Railfan.net Home Railfan Photos ABPR Archives Staff Safari Photos Railfan Links

Railfan.net Forums Railfan.net Forums Railfan.net Forums
Welcome, Guest. Please Sign In or Register. Aug 17th, 2017, 11:17pm
Categories •  FastIndex •  LongIndex •  Help •  Search •  Members  •  Sign In •  Register


New Build Steam
   Railfan.net Web Forums
   Locomotives and Rolling Stock
   Steam Locomotives
(Moderators: Henry, CHESSIEMIKE)
   New Build Steam
« Previous topic | Next topic »
Pages: 1 2 3  ReplyReply     EMail TopicEMail Topic   PrintPrint
   Author  Topic: New Build Steam  (Read 1973 times)
Overmod
TRAINing
View Profile  

Posts: 12
Re: New Build Steam
 
« Reply #40 on: May 16th, 2015, 9:50am »
Quick-Jump   Reply w/Quote   Modify

Let's bump this topic.
 
Alfred Bruce said in the early '50s that the Milwaukee class A was easily good for 128 mph+.  It would not be terribly difficult to modify the design with some form of poppet valve to give better admission and exhaust, and to provide some form of reversible compression control (see Jay Carter's work over on the SACA 'phorum').  The difficulty here is that the locomotive is not truly 'cost-effective' in terms of the size of train it would have to pull to 'earn its keep'.  
 
One of the principal reasons for doing a T1 is that the locomotive is large enough to pull right-sized consists effectively, but has the advantages of divided-drive.  There are multiple effective ways to control whatever percentage of the slipping problem that turns out to be endemic to the duplex-drive concept and not just a consequence of the 'wrong touch' on the throttle and reverser.  If the T1 Trust locomotive is built with outside Type B1 valve drive (as seems likely) it will be possible -- with contemporary late-'40s technology -- to install fast-acting gear conjugation between the two engines.  Of course there are 'traction-control' alternatives too.
 
Be advised that the T1 Trust has four places that have already signed agreements-in-principle to run the locomotive when it is built, and given the amount of careful design attention to reducing various forms of track shock and damage, there should be little if any problem in securing more.  There are also arrangements to develop equipment for 'ferrying' the locomotive effectively in general freight service between places it is intended to be demonstrated or run, and to support it effectively when it is in service.
 
The practical top speed of an ATSF 3460 class is unlikely to be much higher than around 105 mph with a train.  Even a cursory examination of the valves and passages, and the design of the front end, will demonstrate that.  I believe it is easy to substantiate that the latter classes of 4-8-4 (from 3765 to the end, including the 2900s) were considerably faster in a number of important respects than the big 4-6-4s.  Now of course much of this 'problem' can be addressed with different ports and passages (I advocated paired piston valves, similar to David Wardale's approach) and a better arrangement in the smokebox.  With some care the 'surgery' involved can be made reversible (so the locomotive can, as promised, be restored to original visual condition at the end of the CSR/SRI Project 130 effort).  On the other hand, I don't think it's particularly likely that the 3460 chassis will be stable at speeds above the low 120s, and that there will be at least one significant resonant couple before the engine reaches the 'target' speed.  Addressing that is going to involve very great changes in lead and trailing truck arrangements, particularly lateral compliance, even before considerations of overbalance come in.
 
I am surprised that two candidates have not been mentioned, even though one was referenced in passing without further comment.  Those would be the original 86"-drivered Buchanan 4-4-0 999 and the "1905" version of PRR 7002.  In my opinion, someone like David Kloke could build the former locomotive starting tomorrow, and this is one of the few 'small' locomotives that has genuine long-term sustainable interest without having to rely on large numbers of excursion passengers per year.  I for one would love to get my hands on 7002 to re-learn all the secret lore about how to fire and run that locomotive at high speed ... although I can't for the life of me figure out any physical way for it to run anywhere near 127 mph, let alone with a four-car train.
 
I am frankly surprised that one of the J1e projects hasn't come to fruition yet; as noted, a big part of the job resides at Steamtown, and there is continued interest that a trailing truck (for patterns and perhaps some restoration) is somewhere in the Gulf Curve area still.  At least one of these (the one to be built in China) foundered over the issue of 'follow-on' locomotive production; the Chinese were going to be happy to provide J1 replicas at something like 1/10 the price required to build the first one.  What I'd like to see is an organization structured similarly to the T1 Trust, that would do the groundwork for obtaining the necessary drawings (from NYCSHS, which I believe has a reasonably full set), building the 3D model, etc.  The T1 Trust is developing a database of materials and practices that will make much of the development work for a subsequent J1 build much simpler, less expensive, and more assured; meanwhile, over on RyPN, a manual for best steam practices (both for restoration and operations) is in the works.  
 
There isn't much question 'which' NYC Hudson to build: either 5344 (by far the most famous single Hudson) or a "5345" follow-on.
 
Of course, I'm going to put in a plug for the locomotive I'd most like to see restored: the NYC Mohawk in Elkhart.  Here too, developing a coherent organization and a good feasibility plan and volunteer base early on is going to be 'key'.  Elkhart is likely not much different from Nashville in refusing to consider a working restoration until it's demonstrated that hard assets and good planning are in place to assure that, no matter what might happen, the steam locomotive winds up back on display in one piece and cosmetically good condition...


Logged
Henry
Moderator
Historian
Posts: 6107
Re: New Build Steam
 
« Reply #41 on: May 16th, 2015, 12:19pm »
Quick-Jump   Reply w/Quote   Modify

I'm all for a J1e, but I would be perfectly happy with a T1
 
It would be awesome to have a new T1 hit 127 MPH!
 
Being a Pennsy fan and N scale modeler, I have an N scale T1. I need to paint it one of these days.
 
Henry


Logged
Overmod
TRAINing
View Profile  

Posts: 12
Re: New Build Steam
 
« Reply #42 on: May 16th, 2015, 4:00pm »
Quick-Jump   Reply w/Quote   Modify

It's not 'either/or'
 
The T1 is coming no matter what happens.  At some point the 'rising tide' of knowledge will begin to make any effort to build a large replica locomotive easier -- materials science, useful contacts in ereas like CAD advice and foundry practice, methodological best practices and model documents, operating approaches.
 
I don't see the T1 and J1e financing and trip efforts interfering or 'cannibalizing' each other, either.  They are two very different appeals to two very different aspects of fandom, and they have distinct commercial and 'sponsor' prospects.  The T1 is wicked cool, the J1e is a household name and one of the most famous and beautifully proportioned locomotives in the world.  
 
I'm looking forward, too, to a world in which other locomotive projects can receive support and have their path simplified...


« Last Edit: May 16th, 2015, 4:01pm by Overmod » Logged
ClydeDET
Historian
View Profile  

Posts: 4793
Re: New Build Steam
 
« Reply #43 on: May 16th, 2015, 4:28pm »
Quick-Jump   Reply w/Quote   Modify

I think it was Bill Withun who proposed "fixing" both the T-1 and Q-1/Q-2 slippery problems by applying internal coupling (crank axles and light-weight roller-bearing rods) to prevent the two driver sets from getting out of phase, while retaining the advantages of the four-cylinder divided drive. I'm not enough of a mechanical engineer to pick holes in the idea.
 
I expect using poppet valves in any locomotive intended to regularly run really fast would be  good idea. I know the ATSF rebuilt at least a couple (a Pacific and a 4-8-4 with poppet valves, light-weight rods, all roller bearings, etc) post-war, but despite really outstanding improvements in power and fuel economy, not proceeded further with as decision for all diesel-electric had been taken.


Logged
Overmod
TRAINing
View Profile  

Posts: 12
Re: New Build Steam
 
« Reply #44 on: May 16th, 2015, 6:06pm »
Quick-Jump   Reply w/Quote   Modify

Withuhn conjugation (as seen in early '70s Trains Magazine and incorporated into the ACE 3000) is a clever idea, but it has some functional drawbacks (best appreciated if you get hold of a copy of the patent and look at the relevant drawings of the arrangement there!)  You have a double-cranked axle and the conjugating rods on the mains, which are already taking the crank forces from the outside pistons -- that makes for very high unsprung mass on those four drivers.  The arrangement also involves lubrication difficulties on those inside connections; roller bearings would have to be split to be practically applied, and so would a floating bronze bushing as on the later UP FEFs.  While true quarter and throw of the inside cranks is not *that* critical (they are only used for keeping the engines conjugated and only 'see' enough torque to do so) any play or wear is likely to become significant.
 
I looked at Riley Deem's proposed method of conjugating the Q2 class, which involved the use of gears rather than rods.  In order to preserve the benefits of divided drive, the two engines need to be able to take up any phase angle between them, but lock to each other with minimum slip or latency if one of them begins to slip.  Deem if I recall correctly understood the need for a viscous clutch in the conjugation system; I propose using a magnetorheological clutch instead (which happens to be an actual contemporary device; Rabinow's research starting in the latter half of the Forties and well-reported in the industry by 1948, at the time PRR was looking at advanced solutions for things like valve breakage.)  There is room in the transom (as illustrated on the 1947 engine bed and spring rigging drawings) for both the conjugating shaft/bearings and the clutch; while the rear type A cambox is suqarely in the way (tipped up on end!) I don't think anyone would seriously propose replicating either the inside valve gear with derived drive or the 'nightmare box' on a recreated engine.  Remember that the type B-1 Franklin gear differs from the design applied to ATSF 3752 (and covered in the "type C" long compression patents) - it is specifically designed for outside shaft drive but with actuation of eight valves per cylinder (two admission and two exhaust per end).  We're fortunate that the type D installation 'kit' used on USATC 611 has survived, which uses a lightweight version of outside drive ... and is interestingly close to the cylinder dimensions and ratings of one T1 engine...
 
My own use of geared conjugation adds a detail: the shaft is 'detented' so that the two engines preferentially run in 45 degree phase to each other, giving eight power pulses per driver revolution and smoothing torque peaks for starting.  I also use active traction control -- ideally through lateral floating air-over-hydraulic calipers acting on cheek plate surfaces on the rims of the driver centers, although the 5550 is likely to implement this through active elements in the independent brake rigging.  I also expect to adapt the four-Wagner-throttle arrangement hinted at for the ACE 3000, in order to perform some adjustable 'trim' on the forward engine (it is very difficult if not impossible to modify the T1 to have independent front-end throttles for the two engines mounted in the 'normal' position at the superheater header).  
 
Poppet valves are of tremendous importance on a high-speed locomotive, as the results of the Lima K4 installations demonstrate.  (Yes, I have considered the difference between the 'early' and 'late' versions, with the better internal streamlining...)  The problem is that poppets can actually be deleterious to best operating economy if slower speeds -- and on a T1 this means speeds below 85-90 mph! -- are not exceeded in service.  The T1a conversion had very comparable performance (with much less maintenance complexity!) up to this range, but the horsepower very rapidly rolled off above that point and even with large valves was unlikely to reach higher speeds.  One alternative, as noted, is to use two 'modern' piston valves side-by-side, and perhaps operate them so that one is optimized only for inlet and cutoff and the other for exhaust.  But this is NOT an appropriate solution for 5550 (even though a saving grace for 3463)
 
Note that even on a divided-drive, it makes some sense to use Voyce Glaze's balancing approaches - almost zero overbalance (with stiff lateral compliance from lead and trailing trucks) - there is just enough 'overbalance' to account for the vertical component of piston thrust in the main drivers.  That allows the engine to reach upward of 540 rpm without excessive augment or hammer-blow; while there is additional lateral mass at a lever arm represented by the outside valve drive (crank, gearbox, and part of the 'drive arm') it can be fully dynamically balanced.  
 
Nonetheless, I expect there will be critical resonances in the drive, the guiding, and the suspension, which I expect to be significant above about 122 mph and which may progress to destructive amplitude in a comparatively short time.  This is one reason why very fast reciprocating steam may Not Be A Good Idea, as the same sort of shock that leads to high-speed slipping of a duplex-drive engine might lead to runaway vertical or lateral oscillation if encountered at a higher speed.
 
I think the 1948 engineering studies solved the problems with valve breakage that became amplified when the spring return pressure was increased to give more positive seating with no 'bounce'.  Centrifugal casting and selective chamical and heat treatment are likely to be two approaches used.  The difficulty is that, since the T1s were no longer first-line power from 1948 on, most of the fancy valve improvements were never actually adopted, and much of the specific detail design and approach has been lost -- fortunately we can re-create it with more modern materials science adding to the possibilities.
 
One great theoretical advantage to Lentz-style poppet valves is that they can permit much higher superheat. which in turn can improve the locomotive's water rate.  I am of the opinion (without actually running the numbers) that the kind of seat used on the LP valve of a Skinner Unaflow (which levitates the actual seat on a thin film of steam) represents a useful method of providing a tight but cushioned seal at high cyclic engine RPM.
 


Logged
Henry
Moderator
Historian
Posts: 6107
Re: New Build Steam
 
« Reply #45 on: May 17th, 2015, 2:02am »
Quick-Jump   Reply w/Quote   Modify

You have covered a lot of ground in that post. Lots to think about.  
 
I don't know enough about divided drive conjugation to discuss it intelligently at this point. Using gears with some sort of clutching makes sense to me if the drives maintain a fixed distance and alignment with each other.
 
Too bad you couldn't cut out the reciprocating action and cut in a turbine at higher speeds
 
I would think that computer modeling might help with valve gear design and refinement. At least it could reduce some wasted time.
 
What about desmodromic poppet valves? Adjustment might be a nuisance I suppose.
 
I am not familiar with the LP type of valve of a Skinner Unaflow.
 
It's been a while since I thought much about any of this and I'm certainly no engineer. My practical experience with steam was in rebuilding and testing safety valves of all sizes many years ago.
 
Henry


Logged
Overmod
TRAINing
View Profile  

Posts: 12
Re: New Build Steam
 
« Reply #46 on: May 17th, 2015, 9:52am »
Quick-Jump   Reply w/Quote   Modify

Desmo poppet valves are something of a problem on large steam locomotives because of that typical bugbear, differential expansion.  
 
A more likely approach is to use hydraulics to approximate the desmodromic action (Meier -Mattern gear, cf. patent 1763474, being a close thing to the desired action).  Problem is that you start needing a considerable volume of oil displacement at high cyclic rpm to get this to work, and the consequence of either excessive or inadequate steam distribution if there is any failure at high speed  can be dramatic.  This is the same general rule regarding the systems that would use rotary encoding and some form of servo-driven gear instead of a mechanical cam to drive either poppet or 'drop' valves.  There are a number of 'showstopping' failure modes in even the PLC-controlled versions of those systems, and I am uncomfortable thinking of them in 'normal' service that is expected to run at sustained high speeds.
 
I have been watching the various experiments in actuator and injector design to see how reliable a servo-driven system with large inertial mass can be made -- the 'electronic camshaft' systems for large diesel engines being one attractive area of research.  In order to have proper steam distribution at very high cyclic rpm a full-oroportional variable valve gear would be essential; I would propose to implement this with variable followers in an otherwise-mechanical cam-driven design (essentially similar to the method used for type B/"C" Franklin gear)
 
There are a number of hybrid turbine proposals, probably the most attractive of which was Russell Brown's 'asynchronous compound'.  That design was able to use HP pistons with no physical contact with their bores (!), the excess steam being used in a multistage LP turbine driving alternators (and ancillary gear) that was capable of producing excess HP via admission of throttled IP steam.  Meanwhile, the 'other' great PRR experimental turbine program, the original mechanical-drive V1, was going to be tested with one of the great secret weapons of the late '40s, the Bowes drive.
 
The Bowes drive is a bit like a combination generator and traction motor contained in a concentric coupling.  If you have a turbine locomotive, you can operate the turbine at its optimal flow rate virtually independent of locomotive road speed, this getting around the (fatal as it turned out) issues with high slip and flow rate at starting on 6200.
 
The V1 was originally sized to make 'full available horsepower' out of a modified Q2 boiler, which you'll recall was good for 7600-odd horsepower out of the divided drive chassis.  The original mechanical direct drive was intended to be rated at 8000 hp, but some bright promotional genius changed this to 9000 hp, which I suspect could have been achieved with the 'usual suspects' thermodynamic improvements (Cunningham circulators, Snyder combustion-air preheaters, Franco-Crosti economization, and the like) ... but something was being forgotten, the same 'something' that posed trouble for Chapelon's prospective Big Boy if it were actually supposed to be run at its new 'rated horsepower' (which, alas! American railroads would try doing).  Even the 8000 hp V1 had a hellacious water rate, its working range even with the largest PRR 'coast-to-coast' tender being about 130 miles.  You could enhance this a bit with A-tanks, and PRR had access to the N&W experience with how to use these effectively, but you can imagine the working range of a 9000-hp variant.  Achievable water quality from track pans would almost certainly require EXTENSIVE on-locomotive treatment to be turbine-compatible, to the [passenger' version (so famously styled by Loewy in '47 or so) would not be much of an improvement even if it could run 145 mph plus with relative safety.  
 
Now, much of the PRR's apparent infatuation with high speed appears to have stemmed from the late-'20s improvement programs -- the 9000' tunnel under Horse Shoe, the 'new main line' regradings, etc.), just as its design of high-horsepower steam (the Q2 and V1 in particular) were optimized for wartime traffic operations.  The combination of the Depression and the rise of competing modes made superspeed trains uneconomical, and F7s/E7s turned out perfectly capable of extended 'fast enough' running to match what PRR actually required from non-wartime power.  Note that we didn't even see 428-A equipped electric locomotives to Pittsburgh through the new tunnel because of that.
 
On the other hand, there is little doubt that the future of any 'steam speed record' belongs to non-reciprocating locomotive designs, 'unsexy' as that prospect may be.  It's unlikely that any turbine-electric design would be effective in regular service using '40s DC TMs and generators; the N&W TE-1 design foundered almost completely on problems with its main generators and traction motors.  (If you can cook a set of 12 hexapoles in regular service, while only producing 4500-odd horsepower at low speed, there is something impossibly wrong with your engineering and design!)  Meanwhile, the Bowes drive was an integral part of the Ingalls Shipbuilding 2000-hp passenger locomotive -- compare it to contemporary efforts in Europe with lightweight hydrokinetic drive -- and there is little doubt in my mind that the necessary improvements in unsprung mass and truck suspension to make a V1-like turbine capable of true high speed would not be difficult.  For operation at higher speed ranges with restricted load, it would be relatively simple to put in a multispeed transmission, probably running at high speed between the turbine output shaft and the Bowes drive, to improve the 'fit' between steam demand and output hp at speed.  The lower chassis height also promises to reduce frontal area, which is becoming a major determinant of achievable top speed above around 115 to 120 mph.
 
Just for the record: the very-high-speed approach is to use a steam turbine to drive one of the UT-developed 'MegaGen' alternators developed for the ALPS locomotive, with the steam derived from combined-cycle recovery from the ;rincipal gas-turbine and augmented by one of the catalytic direct-cycle methods (which produce steam at appropriate pressure and superheat directly by reaction of hydrogen peroxide with a low-carbon fuel).  All the TGV improvements for high speed translate across reasonably to this approach, and there's a considerable amount of 'costed-down' OTS equipment (now, in fact, becoming functionally obsolete for HSR service) that can be acquired for the purpose.
 
To save a little time for those who are interested:  the Bowes company archives have been saved and catalogued at the Independence Seaport Museum in Philadelphia.  There is a fairly good annotated finding aid at the PACSCL page describing the collection:
 
http://clir.pacscl.org/2010/08/23/thomas-d-bowes-m-e-associates-records/
 
I recommend making a 'first pass' through this, noting the significant patent numbers, and then downloading .pdf copies of those patents through Google Patents.  (I''d post some of them here, but I'm on a different computer from the one with my saved copies.)


« Last Edit: May 17th, 2015, 9:55am by Overmod » Logged
Henry
Moderator
Historian
Posts: 6107
Re: New Build Steam
  PRR_Loewy_V1.jpg - 165912 Bytes
« Reply #47 on: May 17th, 2015, 11:16am »
Quick-Jump   Reply w/Quote   Modify

Once again, a lot to digest.
 
The V1 would have been an interesting starting point. I'd like to see what a decade of development could have done to it.
 
I imagine it would have looked something like this image from the June 1945 Popular Mechanics.
 
I'll have to research all this a bit more when I have a little time.
 
Henry


http://Forums.Railfan.net/Images/SteamGeneral/PRR_Loewy_V1.jpg
Click Image to Resize

Logged
ClydeDET
Historian
View Profile  

Posts: 4793
Re: New Build Steam
 
« Reply #48 on: May 17th, 2015, 4:49pm »
Quick-Jump   Reply w/Quote   Modify

Getting past this former lawyer's engineering knowledge, I fear. But interesting. Or will be, I think, once i digest it a bit.
 
Given that steam handling is a problem in reciprocating engines, i wonder if anybody has modeled sleeve-valves? Or used them - in steam, I mean. Worked well in gasoline engines (including several automobiles - knight and Minerva to my knowledge - and aviation engines - Napier Sabre and several Bristol engines including Taurus and Centaurus). Disadvantages, too. High oil consumption, added expense from the higher precision required.


Logged
Overmod
TRAINing
View Profile  

Posts: 12
Re: New Build Steam
 
« Reply #49 on: May 17th, 2015, 6:07pm »
Quick-Jump   Reply w/Quote   Modify

The 'Triplex 'in the PS article is the V1 design as 'modified' to 9000 hp.  There was a bit of argument between Loewy and Carleton Steins regarding use of the term 'Triplex' (which Loewy originated to refer to a locomotive that had its coal, water, and engine in three separate 'units').  By the time the V1 design was mature, it had nothing much beside the name in common ... there is some correspondence about this in the collection at the Hagley, as there is about the 'secret' Baldwin project to build the turbine-electric that became the C&O M-1 (the upper drawing in the PS article).  
 
Meanwhile, there was a 1947 Westinghouse report that shows what a 4-8-4 version of a direct-drive side-rod turbine would look like.  I have no doubt that such a locomotive, or one with individual axle-motor drive (like the B&O W-1 or the Henschel 19 1001) would have 'succumbed' to dieselization just as any other new big steam did -- but it might have been interesting to watch.  Here, I think, the design would have benefited from a symmetrical Ljungstrom turbine design, with balanced rotors either side of the center pinion, and admission inboard and exhaust outboard (and perhaps some use of bleed steam for feedwater heating and resuperheat).
 
The thing I would look at is how the running gear could have been lightened, and the suspension compliance improved, on a turbine-mechanical locomotive.  Many of the techniques proposed for the 'rebuild' of GG1s in the mid-'70s, such as welded underframes and truck frames and the use of chevron Fabreeka springs for primary suspension, would probably be applicable.


Logged
Overmod
TRAINing
View Profile  

Posts: 12
Re: New Build Steam
 
« Reply #50 on: May 17th, 2015, 6:46pm »
Quick-Jump   Reply w/Quote   Modify

Quote:
Given that steam handling is a problem in reciprocating engines, i wonder if anybody has modeled sleeve-valves? Or used them - in steam, I mean. Worked well in gasoline engines (including several automobiles - knight and Minerva to my knowledge - and aviation engines - Napier Sabre and several Bristol engines including Taurus and Centaurus). Disadvantages, too. High oil consumption, added expense from the higher precision required.

 
One of the best steam designers tried them -- Bulleid, on Hartland Point and then on Leader.  The design used a Meehanite sleeve running in a conventional cylinder, with the piston rings bearing directly on the inner bore of the sleeve.  The dead space was essentially limited to the 1" or so port length imposed by the thickness of the sleeve; in addition, there could be very prompt port opening (since the admission and exhaust porting was in a 'ring' around the periphery of the cylinder, and the sleeve could be rotated through an angle as it was moved fore and aft to help with lubrication and prevent 'stiction').  The interesting thing in tests was that very little actual thermodynamic improvement was observed from this.
 
Naturally the design included a great many rings, and a great number of lube passages, and any one of the rings breaking would pass considerable steam.  It is not exactly clear whether the problems with differential expansion of the cylinder, sleeve, and piston were properly resolved before the plug was pulled on the Leader development project; there is little doubt, however, that a sticking valve on any of the three cylinders would provide an immediate and unhappy stall, without much chance of 'breaking loose' or doing a conventional blocking of the valve to permit the motion to be taken down and the locomotive moved.
 
The valve 'oscillation' was notable for being driven independently of the valve timing, from a sprocket and chain arrangement from the leading driver axle.  The combination of longitudinal and twisting motion was said to be 'fascinating to watch'.  
 
A great weak point was that the sleeve had to be moved.  It was provided with 'ears' that projected through slots in the cylinder head.  I do not know if normal metallic packing can be made to seal the resulting joints as well as a conventional valve rod at the center of the cylinder head can be.  I think it says a great deal about the idea that when Bulleid took up the bogie locomotive again with the Irish 'turf burner', he used conventional piston valves...
 
I would suggest that the sleeve-valve system be used with uniflow cylinders, in particular with the 'auxiliary valves' that benefit flow and compression control at the exhaust.  These sleeves might have to be provided in sections for application, and moved with some form of derived motion, but it might be possible to run them with balanced pressure and thereby greatly improve the tribology and some of the differential expansion problems.
 


Logged
ClydeDET
Historian
View Profile  

Posts: 4793
Re: New Build Steam
 
« Reply #51 on: May 17th, 2015, 9:30pm »
Quick-Jump   Reply w/Quote   Modify

on May 17th, 2015, 6:46pm, Overmod wrote:       (Click here for original message)

 
One of the best steam designers tried them -- Bulleid, on Hartland Point and then on Leader.  The design used a Meehanite sleeve running in a conventional cylinder, with the piston rings bearing directly on the inner bore of the sleeve.  The dead space was essentially limited to the 1" or so port length imposed by the thickness of the sleeve; in addition, there could be very prompt port opening (since the admission and exhaust porting was in a 'ring' around the periphery of the cylinder, and the sleeve could be rotated through an angle as it was moved fore and aft to help with lubrication and prevent 'stiction').  The interesting thing in tests was that very little actual thermodynamic improvement was observed from this.
 
Naturally the design included a great many rings, and a great number of lube passages, and any one of the rings breaking would pass considerable steam.  It is not exactly clear whether the problems with differential expansion of the cylinder, sleeve, and piston were properly resolved before the plug was pulled on the Leader development project; there is little doubt, however, that a sticking valve on any of the three cylinders would provide an immediate and unhappy stall, without much chance of 'breaking loose' or doing a conventional blocking of the valve to permit the motion to be taken down and the locomotive moved.
 
The valve 'oscillation' was notable for being driven independently of the valve timing, from a sprocket and chain arrangement from the leading driver axle.  The combination of longitudinal and twisting motion was said to be 'fascinating to watch'.  
 
A great weak point was that the sleeve had to be moved.  It was provided with 'ears' that projected through slots in the cylinder head.  I do not know if normal metallic packing can be made to seal the resulting joints as well as a conventional valve rod at the center of the cylinder head can be.  I think it says a great deal about the idea that when Bulleid took up the bogie locomotive again with the Irish 'turf burner', he used conventional piston valves...
 
I would suggest that the sleeve-valve system be used with uniflow cylinders, in particular with the 'auxiliary valves' that benefit flow and compression control at the exhaust.  These sleeves might have to be provided in sections for application, and moved with some form of derived motion, but it might be possible to run them with balanced pressure and thereby greatly improve the tribology and some of the differential expansion problems.
 
  

 
Huh - I have a biography of Bulleid, and it has quite a bit on Leader. Had forgotten all about it using sleeve valves. Bulleid was quite an interesting fella.


Logged
Henry
Moderator
Historian
Posts: 6107
Re: New Build Steam
 
« Reply #52 on: May 18th, 2015, 1:19am »
Quick-Jump   Reply w/Quote   Modify

What about air operated poppet valves? It might be easier than hydraulic fluid to move quickly.
 
Henry


Logged
Overmod
TRAINing
View Profile  

Posts: 12
Re: New Build Steam
 
« Reply #53 on: May 18th, 2015, 7:53am »
Quick-Jump   Reply w/Quote   Modify

There is, of course, a great deal of existing work in precise, proportional air-operated valves, going back to the Apollo program in the '60s, and to Formula One racing engines.
 
It is also true that even with the stringent 'control air'-like purity requirements for the supply, and the relatively higher pressure required for clean operation at high acceleration/high peak speed required for locomotive-size operation, it is relatively easy to assure a supply of air on a locomotive, because of the brake requirements.  (This includes the use of relatively high-pressure air for the 'caliper' independent brake)
 
The principal difficulty is that, unlike a hydraulic gear (which is at least partially 'hydrostatic') an air-operated gear would have to be 'full servo', with no controlled valve motion without full pressure and modulation.  If this were to be operated desmodromically, without mechanical valve return springs, the requirements increase.  And it is this characteristic I think is most significant.
 
If there is any failure of the power air, the valves will either close (if spring-return) or float to an incontrolled position or possibly oscillate against their stops if impelled by steam pressure.  The result of this will be admmission of steaqm-chest pressure to the cylinder and piston with relatively little warning, probably when the parts of the engine have very high momentum.  I have my doubts that a modulated compression-control system can accommodate that much steam mass.
 
Meanwhile, it is also possible for the control air to become disordered, with the result that the valve is commanded to the wrong position or operates with the wrong timing.  Again, the results at high speed would be catastrophic, with little recourse even with fast-acting (e.g. computer-monitored and controlled) valves on Wagner throttles close to each steam admission port.
 
There are pictures on the Web that show the effect of valve or timing failure (or water carryover admission) to reciprocating locomotives operaqting at high cyclic rpm.  One set of pictures, on a NYC Niagara, shows the likely effect on modern roller-bearing rods.  There is no guarantee that such an event would not produce a derailment or collateral damage to a passing consist or the track geometry.  There is also the very real danger that such an event would leave the locomotive disabled, and in relatively untowable condition, causing the type of delay that railroads most cite when refusing steam operations on their property.
 
Note that control over the power air will be done with computer systems, reading various kinds of encoders and sensors, executing some form of programmed control that is speed-dependent.  Any of a wide range of failures there will produce a similar effect at the valves.
 
When the system is new, of course, it's relatively less likely there will be failures.  But, as with pulverized-coal firing systems on locomotives, it's only a matter of time before some statistical failure takes place, "and fire looks on and laughs".
 
That is why I favor a relatively simple, mechanical, proportional valve drive and timing mechanism, and why any 'fancy' control of the valves at high speed be done strictly through tripping or other control of the mechanical mechanism.  If there is a failure of the complicated part, the engine is less likely to have catastrophic damage, and it should be relatively more possible to kludge the control system if necessary to keep the engine operating.


Logged
Overmod
TRAINing
View Profile  

Posts: 12
Re: New Build Steam
 
« Reply #54 on: May 18th, 2015, 8:02am »
Quick-Jump   Reply w/Quote   Modify

Now, I might mention that when I was very young, I designed a large high-speed compound 4-8-4 that used spring-operated servo drive to its valves.  This system separated the power used to drive the valves from the system used to time and modulate them, which Ithought at the time was a desirable characteristic.  The valve mechanism was powered by a large 'clockspring' which was wound up by a semiautomatic mechanism driven off one of the driver axles (probably similar to what drove the 'twist' on Leader's valves) and the spring then drove a mechanism that performed valve motion through a complete cycle when tripped (similar in principle to how Corliss valves work, but allowing more complicated movements).  [If anyone wonders, I got the idea for this from how the Varityper typesetting device is powered.]  
 
Note that this gives very high precision and does not require anything other than mechanical control; it can among other things be directly regulated by a cam-type Valve Pilot device to match both cutoff and timing to within the precision limits of the valve gear (probably at least as good as that for British Caprotti, about 3 to 5%).
 
The problem is what occurs if a spring breaks at high speed, which is precisely when I'd expect the thing to be under the most relative stress.


Logged
Pages: 1 2 3  ReplyReply     EMail TopicEMail Topic   PrintPrint

« Previous topic | Next topic »