power plant in the 1920s. Reheat was used in the SNCF 160 A-1 (1940) compound expansion locomotive; the reheater was essentially a Schmidt superheater.
For a single reheat, the optimum extraction pressure from an efficiency standpoint is usually at about 20–25% of the boiler pressure. (Srinivas). Ideally, the steam is extracted when it is no longer superheated but still of high quality (low moisture). Consequently, a second reheat stage is typically used only in plants operated at very high (supercritical) pressures (3500 psia). The efficiency improvement from a second stage is likely to be something like half that of the first stage. (Logan, 443).
Table S3C shows the effect of condenser pressure, reheating, and/or feedwater heating on locomotives modified for condensing operation, on the Stanley Steamer, and on the Besler Steam Plane, the last being the most efficient.
Table S3C: Theoretical Rankine Cycle Efficiency and Steam Quality in Cylinders for Various Operational Cases (Condensing)
condensing variations on locomotive #5266
condensing variations on locomotive #5266+250F superheat
Stanley Steamer and variations
Besler Steam Plane
(If reheat, to boiler temp.)
Remember, cylinder (cycle) efficiency is only part of the picture; boiler efficiency is very important.
Of course, a real life steam powerplant doesn’t achieve the theoretical cycle efficiency because of various “irreversibilities,” such as fluid friction, incomplete absorption of heat from the combustion gases, and heat losses from the steam to the surroundings (Ennis 301ff, Cengel 519ff). For stationary plants, the actual efficiency is typically 40–80% (usually 50–70%) of the theoretical one (Ennis 398); figure 60–70% if high superheat is used. (399). In table S3, in the case of the locomotives for which we have both actual and theoretical cylinder efficiencies, the former was 68–85% of the latter. I am inclined to assume an actual/theoretical efficiency ratio of 70% for back-of-the-envelope calculations.
There are several expedients for mitigating irreversibilities:
Jacketing. When steam enters the cylinder, it meets a relatively cold surface, and condensation can occur, speeding heat loss and also increasing corrosion. Hot air or steam may be introduced into an annular casing surrounding the cylinder, to warm the cylinder walls, and prevent this. The net reduction in steam consumption is on the order of 2–15%. (Ennis 307); in terms of efficiency ratio, the increase is 3–5 percentage points. (399).
Compound expansion. In compounding, the steam is expanded in a series of cylinders. In double expansion, the exhaust of the high pressure cylinder is fed into the inlet of the low pressure cylinder. Triple and quadruple expansion engines are analogously defined.
Compounding has no effect on the theoretical efficiency, but does increase the real-life efficiency, by reducing heat losses through the cylinder wall. (Ennis 319). Ennis (399) reported that for saturated steam, the efficiency ratio increased from 40% (simple) to 50% (double) and 60% (triple) for non-condensing engines, and from 60% (simple) to 65% (compound) and 80% (triple) for condensing ones. Unfortunately, compound engines are larger than their simple counterparts. (Semmens 162).
In stationary plants, compounding is often combined with reheating for increased efficiency.
Superheating also reduces condensation, and a simple engine using superheated steam could be as economical as a compound with saturated steam. (Ennis 402). Superheating also obviates the need for jacketing. (405).
Uniflow. In a normal steam engine, steam enters the cylinder at one end and, expanding, leaves at the other, and then the direction of steam expansion is reversed. In a uniflow system, steam always enters the cylinder at an end but is exhausted from the center. This means that the ends always stay hot and the center cold, so heat losses are reduced. Unfortunately, the fact that the center remained cool meant that the piston could seize up there. There were other mechanical problems, too. Abner Doble declared that it was “unsuited for use in a motor vehicle.” (SCCGB).
It must be emphasized that all of these modifications for increasing efficiency also increase the initial cost, and often also the maintenance costs, and these must be weighed against the fuel cost savings.
The highest efficiency I have seen claimed for a mobile steam power plant is from Cyclone Technologies: “To date the net reproducible cycle efficiency of the Cyclone engine is above 28%, with 31.5% efficiency achieved on the company’s small two-cylinder engine, and 35% confidently predicted to be achieved on the larger 6 cylinder “Mark V” model in the immediate future on the dynamometer.” (Crank). However, the materials requirements (withstanding 3200 psi pressure, boiler temperature 1200oF, exhaust 320oF) are severe, despite use of water as a lubricant. And the engine was developed post-RoF (first patent application filed in 2005).
The energy losses in the boiler include
—unburnt fuel (bottom ash and fly ash)
—incomplete combustion (carbon monoxide)
—(latent) heat wasted by vaporizing water in intake air, in fuel, or formed by combustion of hydrogen in fuel
—(sensible) heat carried off by firebox exhaust gas
—radiation and convection from boiler to surroundings
The greater the combustion rate, the lower the boiler efficiency. Empirically, coal-fired boiler efficiency declines almost linearly as the firing rate increases. For SAR 25NC, boiler efficiency dropped from a bit above 70% at 5,000 kg evaporation/hour to 37% at the grate limit of 32,500. (Wardale 81). The grate limit is the firing rate at which evaporation (steam production) is maximized.
It’s instructive to look at where the heat losses are coming from. Table S4 provides a breakdown, and related information, for a Consolidation (2-8-0) freight locomotive, built 1900, with a 200 psi, single expansion engine without superheat.
Table S4: Effect of Firing Rate on Operating Temperatures, Steam Production, Efficiency and Heat Balance
Consolidation (2-8-0) freight locomotive (1900); boiler pressure 200 psi; single expansion; no superheat.
|Rate of Firinglbs coal/ft2 grate|
Firebox Temperature oF*
Smokebox Temperature oF*
|Water Evaporation Ratelbs/hr|
Equivalent (212oF) Evaporation rate lb/ft2-hr
|Evaporation Ratiolbs steam/lbs coal|
% Heat Loss by CO Formation
. . . by Escaping Combustion Gas
. . . by Unburnt Coal
. . . by Radiation
(Fry and PaRR, data on Series 200, #734). grate area 33.76 ft2; total heating surface, fire side 2541.22 ft2; heating surface/grate area: 75.27; avg. heat value of coal, 14,907 BTU/lb.
*using best fit equation.)
We can see that the decline in efficiency for this freight locomotive is almost entirely attributable to unburnt coal, and we can see the same trend in data for compound expansion and superheated locomotives.
Porta’s gas producer combustion system (GPCS) was intended to burn coal (or other solid fuel) more efficiently. It was, in essence, a primitive coal gasification scheme. Perhaps 30% of the air needed for complete combustion was introduced through the grate, together with a small percentage of the exhaust steam; and the reaction produced carbon monoxide and hydrogen. The reactions of the carbon with the steam are endothermic, serving to reduce the temperature, reducing the likelihood that the bottom ash would fuse into clinkers. And the reduction of the upward draft reduced how much ash was carried upward into the smokestack.
That left the problem of supplying enough air to complete combustion. The remaining (secondary) air was admitted at high velocity through openings at the top (“overfire”) and high on the sides. These were position so that the secondary air mixed with the combustion gases, without disturbing the firebed. (Wardale 78ff).
Because the openings are small, a high efficiency exhaust system (Lempor, Kylpor) is needed to provide sufficient suction without onerous back pressure on the pistons. In a cyclonic GPS, the air and steam are injected in such a way as to cause a swirling action. Porta was able to maintain 78–80% combustion efficiency at maximum output, as opposed to under 50% for a conventional locomotive boiler. (trainweb.org).
As far as I know, GPCS was never used on a working U.S. locomotive (and only sporadically used elsewhere), so the information about it in Grantville is likely to be limited. Hence, even if the steamheads know about it, they are still going to need to work out all the wicked little engineering details in both the GPCS (size and location of airholes, etc.) and the matching exhaust (cross-sections of the mixing chamber, etc.). One source says that “while the GPCS is a simple concept, it requires careful attention to its design and tuning to ensure its proper operation.” (trainweb.org).
Once we deal with the problem of unburnt coal, we might want to try to reduce how much heat is carried off into the smokebox by combustion gases, that being the largest “fixed” heat loss. We can try to increase the heating surface, and engineer the fire tubes to transfer heat more efficiently to the boiler water.
Superheating does reduce the loss of heat in escaping combustion gas. For the “Schenectady No. 3,” working at the normal maximum evaporation rate (12; reached in test 106 at boiler pressure 236 psig and 143oF superheat), 52% of the fuel energy was absorbed in the boiler and another 5% in the superheater, for an overall boiler efficiency of 57%. (Goss 53). On the late steam era SAR 25NC, perhaps one-fifth of the heat used to evaporate water was put to work in the superheater. (Wardale Fig. 21). Bear in mind that because space was limited, putting in a superheater meant reducing the heating surface of the boiler; 22.6% in the case of Schenectady No. 3. (Goss 92).
Stories will ultimately dictate how fast steam engine efficiency improves, but I would suspect that what we have in 1633–1635 for locomotives are simple expansion engines without superheating, and of boiler pressures not exceeding 200 psi. Figure a theoretical cylinder efficiency of up to 14% (non-condensing, for locomotives) or 19% (condensing at 0 psig, for airships). With an efficiency ratio of 70%, auxiliary efficiency 95%, and an overall boiler efficiency of 50%, we are talking about overall cylinder efficiency of 4.7% for locomotives and 6.3% for airships. If we are willing to accept a low firing rate, we might get overall boiler efficiency up to 70% and the airship engine overall efficiency up to 8.8%.
By 1636–1639, I imagine that there will be experimentation with compound expansion, superheating, feedwater heating and, if the steamheads in Grantville know enough about them, GPCS and Lempor exhausts. Once we get the kinks worked out, and I have no idea how long that will take, condensing engines might get up to about 14%. Combining this with high boiler pressures like those in steam cars might take us up as high as 18%.
Power-Weight Ratios for Historical Steam Powerplants
Weight is particularly a problem for steam propulsion; W/P (pounds/hp) for the powerplant on the Giffard airship was at least 100:1. Even the Cyclone Mark V—a supercritical pressure engine that existed only as a developmental prototype at the time of writing and is well beyond post-RoF engineering capabilities—has a W/P of 3.5 (350 pounds dry weight, 100 hp). (Crank).
The Grantville steamheads’ practical experience is likely to be greatest with agricultural steam engines. The first portable steam engines (1849) had a very high W/P ratios; 4 hp was two tons (1000 lbs/hp), and cost $625. (Knapp). A Cooper “Common Farm Engine” (1875), 15 hp, 200 psi, had a weight of 7,000 pounds, for a W/P of almost 467, cost $1,350 in 1881. Note that the weight was primarily the powerplant, as the unit was mounted on cast iron wheels with no frame to speak of, other than the driver seat. (ASME). One comparison of traction engines cited W/P of almost 600 for steam traction, perhaps 225 for steel-wheeled vintage IC, under 200 for rubber-wheeled vintage IC, and under 100 for modern IC. (QuickIH 176). Elsewhere, Quick says that W/P for steam traction was 700–1000, and that IC traction was down to 200 by about 1930. (QuickAT 9).
The W/P ratio is more favorable for a more powerful engine, as boiler weight should be proportional to surface area and power to volume. However, the improvement isn’t great. In 1898, for portable 100 psi boilers of the locomotive type, the boiler weight (without fittings) was 1.35 long tons at 4 hp (W/P 756) and 5.8 tons at 30 hp (433). (Hutton 285). A 25 hp locomotive-type boiler for 1880s oilfield use, mounted on a simple wheeled frame, weighed 6800–7600 pounds (W/P 272–304). (Pees). A Case 150 hp steam tractor (1905–7) weighed 36,000 pounds (W/P = 240). (Leffingwell 33).
Our largest body of data for mobile steam powerplants is for steam locomotives. There are, unfortunately, several problems with using that data as a guide for airship design.
First, the power is expressed relative to the locomotive weight, which of course includes the wheels, transmission and frame, and thus is not strictly comparable to auto engine weight. On the other hand, since a steam-powered airship can’t stop to refill its tender with water every fifty miles like a steam locomotive, it has to have a condenser, and so the weight of the condenser must be charged against it—and that isn’t included in locomotive weight.
Secondly, proponents of steam propulsion argue that there wasn’t a strong incentive to minimize the weight of steam locomotives. A locomotive with too high a power-weight ratio will waste some of that power, because the product of the locomotive weight on the drivers and the coefficient of adhesion (typically about 0.25 for steel wheels on dry steel rail) determines the maximum “traction.” Hence, there isn’t incentive to reduce weight beyond a particular point.
Even so, I found that the ratio of locomotive weight to power was worse for diesel locomotives than for steam. The British Railroad Class 8 Duchess (1937; most powerful steam locomotive built in the UK) had a cylinder power/locomotive weight (P/LocoW) ratio of 39.9 hp/ton. Turning to diesel, the Deltic (1955) had one of 31.1, and the Class 67 (1999) of 33.1. (Blythman). While the above analysis makes steam look good, remember that these steam locomotives had non-condensing powerplants.
What about aircraft use, where a condenser is mandatory? Some preliminary development and design work was done by the US Navy’s Committee on Experimental Power in the 1920s. (Wilson). It concluded that to minimize weight, a flash boiler (no storage space for water or steam) had to be used. Also, the combustion chamber couldn’t be insulated with firebrick; a steel-walled air jacket was used instead. The boiler was intended to operate at 300–500 psi and 800–900oF. In one test, at 325 psig and 772oF, it evaporate 9450 pounds water/hour with 80% efficiency. The Committee estimated that the “finished weight” of the steam generator would be under 2 pounds generator weight/hp. It contemplated that this would drive a turbine that would be 1 pound/hp.
The catch was the condenser weight. A wing-type aircraft radiator weighs about 0.3 pounds/square foot, but a steam power plant requires 11.5 times the cooling area per hp-hr as an internal combustion engine. (Why? The radiator for a IC engine merely must keep the temperature within metallurgical limits. The one for a condensing steam engine must remove enough heat energy from the steam to convert it back to liquid form.) Hence, it would contribute over 3 pound/hp.
A typical aircraft would not in fact have enough wing area and hence there would also need to be a core-type radiator, which would increase drag as well as weight.
The conclusion was that a steam powerplant for aircraft would have 2–3 times the weight as an IC engine of the same power, but be only half as fuel-efficient.
In 1933, Besler flew a steam-powered aircraft for a short flight (10–15 minutes). Reportedly, it had a 2-cylinder, double acting 90o V piston engine with compound expansion, producing 150 hp at 1625 rpm, weighing 180 pounds. The monotube boiler (a Doble Model “F”) operated at 1200 psig, 800oF and the spent steam entering the atmospheric condenser was 4 psig, 215oF. The “auxiliary equipment”(boiler, condensers, feedwater heaters, feed pump) weighed another 485 pounds, for a powerplant weight of 665 pounds and a W/P of 4.43 pounds/hp. (Flying on Steam). A second source (Light Steam Power) says that the engine was only 90 hp, implying a weight-power ratio of 7.38.pounds/hp.
Photographs show that Besler had just a small core-type condenser, and it may be surmised that the Besler steam plane disappeared into oblivion because its condenser was inadequate for sustained flight. (Self, FKP-priorart.htm).
During the 1930s, the Russians had at least eleven R&D projects for aviation steam turbines, and also purchased a Besler steam automobile “as a basis for designing a steam engine for aircraft.” The projects didn’t go well; the designed powerplants were too big and heavy to fit on the available airframes. (Harrison).
Now, putting a steam powerplant into an airship is more enticing. First of all, airship engines typically had higher W/Ps than aircraft engines. Secondly, while an airship doesn’t have wings, it does have a lot of hull surface to accommodate a radiator. I am doubtful that a “hull surface” radiator would work with a non-rigid airship, as its hull is intended only for aerodynamic function, but it could with the keel of a semi-rigid, or even better the full hull of a rigid airship. So, with a semi-rigid or rigid airship, we don’t need to worry about providing a core-type radiator. And 6 pounds/hp for the complete powerplant is tolerable, even if inferior to an IC system.
And then there is another possible twist: if steam is used to provide lift as well as propulsion, then the airship gas envelope serves as a condenser, and the weight of the condenser isn’t “charged” against the powerplant at all. (FKP). We are then talking about a W/P of just 3 pound/hp. (Wilson). Of course, range is limited, as with a hot air airship, as fuel must be burnt to stay aloft, but there may still be a niche.
Weight Reduction Opportunities for Steam Powerplants
In 1830–1870, wrought iron was the standard material for locomotive boilers, typically in a thickness of 5/16″ for 100 psi operation. (White 97ff). Steel boilers were introduced in 1860, but initially brittleness caused by excess carbon was a problem, and even in 1870 they cost more than an equivalent iron boiler. Steel didn’t become dominant until about 1890. Boiler tubes were copper (earliest), brass (from 1851), iron (from 1831) or steel (from 1860s), with iron being predominant in the late 19c. Fireboxes were copper (longer life with wood-burning locomotives), wrought iron (54% the weight and 13% the price of copper equivalent), or eventually steel (five times the life of iron).
Cast iron was the only material used to make cylinders during the nineteenth century. (White 206). Pistons were usually cast iron but occasionally wrought iron or even cast steel. (207–8). By 1913, there were a few cast steel cylinders in locomotive service. (Henderson). However, steel cylinders became an industry standard (Lamb 86), and they are featured on the SAR 25NCs.
There has been modest locomotive experimentation with aluminum, e.g., New York Central’s Niagara (1945): “their boilers were made from nickel steel, their drivers were made of lightweight alloy steel, and aluminum was used on less strategic components to reduce weight.” (Solomon 89).
For pressure vessels like a boiler or steam tube, the principal requirements are high yield strength (permitting high boiler pressure) and a high ratio of fracture toughness to yield strength (to satisfy leak detectably-before-break criteria). (Ashby 160ff). It is very important that these be evaluated at the operating temperature, not room temperature. The tubing is subjected to the same temperature, but the stress is less because of its smaller diameter. The inner firebox is exposed to the higher temperatures of combustion. Note that the thickness of pressure vessels must be proportional to the stress and thus to the pressure difference.
For the inner firebox, and the fire tubes inside the boiler, high thermal conductivity is an advantage. The higher strength of first iron and then steel made it possible to reduce the wall thickness, compensating for the higher thermal conductivity of copper. Of course, for the boiler wall and the steam pipes running from the boiler to the cylinders, you want low thermal conductivity. For all surfaces in contact with steam, corrosion resistance is advantageous. For the fire tubes, which conduct air rich in ash, abrasion resistance is also a plus. Ductility makes tubing easier to flange, and copper had the advantage there.
Aluminum has a high specific strength (yield strength/density)—6061-T6’s is almost four times that of the steel 304SS at room temperature. But the specific strength declines rapidly with temperature, becoming equal to that of the steel at about 475oF. (Burns). This temperature is reached with 134oF superheat at 105 psig or 48oF at 285 psig. While it is possible nowadays to find an aluminum alloy (NASA 398) that can tolerate higher temperatures (750oF), it was developed after the RoF. (Lee).
So aluminum is acceptable from a strength standpoint only for saturated steam (it was used pre-RoF in certain residential boilers) or low superheat applications, and its high thermal conductivity would actually be disadvantageous for a boiler wall, forcing use of additional insulation.
Magnesium is even less suitable for boiler use than aluminum; “the common alloys begin to soften and weaken appreciably on exposure to temperatures as low as 200oF.” (keytometals).
Titanium would be better; the commercially pure grade CP2 has a specific strength exceeding that of 304SS until the temperature reaches about 800oF. (Burns). That said, the preferred material for boilers and steam pipes is carbon steel (coupled with pH control) for low temperatures and pressures, and stainless or other alloy steel otherwise.
For cylinders, see the discussion of weight reduction for internal combustion powerplants in part 2.
As previously noted, a steam powerplant for an airship needs a condenser. For condenser tubes, the principal desiderata would be high thermal conductivity and corrosion resistance (Kutz 185). The traditional material is a copper-based alloy. Here, if available at a reasonable price, aluminum would be quite helpful. Titanium’s higher corrosion resistance compensates for its lower conductivity.
Insofar as radiators are concerned, copper has a thermal conductivity almost twice that of aluminum, but weighs more than three times as much. If the radiator is simply a surface sheet fastened to the keel or hull of the airship, then its structural strength isn’t important. Whether copper or aluminum is used will probably turn on cost.
While the efficiency of a condensing steam powerplant is higher than that of the traditional non-condensing one used in locomotives, and can be improved by superheaters, reheaters, and feedwater heaters, it’s still a bit inferior to that of the gasoline (spark ignition) internal combustion engine, and even more so to the diesel (compression ignition) one.
As noted, the airship steam powerplant designed by the Bureau of Aeronautics weighed in at over 6 pounds/hp. The substitution of aluminum for copper where possible would probably reduce the W/P to perhaps 4, but the “efficiency add-ons” will bump it up a bit.
Steam is certainly a viable option for those airship builders who do not have access to up-time internal combustion engines, or their newly-built counterparts, but there will be a performance “hit.” If fuel is cheap and abundant, the efficiency gap can perhaps be disregarded, and for large airships, the relatively poor power/weight ratio is less of a concern, as required power grows as the square of airship principal dimension, and available power and lift as the cube.
The ability to exploit solid fuel, notably coal, is nonetheless a big advantage for steam. That will be important in countries that are rich in coal and poor in oil, such as England, France or Germany. And the ability to use agricultural waste (e.g., bagasse from sugarcane) would be helpful in India, Africa and South America.
But what if we could propel an airship without any sort of fuel? I will look at exotic propulsion systems in part 4.
To be continued . . .