Can a Computerized System Improve Performance of Miniature Steam Locomotives?

[joanlluch]

Hi, I wonder if someone has experience on steam engines or locomotives, even miniature ones.

Some time ago I had the thought of applying a computerised system to miniature steam locomotive. The idea is to remove the traditional mechanism that drives the steam distribution valves (that's known as Walschaerts or Stephenson distribution) , and replace all that by a system consisting of solenoid valves -possibly steam driven- that would be controlled by a PLC or an Arduino. The idea is to figure out a set of "distribution maps" that would optimise the steam distribution to the engine cylinders -both intake and outtake flows- at all times based on required torque, current speed, available steam pressure and so on.

I do not know of anybody who has attempted this before, so any thoughts on whether this may have sense, would be useful.

Thanks.

 

[Bystander]

available steam pressure and so on

You're probably wasting your time. You'll be trying to balance boiler feed rate, firing, and water level with a control system to "optimize" century old technology that might have been 15% efficient when new.

 

[joanlluch]

Hi bystander, what you suggest is obviously true, but that's not the point.

I build working miniature steam locomotives as a hobby, not because they are efficient or up to date, but because I like them. There is a number of people that does it for the same reason.

Said that, I am not referring to boiler optimisation as this is already very common on industrial boilers. I was particularly targeting real time steam distribution to the engine cylinders. Think on dynamically variable admission and exhaust.

My question is still relevant because: (1) I regard it as a hobby, so I do this just for fun (2) full size steam locomotives are still in operation for tourist rail tours, and new main line steam locomotives have been built as late as 2008 in the UK for that exclusive use, which keeps a great demand. (3) I would be surprised if nobody has attempted or thought on that before, which is why I asked.

 

[Nidum]

As discussed on another forum the conventional steam locomotive in it's simple form probably achieved most of what was possible and to achieve even a little more requires added complexity and considerable additional cost .

The steam engine is always hampered by simple thermodynamic limits - at any conventional pressure and steam temperature efficiency is always going to be low . With a coal fired engine in older times efficiency did not matter too much - there was plenty of coal and it was cheap .

In modern times a steam locomotive would sensibly have to be oil fired and more efficient . Given oil firing it really makes more sense to use a diesel engine and burn the oil directly .

So question is could a blank paper design steam engine ever be designed to compete with a diesel engine . Theoretically yes - but engine would not look anything like Flying Scotsman .

A quasi closed cycle engine running at high pressure and temperature and packaged like a diesel engine could potentially be very efficient indeed and a computerised engine control and management system would be a worth while part of the overall design .

Probably best to abandon traditional ideas about piston steam engines altogether and base any new designs on turbomachinery .

With all electric transmission and catenary collection option fitted as standard a very effective and relatively simpe go anywhere locomotive might be possible .

Problem is though that no one is going to want to go in this direction for a new locomotive design . Diesel and catenary electric trains have already been firmly settled on as the best machines for working UK mainline railways

 

[Bystander]

miniature steam locomotives as a hobby

Equal miniature boilers, equal large upsets in boiler pressure, and what I would anticipate to be a real headache keeping dry steam in lines and valves. As you say, "It's a hobby." Let us know how it works out.

 

[rootone]

You never know, maybe with computerisation of some parts you might get up to double the fuel efficiency of a standard steam engine.
Though it would probably still be less efficient than say a diesel, or a gas turbine using an equivalent amount of fuel.

 

[joanlluch]

Equal miniature boilers, equal large upsets in boiler pressure, and what I would anticipate to be a real headache keeping dry steam in lines and valves. As you say, "It's a hobby." Let us know how it works out.

Hi Bystander. I get the impression that you overestimate the difficulty of it. You can believe that miniature steam locos are actually made and they work !. Search for "miniature steam locomotive" in google images and you will see. As in full size practice, steam superheaters are used to keep steam dry.

 

[joanlluch]

Hi Nidum and rootone.

Well, I doubt any new design of a steam engine could ever reach the overall efficiency of a diesel engine, much less that of an internal combustion gas turbine. To my knowledge, steam turbine based power plants which run 24 hour a day 365 days a year are not either able to achieve the efficiency of a diesel. So to me the superiority of diesels is not under question, and was not at all the purpose of my original question.

In the efficiency range of a -classic- steam engine it is known that the higher the steam pressure and the larger the expansion rate of steam in the cylinders the better the efficiency. To maximise engine efficiency (leaving boiler efficiency at one side for now), a steam locomotive should be operated at full open regulator, and the shorter possible steam cut-off, which will still provide the required power. The higher the boiler pressure the shorter cut-offs must be achieved for max efficiency. This makes me think that a system that (1) would fully and sharply open the steam intake at the beginning of the piston stroke, that (2) would completely and quickly shut it down at a predetermined precise cut-off point, that (3) would keep the exhaust fully open during the full piston exhaust stroke, should behave better -and be more controllable- than the traditionally mechanically driven distribution valves. Now whether this potentially would increase efficiency up to an additional 1% or a 5% or none at all, is what remains to be seen :-)

 

[Nidum]

Designing electronic control systems for a miniature steam locomotive would certainly be an interesting exercise .

Need to start further back in the overall design procedure first though . Most model size engines use steam at 100 psi or so and moderate superheat . Trouble is this steam seldom gets utilised effectively . Tests and calculations show that MEP in model size cylinders is abysmally low ~ 30 psi or sometimes even less . To get anything like worthwhile efficiency this figure has to be raised a lot . Origin of problem is that model size cylinders are usually too big in bore . Reducing cylinder bore would immediately raise MEP to more satisfactory levels .

Valve events in both full size and better model designs are actually quite close to ideal as they are . Possibly a better way to go than using electrically actuated valves would be to have a look at electronic engine management but leave the basic valve gear as conventional .

The computerised system now monitors all functions and automatically adjusts settings of valve gear and regulator for best efficiency consistent with effective working of engine in hauling train .

 

[rootone]

I am sure it will be a fun hobby project, let us know how it goes.
Next project is to send one of the new fangled devices back a couple of centuries.
Steam powered time travel!, imagine where we might be by now!
(Sry, just joking - do have fun with it.)

 

[jim hardy]

Myself i love the idea.

If you can find a Marks Mechanical Engineer's handbook frm 1920's to 1940's there is an interesting chapter on "Carding" of reciprocating steam engines. Basically that's making a PV diagram by actually recording pressure versus piston position. It's how you optimize a stationary engine.
What fun it would be to watch that real time at various loads and valve timing !
snip of an actual "card"
http://www.oldengine.org/members/diesel/indicator/Indicator1.htm

http://trumpetb.net/loco/steampv.html

Check "Florida Flywheelers" antique engine show , February in Zolfo Springs.

 

[Baluncore]

Are these engines all double acting, compound and/or triple reduction engines?
There are plenty of valves to control and optimise there.
Those valves will need to switch high flow and/or high pressure quickly and efficiently.
What type of valves will you use, slide valves, balanced spool valves or diesel common rail type injection valves?

 

[jim hardy]

Union Pacific's steam shop in Cheyenne is open to public one weekend a year, check it out.
A few years back I was stuck in a traffic jam on an overpass there - heard a steam whistle, then a steam locomotive huff-puff, lo and behold (i think it was) a "Big Boy" with a dozen passenger cars roared underneath and sped off to the west... engulfing the overpass in coal smoke.
What a treat !

 

[Nidum]

Null post

 

[joanlluch]

There's quite a number of people building and running these miniature locomotives. Most of these locos are pretty detailed. Unfortunately, there seems to be in the hobby a strong bias towards model perfection rather than physics or Engineering. I am sure that most in the hobby don't even care about the Engineering aspects of their own locomotives. So the result is that all these locos are pretty standard and similar from the point of view of efficiency and construction. In fact, only a few people bring attention or actually understand anything beyond essential aspects of steam technology.

To make things worse, on most countries there are strict regulations on miniature locomotive boilers, which limit capacity, max pressure, building material and other aspects of them. In many cases this prevent users from deviating from the stablished status quo, and this is yet another reason for lack of innovation.

A significant number of plans for these locomotives exist that you can use to build your own, but all them were drawn by hand before the 80's -most date from the 60's, or 50's or even earlier- and in my opinion these plans are all outdated. Not to mention that some of them contain errors that have never fixed.

So the result is that efficiency on these locos is really poor even for a steam engine, the use of modern materials is virtually inexistent, and no industrial standard parts or procedures are normally used.

My intention is to draw plans of one of such miniature locos with a 3D CAD software in a way that parts could be made in a standard way and easily assembled, where no arbitrary impositions are made on design, and where I can try several things. By this I do imply that my loco will be super-performant or super-efficient, as I also aim for it to look more or less standard, but at least I am in for a fresh approach to what is normally made. I will keep posting ideas, questions and thoughts.

 

[Nidum]

One of the problems with full size conventional steam locomotives was that it was difficult to adjust the boiler steam generation rate quickly in response to varying demand . Drivers were very skilled and learned to anticipate demand but nevertheless they generally erred on safe side and maintained a high rate of steam generation all the time when running . This was of course not very efficient .

Another aspect was that engines had a lot of preparation time while steam was being raised and also many engines spent a large part of their working time standing idle - especially UK goods trains which could spend half their working time on standby . All this wasted time while in steam again led to poor overall efficiencies .

It's academic now but devising ways to get faster boiler response times and better utilisation of engines during working day would probably have raised overall efficiency considerably .

On model size boilers there is still considerable scope for improvement - but not with coal firing . Gas or oil firing with a boiler purpose designed to use these fuels and with a proper control system should show considerable improvements in performance and utility .

 

[joanlluch]

Hi Nidum,

Indeed, a fast reacting boiler is key to optimise looses due to idle/standby times and non steady operation with varying demand conditions.

I will attempt to use gas (such as butane) for heating the boiler. My idea is that I should be able to completely and immediately shut down gas supply when the loco is idle. As opposed to coal fired locos, gas will stop adding heat as soon as demand ceases, thus allowing for setting the boiler in standby very quickly without wasting combustible energy. Also, when the steam generation demand does suddenly increase, a gas burner is able to instantaneously provide additional heat with no delay.

On standard miniature locomotives, a safety valve is the main device controlling boiler pressure. Every time the valve opens causes waste of water and energy. Ideally, steam pressure should be maintained high at all times, but the safety valve should never open.

In my case, I may enable a closed loop control system that automatically actuates on the gas burner to keep a target constant pressure in the boiler, just below the safety valve setting, but preventing the valve from opening. A gas fired system has minimal heat inertia, so I assume this is possible in such a system. I can imagine that even a simple on-off system should be able to reasonably control the boiler pressure.

Another key aspect I consider is assuring the boiler is fully and perfectly insulated, so that the locomotive gas consumption is reduced to the minimum -near to zero- when the locomotive is in standby, but still ready to deliver full power at any time.

 

[jim hardy]

Oh no. another boring anecdote ..

Didn't coal fired steam locomotives regulate firebox airflow to control the boiler short term ?

A friend of mine, Gordon J, was a young fireman on UP "Big Boys" in Wyoming just after WW2.
He told tales of the engineer hollering at him for not getting coal into the firebox fast enough when approaching an uphill grade.
Exhaust steam vents up the stack through an eductor that pulls air through the firebox, creating a bit of "Feed Forward" control for steam pressure. More steam through the cylinders increases airflow through firebox, and fuel is mostly air after all.

Since demineralizers were not common back then locomotives used local water which limed up their boilers, requiring them to be cooled down and descaled every few weeks.
http://orr.gov.uk/__data/assets/pdf_file/0017/2636/RSP6-locoblrs.pdf

He also related that when diesels came in in early 50's they lacked the power of "Big Boy" , in fact he rescued some stuck diesel freights with the Big Boy. UP re-routed track around some steep places between Laramie and Cheyenne to accommodate their new diesels..

Diesel replaced steam not because of performance but because it made maintenance less labor intensive.

https://en.wikipedia.org/wiki/Union_Pacific_Big_Boy

...zzzzzzzzzzzzzzzzz... end anecdote...zzzzzzzzzzzzzzzz...

 

[Baluncore]

What type of valves will you use, slide valves, balanced spool valves or diesel common rail type injection valves?

The efficiency and the ability to control the chosen valves will make or break the project.

 

[joanlluch]

Hi Jim,

Thanks for the links and statements. The first link is an interesting read, btw.

On a full sized locomotive it has little sense to use gas, but I am referring at all time to a miniature loco. If steam locos were to return to the tracks I think that one combustible to be considered should be biomass. Not sure in the US or UK, but in continental Europe there's a significant production of biomass. Most of it is used in power plants, but why not use it directly on locomotives, maybe through gasification reactors?

 

[joanlluch]

The efficiency and the ability to control the chosen valves will make or break the project.

I have not looked at that in detail yet, as I am still at earlier stages of my build. I am open to suggestions. It must be a kind of valve that opens/closes reliably and fast and is able to stand superheated steam conditions. My take is that this may have to be designed specifically. One possibility is to use traditional piston valves in the usual way and replace the mechanical gearing by electrically driven solenoids. Better if the valves are steam powered so that the required electrical power is kept to a minimum.

 

[anorlunda]

I think it sounds like a lot of fun. But for me, the fun would be the design and tuning of the control algorithms. I would probably wind up using a simulation of the locomotive, not the physical thing.

You have to decide yourself where you would derive the most pleasure from your project.

 

[joanlluch]

I think it sounds like a lot of fun. But for me, the fun would be the design and tuning of the control algorithms. I would probably wind up using a simulation of the locomotive, not the physical thing.

You have to decide yourself where you would derive the most pleasure from your project.

Yes, control loop tuning is a lot of fun, until you have done it for decades. It happens that I have been in the industrial automation field and the development of automatic control software for 20+ years in the chemical industry, and to me this is like bread and butter. Previously, in the early 90's, I have been involved in developing simulation software to find solutions to physical problems. The software used iterative algorithms applied on math models of physical systems to converge to a solution.

Currently, I am half retired and my interest has switched to making actual physical things. A real steam locomotive (in the sense of something touchable and made in metal) is something that appeals me at this time.

 

[Nidum]

Have a look at how the valves work on locomotive Duke of Goucester . Mechanically driven but probably one of the best ever designs of valves for steam cylinders .

 

[jim hardy]

Caprotti ?

now THAT's something !
youtube.com/watch?v=pIK5mnd_h_U

 

[Jobrag]

I don't know how big your loco is going to be but I think that one of your big problems is going to be the size of controller and IO that you'll need and the power supply for it. As to steam returning to full size locomotives I very much doubt it, Brunel recognised that it is stupid to carry your engine and fuel around if you can put it trackside, his atmospheric railway was a failure due to limits on available materials, but the idea was sound, these days it's called electrification.

 

[jim hardy]

Brunel recognised that it is stupid to carry your engine and fuel around if you can put it trackside,

Big IF there.

 

[Baluncore]

I have not looked at that in detail yet, as I am still at earlier stages of my build. I am open to suggestions. It must be a kind of valve that opens/closes reliably and fast and is able to stand superheated steam conditions. My take is that this may have to be designed specifically. One possibility is to use traditional piston valves in the usual way and replace the mechanical gearing by electrically driven solenoids. Better if the valves are steam powered so that the required electrical power is kept to a minimum.

A “computer controlled steam engine” requires a steam generator and a rotary converter driven by steam expansion. It probably should use pistons to look like a “steam engine”.

The mathematics, software, processors and the steam engine mechanisms are all available today. The missing link needed to satisfy the intention expressed in the OP is the valve technology. It is the control of the steam generator, the pressure to rotation converter and their interface that is critical.

Pneumatic valve technology will be the first place to look for valve technology, followed by hydraulic valve design.

The minimum needed to identify the valve requirements is a computer simulation of a parametrically defined double acting compound engine. That simulation will identify the flow rates and valve timing variation required to efficiently optimise control of the thermodynamic cycle.

What is that thermodynamics simulator called ?
Is it available as freeware ?

 

[joanlluch]

I don't know how big your loco is going to be but I think that one of your big problems is going to be the size of controller and IO that you'll need and the power supply for it.

Hi Jobrag, The miniature locomotive is big enough to hold a small sized PLC controller including batteries to power it. Also you have the option to fit it all in the "tender". To have an idea of how big such locomotives are you must consider that the track gauge is 127 millimetres (or 5 inches). See these two videos for an example:



.

This one is quite spectacular:

.

The second locomotive is not running very efficiently because it is releasing steam all the time through the safety valve on top of it. The first and third ones release a lot of steam only when they stop at the end of the track, which ideally should be also avoided.

These kind of things are very popular in the UK, but people making them have usually a poor knowledge of physics or steam technology.

 

[Baluncore]

Please attach video to your post rather than embed video in your post. My internet is volume restricted and slow. I cannot follow threads that lock up my slow wireless links. I have unwatched this thread.

 

[OldEngr63]

Jim Hardy mentioned the idea of indicator cards. There is a very, very comprehensive book on this subject that has recently been reprinted and is available from Amazon. The title and author are
Twenty Years With the Indicator
by Pray.
It should how a really skilled interpreter could get all manner of information about steam engine operation from studying an indicator card.

 

[Nidum]

A useful tool in electronics is the equivalent circuit where complicated devices and complete circuits are described in terms of assemblies of basic components . Things like current and voltage sources and RLC components .

Another tool more generally used in engineering is an equivalent circuit where everything is described in terms of an assembly of connected 'black boxes' each with a transfer function .

Any thoughts as to what the equivalent circuit of a steam locomotive system would look like ??

 

[jim hardy]

Any thoughts as to what the equivalent circuit of a steam locomotive system would look like ??

i think it'd be the processes in a thermodynamic cycle
adiabatic and isothermal compression or exansion
isentropic expansion
and so forth
there are lots of cycles Rankine, Otto, Diesel, Stirling, Ericcson, Brayton... google finds plenty of images
"Carding" an engine actually plots one for you.

 

[joanlluch]

Please attach video to your post rather than embed video in your post. My internet is volume restricted and slow. I cannot follow threads that lock up my slow wireless links. I have unwatched this thread.

Sorry for that. However I do not see how embedding a youtube video is any different than embedding a photo. To my understanding the video will not play unless you click on it, so you have the option of not watching them anyway and no video streaming will waste your bandwidth. Maybe you should ask people to avoid embedding pictures as well, as they may have the same effect on your internet connection (?). Please (anyone) let me know in case I broke some forum rules or something. I will be happy to fully adhere to any guidelines for better use of the forums. I posted the videos because I thought they were informative, and by embedding them you could watch them without leaving the forums. Thanks.

 

[OldEngr63]

i think it'd be the processes in a thermodynamic cycle

JIm, you are certainly correct as far as the nature of the problem. The mathematical description is something else. With most "equivalent circuits," you can simply apply Kirchoff's laws and have the describing equations. In the steam engine, the best I've been able to do is to describe them indicator card piecemeal, with several different descriptions spliced together to get a good approximation to an actual card (Ive been trying to model a diesel engine card). It is a lot of trial and error (bash to fit and paint to match) sort of thing.