"Jakuba's Compounding article in the Bulletin"

The article claims that multiple-stage expansion (compounding) has no merit in steam cars. It shows that the disadvantages far outweigh the efficiency improvement usually claimed for those cars. See the SACA Bulletin: July-August, 2014, pg 29.
Stan

Hey, I'm sold on the proposition. I started to get very suspicious when all the old arguments favoring compounding kept quoting output in IHP; things always look better when you ignore friction...unless you manufacture tires. When you start to look at the difference between the peak power needed for acceleration compared to cruise the economy gets questionable. The practical implications of shuttling the steam around induce losses, naturally. The HP cylinder is badly prone to blowby and the LP cylinders size yields high friction and the potential for over expansion. As a guy who works in automotive engine development, I sort of cringe when I think of the cost of manufacturing compounds. You have to build more different kinds of parts but you are going to make fewer of them and thus lose the benefits of economy of scale you see in single expansion design.

Ken

Ken,

Finally someone tells the truth about compounding in steam cars Far better to add one or two more cylinders and shorten the cutoff. Boats and stationary, sure; but not in cars.
Only one that is even more ridiculous, the three and four stage expansion with uncontrolled reheat. A subject that eventually defeated Abner and he finally admitted it in a letter to Warren.

Jim

I have to agree when you are looking at a normal compound engine. And as Jim says: "Boats and stationary" It would just perturb the problem of over expansion. Your just draging the lower stages around most of the time.

We do need high expansion to get efficiency and good MPG. But we also need an enormous power range in an automobile. A simplified look at power requirements to speed says that power requirements is proportional to the speed cubed. A 3 to 1 speed range requires a (3*3*3) 27 to 1 power range. At 60 you need 27 times more power then at 20. A 5 to 1 range requires (5*5*5) 125 to 1 power range. At 75 you need 125 times the power then required at 15. That is a requirement of basic aerodynamics. Force required to over come wind resistance is proportional to the speed squared.

Force = constant*Speed^2

Distance = Speed*Time

Work=Force*Distance.

Power = Work/Time
= Force*Distance/Time
= (constant*Speed^2) * (Speed*Time) / Time
= constant*Speed^3

Compound engines have been proven to be of advantage. And st high expansion ratios have been shown to have closer to their ideal cycle efficiency. If fact they have better efficiency then a single expansion engine operating between the same working pressures. The lower temperature drop in a compound stage causes less thermal losses then the high temperature difference of a single stage.

All that said, over expansion is a big efficiency killer. Over expansion is expanding the steam below the external exhaust pressure. The force comes from the pressure acting on the piston. But there are both positive and negative forces involved. On the power stroke we have a positive force while the on exhaust-compression we have a negative force. When we over expand we have some part of the power stroke doing negative work. That is when the internal cylinder pressure is below the pressure acting on back side of the piston. Negative work of compression in the cylinder is not a problem as we get that back on expansion. But negative work to exhaust is a loss. Basicly a pumping loss.

This is not just a problem with compound engines. But is perturbed by the fact that at low power the low pressure stages may be a load on the power. Any throttled engine could exhibit over expansion at low power. but a compound is worse.

When I first heard about the Williams engine I developed a program that could calculate a compression cycle. It used the ASME steam property formulations. A compression cycle is complicated to calculate. In figuring the expansion one needs to know the steam state at the start of expansion. But that is a mixture if the residual steam in the clearance space and the admitted steam. That presents a problem as we do not know the residual steam state when we start. The residual steam state may be compressed or not depending on exhaust closure. We also need to know the exhaust steam state which is dependent on the end of expansion state. This presents a case where the initial state is dependent on the cycle processes. The text book Rankine cycle does not address such a problem. It does not even include clearance in calculating an engine cycle. The text book Rankine cycle was developed before we had automated calculating devices. A real engine has clearance. The text book Rankine cycle is simplified for ease of calculating with pencil and paper. Today we have vary accurate steam property formulation that can be implemented on hand held calculators. Even on smart phones. I have a HP 48 calculator on my android phone and IAPWS steam formulations are available for it. Programmable devices have little problem calculating more inclusive Rankine cycle that includes the mix of clearance volume steam and admitted steam.

Theoretically (based on the text book Rankine cycle) the most efficient cycle is a full expansion cycle. But in the real world we have to consider other things such as flow. It takes a pressure difference to flow steam into or out of the engine. There are restriction and direction changes that also play a part.

In a full expansion cycle there is no room for throttling with out going into over expansion. But how much does one have to compromise to avoid over expansion to get the needed power range required for an automobile to day. We can go to extremely high pressure. But such a high pressure needed for a high expansion and a throttled power range requirement isn't practical.

When you look at all this the usual piston steam engine is not a good fit for the huge power range required for an automobile. The IC engine also has the same problem. That is the reason that electrics have an advantage. Electric motors can be designed for a wide range of power output at a fairly constant efficiency.

In my analysis of the Williams engine I discovered the great advantage that compression has. Basicly you get most of the power it takes to compress the steam as you get on re-expansion. If power can be controlled by varying the amount residual steam looping in the cycle we get a fairly constant efficiency over a range of power. Power can be controlled by the amount of steam being compressed. We have some amount of steam at cutoff. If 98% of that steam is recycled steam we have only 2% being admitted steam. On the other hand if we have 2% recycled steam and 98% admitted steam. That is a 48:1 admitted steam range. That is a lot more range then one could get from throttling at a high expansion range. But how do we keep a constant expansion range while varying the steam rate as described. Well that has a simple answer. We have to vary the clearance. But it's a complicated implementation. We would have to vary the clearance, cutoff, and exhaust close in unison over the power range. That is what I have been working on for years now. I havn't found an engine design that I really like. The best is to be able to vary the stroke. The best would be to do so about the center of stoke. Now that gives a two fold power change. The power is being controlled by the amount of steam going through the engine. By varying the stroke we are varying both the displacement and clearance. As we reduce the stroke we increase the amount of recycled steam in the clearance space. This engine would be operating at full pressure at the inlet. And the expansion is always to the same ending pressure.

This engine is complicated to control and would need a micro-processor to make it work.

What does this have to do with a compound engine. It has to do with the control of the inlet and exhaust valve timing. The electromagnetically latched valves I think would be the best considering the ability to control them electronicly. At very low speed we would still need over lapping cutoff for smooth running. We need a sort of throttled mode to take off and at slow speed. I figure that the electromagnetic-valves could be used as a throttle. By opening and closing them quickly the cylinder pressure could be controlled. But the least amount of time it takes the electromagnetic-valves to traverse open-close limits the expansion ratio as the RPM increases. A three stage expansion is the minim for an over all 30 to 1 expansion ratio of equal power from each stage at 2600 RPM. But even the throttled mode of operation all stages are independently throttled. There would be inter stage receivers maintained at a minim set pressure. So all stages are always producing their part of the power. There would be no dead stages as a simple throttled engine would have.

This compound has come from a theoretical design that operates at nearly full expansion over more then a 125 :1 power range. Initial inlet pressure at 1580 PSIA with inter-stage drops of 15 PSI. The final stage expansion is to 20 PSIA. Note. This is a counter flow full compression engine with electromagnetic inlet and exhaust valves controled to maintain nearly full compression and expansion in each stage.

The reason for going to a compound is that valves have a minim of 4 ms open – close time.

I think that if one throttled each stage and used inter-stage receivers of sufficient volume a compound could be more efficient due to the lower heat losses they do exhibit when used correctly. From the modeling I have done on my compound design it is possible have inter-stage receiver pressures that produce nearly equal power from each stage.

Andy

There nothing like a nice compound running with 200 degrees superheat and 29 inches of vacuum at a constant load hour after hour.
Rolly

Hi

Andy, you often have mentioned having equal power output per stage in a compound engine.

Do you think it is really that important for engine balance?

I'm still for the compound guys.

Best,

Bill G.

Bill,

Are you asking about the dynamic balance or the torque output per cylinder balance?
Dynamic depends on the mechanical balancing and what counterweighting was used.

Torque balance, only if you can vary the cutoff independently between the HP and LP cylinders.
Only one steam car had this, the 1903 Lamplough-Albany. The rest went bumpity-bump at times.

The preferred poppet valve engines present a very difficult problem is you wanted to vary the cutoff between cylinders, very complicated valve gear. Not impossible; but probably not worth the trouble.
On the road with both types of engines, you really don't care.

Jim.

Andy

I noticed that the calculations for the power modulation range, has been made on the basis that the net resisting force on the vehicle is contributed totally by aerodynamic forces. The combined effects of road and aerodynamic resistance represented by a simple power law negates the cubic associated with aerodynamic resistance acting alone. While the power distribution to the aerodynamics is a cubic function, power allocated to the road is linear with speed. Thus, the net power represented by a simple power law requires that the exponent be less than three (3).

The exponent can be found by curve fitting a simple power law to a third degree binomial (power distribution between road and aerodynamic resistance). For a typical 3000 pound family type vehicle, my work indicates that the exponent representing speeds between 60 and 70 mph is 1.77. Thus, the power modulation range between 60 and 20 mph becomes seven (60/20)^1.77.

The effect of representing power modulation range by a simple power law also impacts feed water rate representation between two arbitrary speeds, if the feed water ratio is represented by a simple power law. For example, if the maximum steam generating capacity is 1200 lb/hr which corresponds to a maximum vehicle speed of 120 mph, then at 60 mph the required steam generating rate becomes 198 lb/hr (1200 x
(60/120)^2.6).

Actually, these issues are more complex than normally perceived. In many cases the KISS philosophy is sometimes applied ( in the interest of a quick answer) to represent power and feed water rate by a simple cubic power law. Such an approach normally aggravates actual values. In
your work. have you considered the distribution of power between road and aerodynamic resistance? jerry

Bill
You also want to balance as close as you can the load on the rod journals. By adjusting the cutoff on the high-pressure cylinder you can change the pressure on the low-pressure cylinder.
On my big compound the high pressure was 275 psi with a 3.5 inch diameter piston, that’s 9.62 sq inches or 2645.5 lb on the rod journal. The low was a 7-inch piston or 38.48 sq inches. By adjusting the cutoff on the high I was able to get 50 psi on the low and with 29 inches of vacuum or say another 14 negative lb-per sq inch for a total of 64 lb per sq inches that’s 2462.7 lb on the rod journal. It makes a difference on how smooth the engine runs.
Rolly

Hi,

Jim, dynamic balance with the thing, of course.

I was thinking that a difference of power output between the 2 stages wouldn't be too noticeable though. But, then the old Buick running on six out of eight cylinders did jump around a bit.

I probably have enough cylinders for this to not be a problem. There is no way or reason for me to think of varying cutoff between stages as there is little to no receiver.

Best,

Bill G.

Hi Bill,

The old Buick 6 was dynamically balanced pretty well...dynamic balance referring to the forces caused by the centrifugal forces of the engine rotating parts and reaction caused by the reciprocating. Two cylinders at 90 degrees sharing a common crankpin produce a rotating primary unbalance force, which is readily canceled by the crank counterweighting. That describes the Buick crank and thus it had no primary unbalance. It did have secondary unbalance which manifested as a horizontal rocking couple at a frequency of twice engine rpm. That rocking couple was largely self cancelling, however as it was about 1/3 of what a comparable inline 4 cylinder engine would generate since the secondary forces in that configuration are all mutually reinforcing. A balance shaft would have helped, but honestly, the engine was well enough balanced that it wouldn't have made a huge difference.

The problem with that engine was that it had 3 VEE pairs at 90 degrees and 3 crankpins set at 120 degrees. The spark plug firing was something like BAM....BAM.........BAM...BAM.........BAM...BAM........BAM...BAM.... The power delivery wasn't evenly spaced out and it caused a pulse; why Harley riders love this sort of thing is beyond me but last I heard they quit balancing those engines because the riders LIKED the engine to shake more. Obviously car buyers have more refined sensibilities. The solution to all this was the "split pin" crank. Each crankpin was split in two and splayed apart so that they overlapped to form a continuous crank, the splay produced an even firing order. The balance was actually probably not as good as the old Buick, but it wasn't utterly horrible and a balance shaft or two could remedy this further.

Regards,

Ken

Bill
If were still talking compound engines. If it’s a double acting compound with journals at 90 degrees you need 3.5 to 4 times high-pressure cylinder volume for the receiver between cylinders.
If it a 180 degree engine which most weren’t they would be 170 or 190 degree then the volume is immaterial..
If they were single action 90 degree then it would be two times HP volume.
Rolly

Bill,

The difference in torque output is most definitely noticeable, unless at one throttle opening, load and cutoff, everything is more or less in harmony.

I simply cannot conceive of why you would run a Buick Straight Eight on six cylinders. A straight eight crankshaft has the crankshaft with a 2-4-2 arrangement , while a straight six is 3-3 with the back 3 opposed to the front three.
I had a 41 Buick Century Sedanette and it was smooth as silk at 70 in second gear and good for over 100 in high The XKE, Chevy Stovebolt, two Mercedes, Maserati 300 SI, were all straight sixes and equally smooth.

I think Ken is referring to that Buick (also a GM truck engine I think) with the 90° vee six and I think a three throw crankshaft in the very first ones. GM tried to build the engine on a V-8 production line and someone really goofed.
My boss at Lockheed had one of the first ones, a little teardrop Buick, and it was about as smooth as a paint can vibrator. Two periods one at twice the rpm shook hell out of the car. Horrible engine.
I think GM quickly replaced the engine with one with a split journal, I know my boss got rid of his in two months. Modern V-6 engine are all 60° with split pin cranks and are smooth.

Jim

Jim Crank Wrote:
-------------------------------------------------------
> Bill,
>

> I think Ken is referring to that Buick (also a GM
> truck engine I think) with the 90° vee six and I
> think a three throw crankshaft in the very first
> ones. GM tried to build the engine on a V-8
> production line and someone really goofed.
> My boss at Lockheed had one of the first ones, a
> little teardrop Buick, and it was about as smooth
> as a paint can vibrator. Two periods one at twice
> the rpm shook hell out of the car. Horrible
> engine.

Yea, I drove an early Skylark with that engine, accelerating off a stoplight was quite an eye-opener.

Buick eventually went all the way to the opposite extreme, the crankpins were split at plus and minus 15 degrees and they had to perform miracles on the crank to make it reliable, rolling the edges at the split and the throws.

The 4.3 V-6 Chevy OTOH had quite a different approach, they really wanted to use small-block V-8 parts and production capabilities. As I recall they produced several versions of the motor with different crankpin offsets and installed them in vehicles for consumer evaluation. The result was that they chose a 10-degree offset and reduced the expense and complexity as a result.

My personal experience with the 4.3 is substantial, I owned a Silverado stepside with one and really enjoyed it. Further, I attempted a conversion for an aircraft project which utilized some stock components and a mix of SB racing parts. I eventually sold it and the buyer put it in a ski boat and estimated its horsepower as approaching 300 BHP.

Interesting to note, the block was a refinement of the small-block Chevy and featured updated thin-cast techniques; I don't recall the block weight but it was remarkably light.

I hope "frustrated" will chime in here, I'm recalling most of this information from aging memory banks and I'll bet Ken can provide corrections and additions.

Bill



Edited 1 time(s). Last edit at 08/11/2014 02:45PM by Bill Hinote.

I noticed that the calculations for the power modulation range, has been made on the basis that the net resisting force on the vehicle is contributed totally by aerodynamic forces. The combined effects of road and aerodynamic resistance represented by a simple power law negates the cubic associated with aerodynamic resistance acting alone. While the power distribution to the aerodynamics is a cubic function, power allocated to the road is linear with speed. Thus, the net power represented by a simple power law requires that the exponent be less than three (3). 

The exponent can be found by curve fitting a simple power law to a third degree binomial (power distribution between road and aerodynamic resistance). For a typical 3000 pound family type vehicle, my work indicates that the exponent representing speeds between 60 and 70 mph is 1.77. Thus, the power modulation range between 60 and 20 mph becomes seven (60/20)^1.77. 

The effect of representing power modulation range by a simple power law also impacts feed water rate representation between two arbitrary speeds, if the feed water ratio is represented by a simple power law. For example, if the maximum steam generating capacity is 1200 lb/hr which corresponds to a maximum vehicle speed of 120 mph, then at 60 mph the required steam generating rate becomes 198 lb/hr (1200 x 
(60/120)^2.6). 

Actually, these issues are more complex than normally perceived. In many cases the KISS philosophy is sometimes applied ( in the interest of a quick answer) to represent power and feed water rate by a simple cubic power law. Such an approach normally aggravates actual values. In 
your work. have you considered the distribution of power between road and aerodynamic resistance? jerry

Hi Jerry

The aerodynamics low applies to simple shapes. An automobile body can be designed such that it exhibits less resistance at some higher speed. That is relative to the cube law. The coefficient of aerodynamic wind resistance varies with speed. Some cars have active aerodynamics and would exhibit more aerodynamic resistance when producing increased down force. Rolling resistance decreases with speed initially but stabilizes as centrifugal force overcomes tire deformation. So yes there are other factors involved. A friend of mine did track his fuel usage and found that at 50 MPH he used less then 1/2 the fuel as at 65 MPH. (65/50)^3 = 2.197 That is in agreement with the cube law.

The cube law alone is only part of the equation. But my intent was a simple explanation of why a very wide power range is required of an automobile today. With simplified cubic speed to power model and some build experience we can simply adjust the parameters to get what we wont. I.E. If we are trying to get a 20 to 75 speed range. Experience might tell us we need to design for a 20 to 95 range to get 20 to 75. I wouldn't really care if we got more range.

I am a bit confused that you found the exponent to be 1.77. That does not seam right. The rolling and aerodynamic, in fact all resistances to movement, should be additive. The fact that one decreases with speed should not result in an overall value that is lower then any single contributing factor. If A and B are positive values how can the sum of A and B be less then B? Or A?

Bill asked:

Andy, you often have mentioned having equal power output per stage in a compound engine.

Do you think it is really that important for engine balance?

To me balanced power and a balanced engine are not the same. I go along with Jim on that. But at low speed I wont as soothe a torque profile as I can get. A lumpy running engine a low speed would not be good. If you have several pistons per stage then power balance would not be so important. The main problem woud be if at low power the low pressure stages are not being utilized. That would just drag down efficiency. That is the reasion for having the interstage recievers and throttling each stage. The Recievers would be maintained at some pressure that works over the entire power range. All pistons at all times are contrubiting to the output. No dead cylanders being draged around at low power.

Analyzing the compound design I described over it's power range. That is calculating thousands of cycles adjusting the inter-stage pressures for the closest to equal power out of each stage. Seams to come close with fixed interstage pressures. But using MathCad you can not analyze time dependent processes. Flow through valves etc.

Still working on my VisSim plugine to do that.

Andy

Jim's right, I misread the comment about the 6 cylinders. Actually, however, the point is still valid. Dynamic balance is all about inertial forces and nothing else. An eight cylinder engine, whether straight or vee, is inherently balanced. Well, the V-8 is sort of a special case, the unbalances resolve out into a rotating couple which the endmost counterweights cancel...if anyone wondered why they are often bigger than the others. The balance is unaffected no matter how many cylinders you use, it is the steadiness of the power delivery that will cause vibration. Engines that shut down half the cylinders to save fuel are pretty common, and they all run very smoothly. In fact, there has been talk of using systems that will run with any number of cylinders (and experimental models have been tested), they are still very smooth because multiport fuel injection allows the computer to vary fuel delivery to each cylinder in such a way as to smooth out the torque variation.

There have been at least two different GM split pin 6 cylinder arrangements, one a 60 degree bank and the other a 90 and both run very nicely; just shows that if you do the math right and add a balance shaft or two that you can generally come up with a viable option. I know that if I were building an inline 4 cylinder steam car I wouldn't use a 4 cylinder IC crank if I could afford to have one custom built. The IC crankpin layout is 0-180-180-0, which is balanced for primary forces but causes all the secondary forces to move in the same direction at the same time. This layout works nicely because the 4 stroke cylinders fire every other revolution although modern engines are adding two balance shafts running at twice engine rpm to take out that secondary shake. In a steamer you get 2 cylinders "firing" at the same time, nothing necessarily wrong with that other than that you could smooth out the power delivery and reduce peak stress by putting the pins at 90 degree angles.

As I understand it, the Williams brothers used a 0-180-90-270 layout (basically half an inline 8) as it gave the least unbalance; which is not to say it is smooth. I would go with a V-8 style 0-90-270-180 which has a bit more unbalance in the form of a rocking couple. My reason is that this layout fully cancels the secondary unbalance leaving just the primary while the Williams method leaves you with both forces superimposed on one another...making for a "lumpy" output. Also, a single balance shaft turning at engine speed can easily convert this large fore and aft couple into a much smaller side by side couple while two counter-rotating shafts could totally cancel out all the unbalance....The Williams design would take 3 balance shafts to reach the same level as the V-8 layout with 1 shaft and they would need 4 to do what 2 would accomplish with a V-8 crank.

All this relates back to Stan's argument against compounding. Assume an inline quad compound. The reciprocating mass on each piston is going to be different unless we add a lot of weight to some of the pistons, which is counter productive. We can counterweight each crankpin to the rotating mass plus half the reciprocating for each associated cylinder and reduce the unbalance as much as possible but the different inertial forces arrayed along the crank are going to create some kind of unbalance. We can handle the unbalance with balance shafts as with my V-8 crank example, above. The problem gets even trickier, however, because the different piston diameters cause the crankpins to be unevenly arrayed along the crank; now you not only have different forces for each cylinder but you also have different leverage (in two directions). The problem gets annoying when we consider actual engineering. The quad 4 crank is going to be longer and "whippier" and we are adding unmatched forces along the length; this leads to the crank bending which results in crank fatigue and premature bearing failure. This can be addressed by the same means long used with inline 6 cylinder engines, make the crank heavier and beef up the block lower end. Of course, modern automobile design is trending in the opposite direction for reasons of cost, operating economy, handling and so on.

As the old DI told me in boot camp: "Ain't nothing never easy."

As for Bill's comments on the 4.3 V6, they sound pretty much as I recall the information.

Regards,

Ken

Regarding the discussion on friction and power requirements. You could take a car of comparable size and configuration, calculate the weight with driver and fuel, take it onto a stretch of abandoned level road, run it up to a given speed then pop it into neutral and measure how long it takes do decelerate to a lower speed (say 10 mph less) with a stop watch. Calculating the kinetic energy at the two speeds and knowing the amount of time it takes to transition from one energy state to the other should give you the combined road and aero resistance.

We seem to discuss a lot the causes of vibration associate with multiple expansion engines. Observing the interest, I should like to point to the exhaustive treatment of this topic in the SACA Bulletin in 2005. Four issues were dedicated to the balancing topic, culminating in the Oct.- Nov 2005 (Vol 19, No 6) issue.
Stan

Hi Stan,

I'm gratified to see that someone read that article all the way through. I posted that article on the Phorum a few years ago when the topic came up. Anyone wanting to read it can go to:

[steamautomobile.com]

Regards,

Ken



Edited 1 time(s). Last edit at 08/11/2014 04:12PM by frustrated.

Alright,
Ken's contribution certainly takes care of the dynamic balance problems and choices for cylinder arrangement. Incidentally the Napier Lion W-12 of WW-1 was a smooth and very powerful engine, as is the new one Bentley, VW and others now use, plus a W-16 version too.

Balancing torque output in an automotive compound is not so easy unless much added complication is really warranted. Then you need to decide if absolute balance is really worth it. Fourteen years driving the OO White proved it was just not critical.
Then consider that very often you try to start from rest with the HP cylinder on either TDC or BDC so a simpling valve must be included, unless you want to stick your foot out the door and give the car a push.
Dobles went to great length to have an automatic simpling valve in the Series D cars and never got one right. So a separate foot pedal is seen in the photos of the four D cars.
Hence the interest in a radial inflow turbine for the second stage, giving a 20-1 expansion ratio if wanted.
Jim

Andy

Of course the format here is not friendly toward equations. Also, readers do not get excited about equations. So, I will try to
sum up my mathematical take on engine power ratio real quick.

Two equations are required: one for state point one (1), the other for state point two (2). Combining and rearranging, an equation
results which represents the horsepower ratio between (1) and (2) as a function of speed ratio(S). This equation has the form:
KS + CS^3 = S^m, where (K) and (C) are constants less than unity. Specific values of (K) and (C) are such that (m) must be less
than 3 in order that the equation hold. jerry

Jim --

When you mention a radial inflow turbine as a LP expander, how do you see the total expansion being divided between the engine and turbine? Would there be limitations on the dryness of the steam coming out of the engine?

Just to make sure I understand what you are describing, is the turbine you are looking at similar to the turbine in a standard turbocharger?

Thanks,

Tim

Most steam turbines are not happy with liquid water in the steam.

Condensing steam locomotives had lots of trouble with cinders cutting the exhaust fan blades.

There was at least one condensing steam locomotive that also had problems with the exhaust fan drive turbine shelling out unpredictably. They finally realized that this was due to water in the steam. They substituted a drive turbine that was water friendly. In fact, I believe the drive turbine was pretty much a pelton water wheel used with steam instead of water.

Kerry

You could handle wet exhaust in a few ways.

First, use a separator, this could be a simple chevron/cyclone rig.
Second, perform a mild reheat. Since the exhaust is going to do work in the turbine, the penalty isn't as big as it would sound.
Third, don't expand the steam quite so much in the cylinder and leave enough superheat to ensure dry steam into the turbo.

Ken

kerry Wrote:

> There was at least one condensing steam locomotive
> that also had problems with the exhaust fan drive
> turbine shelling out unpredictably. They finally
> realized that this was due to water in the steam.
> They substituted a drive turbine that was water
> friendly. In fact, I believe the drive turbine
> was pretty much a pelton water wheel used with
> steam instead of water.

A little off-subject here:

I live in Tehachapi, California--home of the "Loop", one of the civil engineering marvels of the late 19th century. Google for it, you should see the massive train lengths which overlap each other on their way "over the hill".

More amazingly, there currently is an average of 36 freight train passages through this pass each day--and the railroad is being upgraded to handle an increase to something like 56 trains per day within the next year or so.

For me, the most fascinating part is the performance of the diesel-electric locomotives which haul these loads up incredible gradients. It's just so hard to imagine how 3 or 4 locomotives in front and one or two in the rear, can handle the tonnage they're managing along with the road conditions.

Sometimes they smoke a little--but when you consider that a 100-car train is eliminating the same number of diesel trucks pulling the mounted trailers and containers ( as well as eliminating all those drivers, the whole train is managed by just 2 people!). This is a technology which we don't even bother to acknowledge because it's so common and well-accepted.

Right now, I don't see any hope for steam to compete with the current locomotive technologies as they exist.

FWIW.

Bill



Edited 1 time(s). Last edit at 08/11/2014 09:14PM by Bill Hinote.

Ken

You are right on all points, but turbine failure occurs in the blink of an eye. A turbine can be just fine one instant and scrap metal the next. In fact, shutting down a turbine with a problem is often the kiss of death. You probably have one or more critical speeds to pass through on the coast down. Better to keep the turbine online until you recover control and then perform an orderly shutdown. Turbines I have experience with work economically only with the blades at transonic speed. Better to design a machine that is tolerant of conditions than to live on the ragged edge.

Kerry

Hi Kerry,

I've worked with turbomachinery from about 50 HP up to 70,000 HP, and you are absolutely correct about critical speeds and rapid failure. On the other hand, I occasionally downshift my car to up the revs and get into a better torque band while I floor it---sometimes to beat a light and usually when I pull onto the freeway. The sounds of the turbocharger winding up and down is distinctive over the piston engine but it takes the abuse and doesn't skip a beat. The power turbine section of that unit is also a radial inflow turbine and in general about what you would use as a second stage expander for a steam engine, although probably with a higher expansion ratio....the cross section would likely resemble IC automotive turbocharger compressor sections more than the power turbine end. In fact, it does tend to resemble a Pelton wheel more than an axial flow tubine, the admission is perpendicular to the shaft rather than parallel.

Ken

Tim-Ken,

The reasoning is this. Having lived with several compound steam car engines (White, Doble, Besler) I don't like the complication, bulk, torque roughness at varying cutoffs at slower speeds, and being restricted to how the cylinders are arranged to give good dynamic balance. Torque impulse balance is still one vexing problem. (I must confess, I am still very attracted to the Wankel as the main expander for a ton of reasons now that Mazda has apparently developed oil free seals. I would much rather go that way than have to develop some compound.)
Since the desired expansion ratio between cylinders is frequently stated as 4-1, the turbine can do a lot more for sure, are small and balanced and if designed with variable inlet nozzle vanes, common practice now with turbos, is more efficient by a long shot over a wide rpm range.

First of all, the turbine would be a radial inflow with the expansion ratio as large as dyno testing would allow, the blade shape determines this.
With modern turbocharger wheels, the compressor side has a better ratio because of the need for high boost pressures at the speeds they will get to in service and a high flow rate. The turbine wheel only has to supply enough power to run the compression section. So for fan turbine or draft booster use, the compressor side of a turbo is better.
Inflow turbines can go as high as 20-1 and probably higher. Barber-Nichols has a great web site with a good download on this type of turbine. Naturally our SACA web site will not allow it to be attached, too big. Their chart of who works best and where is attached. Download their web site and read the radial turbine part.

As it was explained to me be three real turbine experts, you always want the exhaust from the piston engine to be above atmospheric, seems whether slightly superheated or just dry doesn't matter and definitely exhausting to about a 24" vacuum. Stretch that usable PV diagram out as far as possible
No turbine likes water drops, but axial flow impulse suffer fatal damage and errosian far easier than radial types. So the cutoff would be limited to always insure that the steam does not condense into water in the turbine.

My feeling is to leave the beautiful open compounds and triples to steam launches where hey are so delightful to watch and stationary constant speed uses. Vehicles present a much greater challenge.

Jim

My feeling is to leave the beautiful open compounds and triples to steam launches where hey are so delightful to watch and stationary constant speed uses. Vehicles present a much greater challenge.
Jim

Right on the money Jim
I built four compounds and designed a few that never got built, and all were used in boats. That’s where they belong.
Rolly

[home.comcast.net]

A wise man (Churchwald, I think) once said words to the effect that anyone can get steam into a cylinder, the trouble is in getting it out. High expansion ratios or expanding to low pressure in a cylinder exacerbate this problem.

I often wondered if the Erie and Virginian triplex locomotives might have been more successful if they had a Pelton wheel expander on the exhaust of the rear LP cylinders. This could have balanced, or overbalanced the back pressure caused by the exhaust nozzle on the front LP cylinders. The Pelton wheel might have been used to drive centrifugal water pumps or centrifugal air compressor for braking, to supply over fire air, perhaps to drive the stoker engine. Then again, the Pelton wheel might have been excessively large and given the rightful conservatism of railroad engineering departments, the idea is probably a nonstarter.

Someone (on another forum), sometime ago promoted the use of a turbine for second stage expansion of steam discharged from the cylinders a conventional steam locomotive. The turbine was proposed to drive an electrical generator/alternator to drive traction motors on the pilot, trailing and/or tender axles.

A while back, I proposed (to no response) expanding across the phase boundary to recover some latent heat and perhaps condensate.

To my mind, the draft inducing front end of a conventional steam locomotive and the draft booster of some Doble cars qualify as compound engines. I actually like the idea of a compound locomotive or automobile engine. What I don’t like is for both stages having to drive the same load and thus operate at the same speed trying to maintain some equality of expansion.

Ken- when I tried to open your link I got an error message. Am I doing something wrong?

Kerry

All files from this thread

File Name	File Size		Posted by	Date
Expander and turbine_chart copy.pdf	676.4 KB	open | download	Jim Crank	08/12/2014	Read message
Engine_Balance_2003.pdf	623.6 KB	open | download	frustrated	08/12/2014	Read message
ENG0016.jpg	69.2 KB	open | download	Rolly	08/18/2014	Read message
ENG0015.jpg	125.5 KB	open | download	Rolly	08/18/2014	Read message
Scroll generator..jpg	78.5 KB	open | download	Rolly	08/18/2014	Read message
RollyEpnVS.JPG	47.9 KB	open | download	steamerandy	08/31/2014	Read message
wobv2.mpg	533.8 KB	open | download	steamerandy	08/31/2014	Read message
1-2 comp 6.jpg	796 KB	open | download	sidrug	09/14/2014	Read message
BillPlots.JPG	194.9 KB	open | download	steamerandy	09/19/2014	Read message
Doble E Valve and sleeve Model (1).pdf	81.4 KB	open | download	Jim Crank	10/02/2014	Read message
Stanley oil pump.jpg	86.7 KB	open | download	Rolly	10/05/2014	Read message