It's not exactly 'inertia' but for many years I sat through explanations where people used the gas mass in momentum calculations -- complicated by the fact that the gas DOES have inertia and momentum when in motion and WOULD exert force when 'decelerated'. The point I'd nake is that the elastic effect of molecular motion at these densities and temperatures far outweigh the "momentum" contribution directly.
When the valve closes faster than a certain rate, steam will compress slightly (and probably reflect a shock back up the tract or 'ring' around in the chest a bit) but I don't expect inertial force to hold pressure for the length of time back to same-side opposite port admission. One of the things to be tested at high speed on 5550 is the actual gas kinetics at high speed with the eight-valve setup; at least some of those data will be highly applicable to the situation with large-port long-travel piston valves.
(If you have not seen a video explanation of long-lap long-travel valves, find and watch one. Some people think piston-valve travel is very short, as with older slide valve/riding cutoff practice where low travel and minimized mass was advantageous (and easily driven by eccentrics on an uncranked axle)).
The first place I recall seeing momentum effects at a macro level during admission involved comments in the test records of the PRR Q2, a design "infamous" for having been designed with colossal dead space by any published design criterion. It is difficult to imagine an actual mistake of that magnitude, especially on a carefully-considered engine with reduced steam-mass-flow rates into the smaller cylinders of the divided drive. Look critically at the series of surviving test results and think. (Keep in mind that the Q2 developed horsepower and the slightly higher V1 power were about the practical maximum sustainable horsepower for a single-unit locomotive, set by the water rate and practical cistern capability -- about 8000hp at large-locomotive efficiency...)
One of the points of piston valves is that the port volume is arranged circumferentially so any increased pressure is balanced on the rings and any momentary starvation is made up by very high and, ideally, reasonably flow-streamlined acceleration of mass into the passage through the port. What you're looking for would be some recovery of indicated cylinder pressure rather than continuous fall across the period of admission, to be 'as high as possible' at the moment admission stops.
Any pulse effect induced after admission ends will likely not affect potential admission elsewhere while it still has the effect to enhance flow -- which for steam is likely to be slight before the pressure effects move away from the immediate neighborhood of the port by the time subsequent admission starts. Compare this with either pulse ram induction or the principle of the pulsejet.
Poppets if a bit overenthusiastically closed may lead to worse flow issues, especially if the admission tract is as weirdly convoluted as in, say, most of the Franklin type B seems to be (the reasoning as far as I can see is to have all these bent volumes act as a 'steam chest' that recovers smooth pressure by the time the admission valves pop open -- not necessarily achieved in practice according to some accounts I'be seen go by. I was taught very early to use nothing but modified-trapezoid cam profiles to drive steam poppet-valve gear (no jerk!) and the form of the cam can determine admission opening asymmetrically to establish best early mass flow vs. pressure drop. Closing, though, in many Franklin applications is not desmodromic and excessively accelerated to get positive and debounced poppet closure... look for weird pressure effects there and then.
Balance chambers in slide valves are often fed via a different path than the main steam, and this may determine what the effect on the valves is. Anything that adds to the (differential) pressure on the chest side of the valve, even for a brief time, may upset the tribology or increase the resistance of valve motion -- even fretting motion may start to ruin the lapped surface
OvermodBest short-term answer: see the N&W 610 testing on PRR. Verified speed in excess of 110mph...
is this for a passenger train? what about a 100 car freight?
greg - Philadelphia & Reading / Reading
LastspikemikeWere these fast steam locomotives steam flow limited or driver rpm limited?
In the days before Woodard and Eksergian, the idea of 'diameter speed' as a general rule of thumb was thought to apply: the driver diameter correlated with the top service speed of the locomotive. In practice that number increased dramatically within the first few years of lightweight rod 'understanding', to the point that Alfred Bruce (who was in a position to know if anyone was) thought the Milwaukee A design was good for 128mph or better (on 84" drivers without Timken rods) and Kiefer after WWII spun a J3a up to the equivalent of 161mph without damage doing augment testing.
On the other hand, relatively large drivers were not always indications of high speed; Golsdorf's 2-6-4s (which had Krauss-Helmholtz bogies and therefore better guiding than Bissel engines) used the large drivers to reduce machinery speed and lubrication requirements, not obtain higher balanceable rotational speed. Both Mallard and the T1 used drivers no larger than 80" for demonstrable speed well above that. And large drivers combined with short stroke can hurt high-speed acceleration capability... without which you can't reach high speeds you might otherwise be able to sustain. Meanwhile there are critical speeds and resonances which have a very practical speed limiting effect, too.
CPR fastest locomotives as tested were the Jubilee class 4-4-4 (come on Rapido, these just have to be icons of Canadian Steam)...[/quote]These are the moral equivalent of the Milwaukee 'heavy Atlantics' -- we're talking of course about the F2s, not the little wind-up toy F1s which are light Hudsons with a driver pair left out -- which had 80" drivers -- perfectly adequate for speed. I suspect an A could run away from one, but that's more an artifact of being able to oil-fire rather than having greater heating surface, etc.
They looked to be large locomotives also.
Throttle restrictions and steam flow issues matter not if the locomotive can't increase driver rpm anyway.
ICE have rpm limits that can be reached before the otherwise maximum vehicle speed can be reached, for example.
But IC engine limits are at far higher rpm and hence have much higher inertial peak forces than a properly-high-speed-balanced long-wheelbase high-polar-moment American locomotive involves. By analogy with the balancing method Riddles used on the 9F class in Britain (and the Australians tried on some surprisingly unexpected wheel arrangements!) a reciprocating locomotive can receive zero overbalance. The only hammer-blow tendency in such a locomotive is driven by the couple between counterweight and rod, which are within an inch or two of being in a common plane; the absence of overbalance means that the rotating balance of the side rods (and rotating part of the mains) can be 'perfect'. While the resulting wheels have enormous rotational inertia compared to parts of a vehicle IC engine, they also rotate at comparatively low rpm -- AAR recommended highest speed was about 510rpm but this could be practically exceeded in service at times; IIRC Glaze's J balance design (which was NOT even remotely zero-overbalance, although I was taught years ago that it was and that the lateral component of unbalance had to be absorbed in stiffer lateral in the lead and trailing trucks) was good for about 540rpm, and modern practice could well do better ... on smooth level track without the usual and perhaps frequent sort of perturbing effects.
There are people who will affirm that very high drivers are necessary for high steam speed. One justification is keeping machinery speed "low enough" (but they will often quote machinery speed, e.g. peak piston speed, that is conservative by late observed practice, or by what advanced design and tribology can support). A problem with higher drivers is that they have difficulties with starting torque at even relatively high peak piston thrust, and especially when combined with 'lowest possible stroke' as in the T1 (which only got to its stroke by absolutely minimizing the same metal web thickness between axle and main pin, and even then ground the main pin to an eccentric profile so the main rod-eye circle was less than that for the side rod). Doing this can produce the somewhat counterintuitive model of a 6400hp locomotive that cannot start its own train on a grade without assistance ... and was intentionally built without a proper reversible booster to assist with that issue.
gregcat some point, with wide open throttle and significant cutoff, steam chest pressure will be insuficient to go any faster pulling a train on level ground.
Very likely mistakes here are the 'valve problems' observed in some otherwise-fairly-competent-looking designs that ought to have been able to reach very high speeds but most certainly didn't -- the ATSF 3460 class perhaps being the poster child of all poster children, but the C&NW E-4b a very close (and prettier sister!) runner-up. The 3460s were easily capable of speeds right up to the 100mph sustained range, but the performance then fell off a cliff, with great difficulty getting to 102-3, and the peak achieved being about 105. This very clearly suggests ports as being the trouble spot ... possibly the only limiting restriction on getting to T1-like levels of sustainable speed, but in the event a sufficient limiting restriction. The North Western certainly didn't lack for good high-speed designers, but AAR couldn't get those Hudsons more than about 7mph faster than fairly stock PRR K4s(!) which are nowhere near the design an E-4b otherwise is.
LastspikemikeThose cylinder heating steam jackets would reduce pressure loss on transit of steam from steam chest to cylinder.
Wardale in fact makes a fairly strong case for nothing more than good lagging and insulation around that area, that anything added by extra jacketing (and its complications) or tracers doesn't justify the trouble and expense. (He uses steam cooling of the valve liners to solve tribology concerns there due to high inherent superheat, some of which may help wire-drawing condensation concerns as the ports open to steam in gregc's period of interest...)
There are some interesting Google Books treatises on steam-engine jacketing (in smaller road-engine practice, where it becomes more essential and the bulk involved less loading-gage critical) which you might want to read critically and assess.
Pre heating the Newcomen cylinder would have been even more efficient, by exhausting the still hot steam into a cylinder jacket first before condensing it.
Likelier better is to recover heat in the exhaust steam somewhere in the condenser, if possible, and use that in the Rankine cycle somewhere (probably as combustion-air or feedwater heat). Historically though the full 'sinking' of that recovered heat is either uneconomical or saturates the available means long before high sustained engine power via mass flow is demanded.
In a Newcomen engine, one of the early 'improvements' was better lagging/insulation of the outside of the cylinder, and it would probably make sense to put one of the stages of the "condenser" on the outside of this lagging, to 'thermally float' the acting cylinder as much as possible and to cut down heat loss outside the lagging (ideally to negative flow, but that would not likely balance at the inner wall of the cylinder during air exclusion, which in a Newcomen engine is what you're optimizing).
I am a bit chuffed to have understood that enlarging the steam chest to match that of the cylinder to be filled improved efficiency.
The idea of reversing the "low" and "high" pressure cylinders in terms of flow was particularly interesting, in effect superheating twice to lower maximum temperatures but extracting and employing at least the same amount of heat from the fuel.
This is a sort of flow-optimized version of a 'large steaqm-chest volume' approach, and if you look carefully you can see that the effect of superheat present in the admission steam (there is very little by design in Chapelon's arrangement as built) is thought to contribute little to gregc's admission pressure drop of concern.
The exhaust Venturi stacks were a clever application of the same idea utilized for injectors feeding water from the tender to the boiler.
Here it pays to study the Master Mechanic front end and its principles, which are very different from most British boilers. Carefully chart the flow path through one of these things, making careful note where the baffle plates go and what they do -- one of the effective purposes is to constitute a convergent-divergent nozzle arrangement on a grand scale in the combustion-gas flow, incidentally greatly increasing the actual gas-path length ahead of the front tubesheet. You'll see much mention of the mysterious 'lip' that was apparently so sensitive in early empirical front-end tuning; I doubt you will have any trouble winkling out both what its purpose was and why it is such a sensitive thing to tune -- and yes, you should be even more chuffed after realizing those things.
Makes you wonder how far the steam locomotive might have progressed had it not been for the nuclear power plant, which converts steam energy to electricity...hmmm mini nuclear plant powered locomotive anyone? Postwar, France invested heavily in nuclear powerplant generation which made the TGV feasible if not economically sensible.
Incidentally the thermodynamics of nuclear stations are important -- they are wretched even in comparison to traditional locomotive practice, only made practical by lavish but very cheap heat flux. (The same thing that makes nuclear blue hydrogen an attractive prospect ... but I digress.) While it is possible to superheat nuclear steam, it is very difficult to do so safely (the only Russian approach would curl your hair to see, and at the original Indian Point steam station it was projected to do it with separate oil-fired superheat as in marine practice) so all the PWR/BWR designs in the 'golden age' of power construction ran on saturated steam (admittedly at very high throttle pressure) and this is a major contributory reason why nuclear plants "can't" be rebuilt with renewable or conventional steam generation and run cost-effectively. The advanced gas-cooled reactors can have less limitation in this respect, but it remains to be seen if those can make a return to large markets in an age when 'zero-carbon' becomes essential but FUD regarding anything with 'nuclear' associated with it anywhere is still strong and amply advocated...
at some point, with wide open throttle and significant cutoff, steam chest pressure will be insuficient to go any faster pulling a train on level ground.
gregchow close to boiler pressure is steam chest pressure or MEP cylinder pressure at max speed on level ground? can this be quantified if we know the rolling resistance at that speed (see Armstrong's charts)?
can this be quantified if we know the rolling resistance at that speed (see Armstrong's charts)?
i believe yes
would you mind commenting on the difference in steam chest pressure from boiler pressure and the reason for that difference at different rpm?
This is of course germane to the original posted question.
It helps to remember what an indicator diagram represents (and, a little more esoterically, why we use integration to recover information from an indicator diagram like the one modeled in the image gregc provided). In the original indicators, calibrated manometers were used; in more modern versions, fast-acting pressure transducers -- these have some lag and overshoot concerns but of relatively small concern for most work. At any instant the reading from such a device shows the instantaneous pressure; it has been trivial for a long time to implement an 'opposite of a stroboscope' to capture very short-duration readings coordinated with physical excursion or position of parts of the engine (e.g. the valve rod or engine crosshead, or with a little more sophistication the physical or effective steam edge of the valve) and observe pressure deviation over time under different conditions. Only incidentally are these readings 'integrated' by the indicator's charting engine into a 'continuous' trace or line.
It is pretty obvious that without some careful or exotic remediation the 'spot pressure' as admission just begins will drop, as mass in the chest begins to be 'accelerated' into the opening port faster than gas kinetics can support. (In the admission diagram this is reflected by the little pink wedge 'b', although it need not, and probably should not, be progressively lower through the full duration of admission as drawn).
Having more hot steam in a 'properly sized and shaped' chest adjacent to the valve minimizes this drop, but will not eliminate it. Fortunately perhaps the effect on engine performance is relatively slight, because the effective mechanical advantage of piston thrust in this range -- depending of course on the point of 'limited cutoff' less the effect of any slot/Weiss porting, Herdner valves or the like -- will be limited up to the point flow down the port into the passage and cylinder volume 'equalizes' with the flow into the chest from the tract so that EP falls little with admission mass flow as great as needed to fill the swept volume (behind the mechanically-accelerating piston face).
The thrust concern of the drop here is relatively small compared to the mass-transfer consideration. We have not yet looked at the individual pressure trace at the valve in the instants after cutoff (where flow from "the tract" into the port, which was ostensibly flow-smoothed, 'streamlined', assisted by steam conditions in the chest, etc. in passage through the chest, is quickly interrupted in the last moments of crossing the port edges (see the 'trapezoidal port' issues in general for more) but as mentioned the 'practical' effect here is of no particular interest to the engine designer unless any deleterious effects persist until the instant of admission at the other port -- and that time is an eternity compared to the recovery time over admission-flow establishment and then actual admission flow at any reasonably short cutoff. The pressure drop will have recovered and any oscillations or shocks 'ringing' in the chest volume will have either been damped or evenly distributed in steam kinetics by the time of next admission, and any attempt at 'tuning' to achieve enhanced early mass transfer a la cross-ram induction is likely to be difficult across the required range of speed vs. mass flow in this size and cycle time range.
I would be tempted to consider this as if it were a structural loss like nucleate condensation, because many approaches that might reduce it will either impact the cost of the engine or its flexibility in operation -- or its maintenance. Some of the prospective "solutions" introduce the same potential for catastrophic consequence upon derangement or failure posed by various kinds of nonproportionally-linked servo-driven valve gear, a 'design improvement' that armchair thermodynamicists perennially propose as if it were their bold new insight into a better reciprocating locomotive.
Most of the rest of the comments are right, but the concern is not that they 'exist' but what can be done about their effect. The 0.85 factor in the classical formula reflects these much more than putative 'heat loss' from boiler to valve that some people seem to assume.
OvermodThere is a bit, in passing, on steam chests and cylinder filling (and several nested reference lists!) in this little downloadable article: The Development of Modern Steam 1: Andre Chapelon and his Steam Locomotives
The Development of Modern Steam 1: Andre Chapelon and his Steam Locomotives
i believe the indicator diagram in the linked article is misleading to suggest steam at boiler pressure typically enters the cylinder. The diagram below more accurately illustates that steam entering the cylinder from the steam chest is lower than boiler pressure.
as both diagrams illustrate, pressure drops during admission (the Chapelon diagram shows a very significant drop). this drop illustrates the steam chest pressure fluctuations mentioned in this thread.
while the steam chest pressure will be at boiler pressure at very low rpm, i believe the difference is greater at higher speeds as flow needs to match consumption by the cylinders and a greater pressure difference is required (i.e. flow is proportional to pressure difference).
consumption is of course reduced as cutoff is reduced and steam chest pressure may be higher
LastspikemikeThere is a bit of a catalytic effect in that posts I read but don't understand sometimes become comprehensible when I read another post some time later.
Ed almost certainly has that page of Angus Sinclair's where he rips crackpot patent drawings with arrows that show the steam flow ... but nobody told the steam it had to follow, and so in practice it doesn't. Quite a bit of common-knowledge steam "knowledge" is in that sort of category, and more... including quite a bit from Porta, who really knew better... does not work as 'advertised' for similar reasons (the intricately, arcanely calculated Lempor that never seems to work 'as designed' in practice being one that comes to mind).
There is a bit, in passing, on steam chests and cylinder filling (and several nested reference lists!) in this little downloadable article:
http://static1.squarespace.com/static/55e5ef3fe4b0d3b9ddaa5954/t/55e6373fe4b04afd122b821d/1441150783767/%23+DOMS-1_Chapelon.pdf
You may enjoy reading some of the others in that 'white paper' series.
LastspikemikeI just know you have a sense of humour. Be assured that I do also.
At the end of the day, it IS just a sandwich.
LastspikemikeWell, the common V8 would be one cylinder then since all eight holes are bored in one chunk of metal.
Say for argument that we take a couple of 210As out of early Zephyr-style motor trains. These already have the scavenge ports, belt passages, even proportional scavenge-blower capacity to work as desired for an OP compression-ignition engine; it will not change anything in the prospective 'nomenclature' to replace the two blowers with one driven preferentially off one of the cranks (as in Stonehocker's 38 ⅜ engine). Now we need to replace the power-assembly cylinder heads, valves, and central direct injector with ... something. The first significant detail emerges at this point: the engine manifestly retains the separate liners as well as structure of the 'cylinder' part of the 201A (i.e. 'power assemblies') These are frankly obviously 'cylinders' in Winton parlance, and that would and perhaps should carry over to describing the engine thus far.
The FM engine features (and it is repeatedly described historically as a feature) a one-piece liner forming the effective 'cylinder' that the pistons run in and seal against; we do not have that yet in our hypothetical 201A-derived powerplant.
We now have a choice: we can bolt a plate between the two upper ends of the engines (keeping the manifold geometry and attachment sealing a bit more straightforward; this would probably be better for engines cast en bloc which the 201A isn't), or we can make individual units with required fire rings, ports, and injectors for pairs of the power assemblies (comparable to a double-sided 201A head with circumferential instead of poppet valve-in-head exhaust and lateral instead of inline injection) which would be easier to work with.
Technically there is a third alternative, and here is where the way I see the semantics may be clearer: you can make up the equivalent of two opposed power assembly cylinders with the porting and injector bungs, etc. incorporated, perhaps by welding, and then honed as necessary to seal pistons without scuffing as the engine heats in service (analogous to that Stumpf uniflow). Now to the extent this retains separate liners (they would have to be extracted from the bottom) you could still consider this 'two cylinders' in terms of some of the parts and maintenance, but the outer part of such a power assembly, and its function in normal operation, is still one cylinder, with one injection location's worth of pump tuning or whatever, and it runs and could be considered exactly as one long cylinder with two opposed pistons working in it, and could rather easily be phased 12 degrees, etc. just as in the FM engine's single long liners and cylinders.
Glad you liked my 16 44 reference....the 24 66 had 12 cylinders ... coincidence? I think not.
As for cylinder count I was referring to the way the double acting engine functions as a four cylinder would. Four expansion strokes per revolution of the crank.
If there are only two cylinders how come it needs four exhaust ports?
The FM design has only 8 power strokes.....but 16 power inputs to the final drive per crank revolution (two cranks geared together rotating as one crank)
It is not that hard to find either accounts of the design of this engine or the practical implications of some of its design details on the Net ... and you'll have some interesting reading in the meantime.
PM RailfanAbout most all the incorrect info being given here. I'll wait till the dissertation is over, then fill in the blanks with the correct answers, again.
Do keep posting; you might have become confused by some of the strange hypothetical things in this thread, and been judged unfairly.
LastspikemikeMy reference to the FM engine was by analogy. In the case of FM opposed 8 the engine is actually 16 cylinders. Indeed the makers thought so, hence 16 44.
Now if you really need to ease your confusion by thinking of the swept volumes of the trunk pistons as separate 'cylinders' that is fine; just don't go confusing the people who need to know the actual design principle and detail design compared to engines with discrete cylinders or power assemblies. In some other OP engine designs there may in fact be more support for your analogy; you could think of the Deltic as three V engines arranged so all the pistons wind up in mutually-opposed sets, and then the analogy makes better sense if an engine of that configuration were in fact assembled with cast monobloc cylinder-and-crankcase units 'en V' like any good 60-degree V engine, assembled so there are physical joints between them where heads would go on a regular V engine. On the other hand if the cylinders are in one block and bolt to the crankcases below the swept area for assembly (and I think that mechanical triumph the Commer Knocker was made that way?) you'd still be better off with the two-pistons-in-one-cylinder interpretation.
In my opinion you might as well use the analogy that a Stumpf uniflow engine has two cylinders, since the ends of the trunk piston sweep little common wall area. I prefer to treat this as one long cylinder with a long DA piston in it, but if it's easier to understand as two I'm not going to insist.
The point about the pressure excursions in the tract is easily addressed by indicating several points with fast-acting transducers and then examining the resulting data relative to a common (fast-clocked) timebase. I have not heard that the phase of intermittent flow in one branch of the tract materially affects filling during admission in the other, but that does not make the idea 'wrong' and indeed other assumed effects in steam distribution have proven historically unjustified. It might be interesting to look at some of the more extreme Q2 tests, which had some tract and passage pressures measured continuously, as it might be possible to deduce results.
"I'm having a hard time accepting that ...." - Dont worry, I am too. About most all the incorrect info being given here. Ill wait till the dissertation is over, then fill in the blanks with the correct answers, again.
PMR
OvermodSorry to keep harping on this, but it's only the swept volume of the part of the cylinder up to effective cutoff that 'matters' to steam-chest pressure fluctuation.
yes, and that's why i ask about the impact on average steam chest pressure when cutoff is reduced
LastspikemikeIf the volume of each side of the steam chest is equal to or greater than the swept volume of the cylinder it feeds then that should be more than adequate to maintain pressure at say 50% cutoff.
Plus the two cylinders are 90 degrees out of phase. Pressure will be available from either of the two connected steam chests.
From a pressure perspective each cylinder is supplied by both steam chests...If there is any pressure effect at all, it will be reflected pressure or shock. There will pretty obviously be negligible actual mass flow 'from one chest to the other' no matter how black you draw the arrows... Due to the efficient design a two cylinder locomotive is really a four cylinder.
If there is any pressure effect at all, it will be reflected pressure or shock. There will pretty obviously be negligible actual mass flow 'from one chest to the other' no matter how black you draw the arrows...
I'm reminded of the 8 cylinder FM opposed piston diesel which has 16 pistons...
[quote]Point being that each steam chest fills two cylinders to the same cutoff level each revolution of the drivers.[quote]But we take this into account because the period of concern is over a stroke, not revolution of whatever the piston thrust is driving.
Part of why this is important is that there is part of the stroke that y'all are ignoring, and that I've alluded to but perhaps not described explicitly enough to recognize. This is the whole part of the 'first' stroke -- let's arbitrarily say the one going to BDC --after the practical point of exhaust closure (which, remember, is affected both by limited and shortened cutoff with piston valves and most radial gear) back to BDC. There is a residual steam mass remaining in the cylinder at this point, and it stays there to the extent the piston seals properly; it is critical to practical engine operation, in fact, that it is there.
Now take the 'second' stroke, still considering the engine as non-self-starting. Admission to the cylinder ideally comes with the piston still very close to BDC, but the valve has been moving for some time already (the magic of long-travel) so the ports come open, the steam flows in for the admission profile (note in most engines this isn't symmetrical as the piston rod, etc. is present in the back volume but not the front unless tailrods are fitted) up to effective cutoff, and the expansion then occurs between cutoff and release isolated from the chest as we have seen. What is still there, though, is that steam mass from the previous stroke -- it is still there, and it becomes more and more compressed as the piston moves forward. This compresses the steam in the dead space at the forward end, too, and it is not hard to see that at some point the EP of compression will equilibrate to the EP if expansion on the other side. From that point physical work has to be done on the piston to get it through FDC, and there is utterly no guarantee that the peak pressure right around FDC won't rise substantially over the nominal boiler pressure -- in fact it was common for such high compression and its effects to be observed, at least to the degree that Okadee made a product line to mitigate problems from it. Now one of the principles of engine design is to use this compression to decelerate the reciprocating part of the mass; it's been argued roller-bearing rods wouldn't be practical without it and even balanced conjugation doesn't address inertial stress in the mains.
Now consider what happens at the port and the volume in the chest adjacent to it when the steam edge of the valve cracks open to start admission -- with many psi of differential pressure for a moment going the wrong way. That this screws up establishment of smooth admission mass flow for at least a couple of milliseconds goes almost without saying; when the entire practical time for admission at high cyclic and short cutoff is already only measured in milliseconds you can realize the potential horror...
And here the point of Carter's reversible compression control becomes more clear. If the compression pressure is at chest pressure at the moment of admission, the communication of pressure from chest steam to all parts of the piston will be immediate, and the subsequent mass flow (if there is to be no drop in chest pressure) will be exactly the same down the tract as is going through the port for the duration of admission.
Obviously you want more reserve at high pressure close to the valve if there is inadequate effective mass flow for makeup; equally obviously you want as little flow restriction down the tract as possible. And this brings us briefly back to using sliding-pressure firing rather than throttle flow impairment at controlling admission EP/MEP and hence power per stroke...
The two steam chests feed four cylinders of capacity.
At high speed the cutoff will be shorter than when accelerating to speed.
This is where the magic 'low 40s' number comes in -- there will be some point where no more steam mass can be admitted to make better power, and here acceleration falls to zero and the train will be running at maximum achievable speed. But that will be far from shortest cutoff.
In practice the shortest effective cutoff with controlled timing and duration (but no pilot or divided admission or modulated injection) was about 15% for British Caprotti, which is both precise and reasonably fast-acting without spring-return issues. It is notable that the valve arrangement with by far the lowest dead space and lowest port restriction of a reciprocating double-acting design (Bulleid's original sleeve valves as applied to Highland Point and the Leaders) showed no thermodynamic advantage (and in fact Bulleid reverted to piston valves on the follow-up 'turf burner'. Some of the theoretical design work by Doble and the Beslers indicated that very short cutoffs (in the range touted as mechanically achievable with the evolved British Caprotti, about 3 to 5%) will only be of use for substantially higher pressures than are practical at large-locomotive scale or noncondensing configuration.
Lowest steam chest pressure would be at maximum valve opening duration.
If we make the assumption that steam mass flow through the tract is insufficient to sustain steady chest pressure at the ports (as in gregc's example with partial throttling) then progressive droop in admission might be observed, but the droop might not be linear. Remember that piston speed is not linear with stroke progression, and during the entire range of practical cutoff the admission volume will be increasing over time... so any deficit in mass flow will produce lower admission pressure as expected right up to shrouding ahead of cutoff.
I assume this drop would not be significant given the volume behind the push of steam into the cylinders.
One of the classic approaches that 'might' be used is a variant of desuperheat: inject a small mass of overcritical water (for argument's sake, distilled or demineralized water 'floating' at boiler saturation temp and pressure) into the chest where it will not enter the ports. This would flash at 'lower' pressure and then take up some of the superheat (at a time in the stroke it can't be used) to keep EP high at the period through shrouding and cutoff.
The shape of the steam chest castings in the photo at the link provided (UK practice I note) seem patterned for convenience rather than from any knowledge of gas flow.
Sometimes functional streamlining doesn't 'look' particularly elegant (Jaray bring one of the masters of that... and do you remember the 'banana car'?) but still highly efficient. That would certainly be true of the shape of all parts of the exhaust tract through the nozzles and diffusers through the shape of the airfoil at the bottom of the petticoat (Wardale in fact having instructed me on reasonable flow shaping to direct combustion-gas flow from disparate directions around to be coaxial with the steam-jet flow for entrainment...)
~~bingo! so how would steam chest pressure be affected if cutoff were reduced to half? or if only half the volumne of a cylinder needed to be filled from the steam chest/
Here might be the point to take up the other half of the discussion: are there circumstances where interrupted flow produces an effect analogous to pulse tuning (previously alluded to here and a few places elsewhere) and, perhaps, be so tuned as to cause an effective pressure rise in the steam chest during admission that helps 'snap' effective mass flow as admission times become shorter and shorter and transient flow restrictions more and more significant, especially at very high cyclic? That the effects are present in obtaining higher mass flow in IC reciprocating engines is beyond question. I have not seriously looked at this (primarily because there are many ways to get it wrong vs. any that might be effective, and much of the required technological complexity is probably unsustainable in service at any reasonable price, let alone be cost-effective. But do not let that interfere with the fun.
Incidentally, here is the time to put in a plug for the 'full version' of the FDCs as produced by the 5AT organization and (AFAIK) still available from the successor-in-interest. There is probably no better overall reference on practical design of a locomotive, and while it is likely still 'expensive' (Wardale has stated reasons for that) and may need some judicious interpretation for use in a North American context outside 5AT size, you should have -- and pass along when time -- a copy.
There is still not, to my knowledge, a version of Charlie Dockstader's 'valve gear on the computer' that extrapolates the geometry to CFD-derived steam-flow changes as it is adjusted. That does not diminish the goodness (or continued usefulness!) of those programs, which were ported at least through early Windows to XP compatibility (and will still run effectively in emulation).
While you need certain acquired acumen to interpret the results, Bill Hall's programs (including Perform and Perwal) are highly recommended as tools.
OvermodInherent in the explanation is the idea that there be enough volume aa near 'boiler pressure' as possible to get admission filling
at speed, i assume average steam chest pressure is far from boiler pressure
1881
gregc... getting off track
Note that to Wardale it's almost a truism that the 'volume' throughout the steam chest not only have 'recovered' from any transient pressure drop during admission during the subsequent period of cutoff but that it represent a reservoir of high-pressure steam equal to (or as noted 'greater than') the swept volume to be filled during admission. (The peculiar thing is that at least the earlier version of the 5AT collateral implied that in SGS that followed the FDCs the chest volume should equal the full swept volume, which is unnecessary with even reasonable tract-flow streamlining down to the chest volume.
Inherent in the explanation is the idea that there be enough volume aa near 'boiler pressure' as possible to get admission filling -- you will have recognized that this implies prompt and complete mass flow down the tract to 'make up' the volume as it transfers out of the 'reservoir' and that it do so with enough near-instantaneously-communicated pressure to keep the EP from falling at the piston face (that being net of energy extracted via thrust during the admission period, when any 'expansion' in the steam leaves the equivalent of money on the table at the instant of cutoff).
In that context you can see the relative artificiality of discussing a steam-chest pressure that "fluctuates" downward during parts of the stroke. For it to actually drop during admission is an indication that either the throttle is artificially closed or that the tract is defectively restrictive. I don't have my copy of Carpenter's translation of Chapelon's handy, but a great deal of the theory behind 'Andre the Giant's' theory of steam 'streamlining' was actually reducing impediments in mass flow to the valves that accomplishes the necessary pressure-maintaining through admission with no need of recourse to a local 'reservoir'.
If I recall correctly Porta explained a bit about the use of Wagner throttles (he charmingly spells them 'Waggoner' and does not explain what they are, classic LDP) in the patent description of the ACE3000. The idea as stated involved provision of receiver volume very close to the valves (which in a full proper high-speed design would be adjusted via IP injection to full design pressure at least over the commanded range of admission) with little flow-smoothed proportional trim throttles as near the ports as desirable. (A variant of this system is one of the correct answers to high-speed slipping on unconjugated or clutch-conjugated duplexes, but we can take that point up later as it is a digression here.)
The moral, to the extent there is one, is that 'fluctuations in steam-chest pressure during admission' are a bit like the old UNIX International joke that if you see a command-line interface you've failed utterly as a systems designer. To the extent there is any droop in indicated pressure at any point during admission, some of the available power is being lost, and unless full pressure (or at least as high a pressure as possible) is achieved in the cylinder at the moment of cutoff, the 'best' efficiency against somewhat fixed 'exhaust resistance' will be compromised from 'what it could have been'.
While it doesn't help the steam-chest thermodynamics, I'm going to reiterate why net decrease in MEP at moment of cutoff can be valuable: a number of effects both in propensity to slip/spin and to sustain slipping for longer degree of rotation per 'incident' can be lessened with restrictive steam-tract conditions alone. In a slip the rate of volumetric increase in unit time increases, just as it would for the higher cyclic associated (with necessarily higher mass flow against greater 'piston resistance') at high speed, and therefore an 'automatic action' that causes admission MEP to fall at such a moment will assist in recovering the slip. I would gently suggest that one of the lessons taught by a Franklin type A equipped duplex is that perfectly high chest pressures can really become annoying if you can't trim for chronic slip...
Overmodgregc, can you provide a link to the other 3D detail pictures of that arrangement?
https://www.advanced-steam.org/5at/technical-terms/steam-loco-definitions/steam-chest/
... getting off track
PM RailfanYour picture is wrongly termed. "Steam chest extends across twin valves" makes no sense.
Perhaps now would be a good time to note that this is HIGHLY atypical setup, not at all characteristic of a normal piston-valve arrangement. (It took me a moment to recognize I wasn't looking at a Wagner bypass-valve cutaway (see for example the restoration pictures of ATSF 2926) as the 'twin' arrangement is so unusual...)
gregc, can you provide a link to the other 3D detail pictures of that arrangement? I'd like to see whether those will be articulated valves with many thin rings, and how the four of them will be driven.
The "Steam chest" is not a pipe. It is the chunk of cast metal that houses the valve cylinder AND piston cylinder AND contains the porting necessary to operate the cylinders.
"Does the superheater act as a kind of buffer rising and falling in temperature in opposite phase to any pressure effects?" - NO, the temperature of the superheater is controlled by the firebox. And is used to 'super'heat the steam.
I am not sure where that 'controlled by the firebox' business came from. Superheat is a consequence of combustion-gas flow. If you are not aware of superheater dampers ... and why later designs stopped using them ... I recommend that to your attention.
You dont want the superheater temp to be able to fall. If anything the whole idea of a superheater was to rise, and stay as hot as you can get it!
The formula goes: higher the heat = hotter the steam = more work can be done.
I suggest you read up on some of the characteristics of highly superheated steam -- and the practice of desuperheating in power-station operations -- before pontificating that moar superheat is a design principle.
*If you can find any of the discussions of the original enginion AG ultrasupercritical engine design, you can read about a version of extremely high superheat (950 degrees C being the number that first caught my attention) produced in a separately-fired superheater for use in a 'positive-displacement expander' (really just fancy jargon for a piston engine). This design could use high superheat because it metered (as in modern low-fuel-mass direct fuel injection) a very small amount of (again going by memory) circa 7550psi USC steam with pilot-injection techniques, supposedly long-expanding it down to where it could be expediently recompressed after long expansion to go back into the BFP feed train without ginormous heat-exchange condensing. See if the 'ZEE engine' technical paper is still up; that was a somewhat less extreme but still fascinating version of the technology, and (if you believe some of the hype) was slated to become a reliable tech for distributed electricity generation, 5kW per unit...
"just to be clear, the cylinder valves sit in the steam chest ...." - No. There are no cyclinder valves. They are called "valve cylinders". One each side.
Yes, the 'valve cylinder' resides in the 'steam chest', right above the 'piston cylinder'. The 'valve cylinder' is named so, because it is a cylinder, that is used as a [sliding] valve. Not cylinder valves, that is a whole other type of valve.
Your picture is wrongly termed. "Steam chest extends across twin valves" makes no sense. The "Steam chest" is not a pipe. It is the chunk of cast metal that houses the valve cylinder AND piston cylinder AND contains the porting neccessary to operate the cylinders.
As a clothes 'chest' holds drawers, a steam 'chest' holds cylinders. It also holds the "yoke" or "saddle". Thats the part the smokebox rests on, on top of the front of the frame.
"Does the superheater act as a kind of buffer rising and falling in temperature in opposite phase to any pressure effects?" - NO, the temperature of the superheater is controlled by the firebox. And is used to 'super'heat the steam. Its a controller, not a controllee.
just to be clear, the cylinder valves sit in the steam chest or are fed directly from it thru a large opening.
OvermodWHO CARES???
isn't it the starting pressure of the steam entering the cylinder?
LastspikemikeAnd if it cannot recover?
If each time the cylinder is opened to the steam chest the pressure in the steam chest or indeed the whole system drops can it keep dropping?
The superheater still has me puzzled. The hotter the superheated steam gets the more volume it occupies. Pressure doesn't rise. If a low pressure wave travelled from the valve opening back to the superheater tubes then the superheated steam would cool a by a small amount. It should be heated almost immediately expanding as it does so cancelling out the pressure drop caused by the valve opening to the cylinder.
Conversely, when that valve shuts, particularly when due to a shorter cut off, the resulting wave should try to compress the superheated steam resulting in less instantaneous heat absorption. The buffer effect imagined might actually occur. Net result would be a tendency to cancel the pressure oscillations that are being considered as possible.
gregcbut the average pressure is higher
And falling admission leads to lower EP at the moment of cutoff, after which no one cares again what the steam chest does for a while.
of course it fluctuates, it drops during admission and recovers in between. it drops less when cutoff is reduced.
but the average pressure is higher