Re: Steam Physics

Have you given it a go? Post your experiences, particularly things that worked well. Let the team know what works and doesn't work as they cannot test the software with every combination of routes, activities and content available for MSTS. OpenRails is actively being improved and is quickly approaching v1.0. This is a great place to discuss what you can do with OpenRails.

Moderator: Moderators

Locked
User avatar
DarwinS
Very Active Forum Member
Posts: 1355
Joined: Fri May 28, 2004 10:08 am
Location: York

Re: Steam Physics

Post by DarwinS »

Steam Locomotive Physics

Postby DarwinS on 11 Feb 2017 09:03
I would like to start looking at eng and wag files for locomotives and rolling stock and adjusting them to get appropriate performance in Open Rails.
It seems as though many MSTS parameters may no longer be needed, whilst other new factors specific to OR are being introduced.
As I am looking at running in Open Rails only I am happy to rationalise files, delete unnecessary MSTS parameters and perhaps start from scratch using only those things that OR needs. (Recognising that eng files may need to be updated as OR develops.)

It would be good to get a complete guide to what is needed and what is not. I have started using the guide at Coals to Newcastle. http://www.coalstonewcastle.com.au/phys ... meter-std/, The excellent manual provided with OR and the document on steam locomotive physics have all been helpful.

Here are some findings so far. I am not sure that they qualify for reporting as bugs, but they are definitely issues requiring further development.

(1) At present steam locomotives seem to have an "engine only" mode rather like diesels! Try opening the regulator in neutral gear and you can make a huge amount of steam and smoke without going anywhere. It seems that the amount of exhaust is linked only to the regulator setting. As exhaust steam other than through the blower has to pass through the cylinders before going up the chimney probably the exhaust should be based on Cylinder Usage which is controlled by both regulator and reverser.



(2) There is certainly a need to remove some of the "not required" lines for MSTS, and also to make good use of the new ORTS instructions. In an early test with a mixture of ORTS and MSTS physics lines two interesting things were observed. #
(a) It is possible to happily start and drive along locos with the reverser in neutral! [I have not done detailed experiments to find out which MSTS lines caused this but when unused MSTS lines were removed the effect disappeared, so although MSTS and OR eng files are compatible, I would suggest having separate eng files for OR.]
(b) The default calorific value for coal is very low (by British standards!) at 6234Btu/lb. So definitely needs to be raised for UK locos.
A good general value for UK seems to be about 12000Btu/lb, with the best Yorkshire coal on the LMS going up to 13700Btu/lb, best Yorkshire and Durham coal for the LNER top link perhaps at 14000Btu/lb and the GWR Welsh steam coal used before WW2 making 14200Btu/lb or more.
Possibly other people can help with some values for the poorer quality coals. I can remember reading in David Smith's book on the GSWR that when CR locos were sent to the GSWR they were not happy with the low quality Ayrshire coal and the Northern Division sent coal from elsewhere in Scotland for them.



(2) Performance in OpenRails seems to be fantastic if you drive or fire well. However the sim seems to be far too kind if you do your job badly. I had a try with setting parameters for a Princess and driving it flat out, ignoring the OR warning that the fireman was not able to shovel fast enough. Here is a screenshot after 25 miles:



20 miles further along we seemed to reach a steady state with the boiler pressure down to 37psi and the fire mass down at about 50% of ideal. Still we were going along at 70mph with steam production about 80% of maximum and looked as though we could have completed 200 miles in such a state!

In this situation there should be more feedback such that things will get worse and worse unless corrected.
(a) Steam production must surely be related to fire mass and fire temperature as well as grate area and calorific value of the fuel. In this case certainly the fire would have been hot enough until it became thin enough to develop holes, but surely the steam production should have declined more as the fire mass decreased.

(b) As pressure falls I would imagine that the amount of expansion possible in the cylinders would also fall so surely the locomotive could not pull so hard and develop so much horse power with the boiler pressure so low.

(c) A fall in boiler pressure should eventually result in train brakes leaking on and ancillary devices such as compressors, generators, feed pumps not being able to function. It seems this is not yet implemented in ORTS, as in this case the eng file did contain the MSTS instruction:

VacuumBrakesMinBoilerPressureMaxVacuum ( 110 )

(d) Another note with the big pacific was that it did not seem to slip as easily as I would have expected for a loco able to develop such high TE compared to the mass on the driving wheels. The key line in OR being:

ORTSDriveWheelWeight ( 67.5t-uk )

It proved to be virtually impossible to slip the loco on dry rail even with full regulator and full gear. (Thought here did locos tend to slip on departure because rails in busy stations were often wet and greasy where locos stood frequently even on dry days... :-? )
With the advanced option "adhesion proportional to rain/snow/fog" checked and correction factor at 100% (rather than default 130%) there was indeed some slipping in wet conditions.
I would hope that when an activity editor becomes available we could adjust the adhesion correction factor there.

Resuming after a save I was surprised to see that the fire mass and the boiler pressure were restored to very healthy values:



So it seems that there is a small problem to solve with save and resume. (Same location, same velocity, same amounts of coal and water, but boiler and fire conditions have reverted to original having had a boiler pressure of 37psi and a fire mass of 680lb when saved, I did not check if the same applies to boiler water levels.)

OR does not use the initial conditions set in the eng file in MSTS (boiler pressure, water level, fire mass, fire temperature, amount of coal, amount of water...)
To be honest it would be better for these to be defined in the act file for an activity rather than in the eng file anyway. So another input for the forthcoming activity editor. There does need to be some way for the user to vary these however (more than just the hot or cold start choice).

For manual firing I noticed that you do not have to open the fire doors to put coal in :o :-? This might be fine for a mechanical stoker but I did appreciate in MSTS having to open the fire doors to shovel in coal... and also being able to end the shovelling by closing the doors. The control+R function for adding a shovel full at a time is really nice. (I have not looked into sound 'triggers' yet but hopefully we can edit sound files to coordinate the shovelling noises with this function.

Possibly the addition of two disaster functions would be an enhancement -
(1) end of activity due to damaged cylinders if cylinder cocks are not opened on starting
(2) end of activity due to blowbacks if blower, fire door and damper are not used appropriately when regulator is closed and in tunnels.
Regards

Darwin
Regards

Darwin
Regards

Darwin
User avatar
rickloader
Well Established Forum Member
Posts: 675
Joined: Mon Apr 11, 2005 5:11 pm
Location: Hampshire

Re: Steam Physics

Post by rickloader »

Thanks Darwin, very interesting. I keep meaning to investigate OR physics, but find it rather daunting.It isn`t clear whether the current OR physics are finalised, or merely an interim stage. It would be awful set up a whole trainset with new parameters, and then find the physics revised.
Using Danny`s Merchant Navy with 12 on, I found performance poor, and unable to maintain the timetable . (and the 1950s timetable wasn`t exacting). Perhaps you need to load up your Lizzy!
Vacuum braking is woefully under powered, and you can`t stop in the typical 1000 yards between distant and home signal. Some work is being done on Vacuum brakes, but mainly represent the ejector.
One feature needed is an engine steam brake. Shunting is very difficult with the long lag on the current system.
With 3000 hp theoretically produced by a class 8 loco, some sparkling performance should be available. On the other hand, bad driving/ firing should result in failing performance.
I guess part of the trouble is that most train simmers are focused on long, airbraked, diesel trains. Those interested in vacuum or loose coupled, steam trains are rather a minority.
So any guidance for modifying eng and wag files would be very welcome.
thanks, rick
[album 71307 OR_Supporter_Logo_Black.JPG]
User avatar
DarwinS
Very Active Forum Member
Posts: 1355
Joined: Fri May 28, 2004 10:08 am
Location: York

Re: Steam Physics

Post by DarwinS »

Hi Rick

The Lizzy was 14 bogies (449 tons) behind the tender. Able to keep the LMS Special Limit timings without overstretching either the boiler or the fireman.
Generally keeping steam usage to 30 000 lb/hr and firing rate to about 2 600 lb/hr. (Cruising if you can call it that around 70 mph sometimes up to 80+ with 67% regulator and 22% cut off.)

I could perhaps tidy up the eng file and share it. From what I read the boiler and engine physics seem to be at an advanced stage. I imagine there may be some tweaks in the code, but I doubt that there will be any need for content creators to make significant changes to these parts of the files.

Vacuum brakes are still to be worked upon, I need to do some testing to see what can be achieved at present, but there may be changes to come when the OR development team get further with this.

I think things have reached that stage where I can start to edit files and complete the bulk of the work. (Tidying up the file contents as suggested on the Coals to Newcastle site should make future amendments relatively easy if they are needed.) This should allow me to upload alternative OR files for the stock that I have created previously for MSTS.

The big issue that faces me on steam performance is how best to limit the rotational speed of older locomotive types with short travel valves. Maxvelocity used in MSTS is not useful in OR it just shows up in the track monitor as your maximum permitted speed but does not influence performance. No doubt some of the advanced parameters would help but I have not yet worked out the appropriate combination. (Due to limited understanding of them!!)
Regards

Darwin
copperpen
Been on the forums for a while
Posts: 120
Joined: Fri Feb 07, 2014 8:21 pm

Re: Steam Physics

Post by copperpen »

I can speak as one who has been involved in testing the steam physics as they have evolved so far. Only one person is doing the code work, in his spare time. Vacuum braking is being worked on, in between other items. The initial testing was based around the rebuilt MN and an American Atlantic type, both of which had been put through a test plant and documents were available for direct comparison. We arrived at the point where both locomotives performed correctly in OR but with two things outstanding. One is to rework the firing to change from continuous aka mechanical stoker, to intermittent, aka fireman. Once that is done, the stack output can be more reliably linked control settings where volume is concerned, and firing where colour is concerned, probably to include an element of draft from the fire as well.

Turning to the two express locomotives here, I would expect the Lizzy to eventually fail due to blown plugs as the injectors will fail to keep up with demand. The MN will find it hard to maintain a schedule if only using an MSTS type eng file. With an OR eng file there should be no problems in maintaining a schedule.

Locomotive performance can be adjusted using cylinder efficiency, boiler efficiency and exhaust port opening time and size.
User avatar
DarwinS
Very Active Forum Member
Posts: 1355
Joined: Fri May 28, 2004 10:08 am
Location: York

Re: Steam Physics

Post by DarwinS »

Thanks for that information Copperpen.

I think what has been achieved so far is very impressive. It is also really helpful to see the model / flow-diagram to understand how things are set up.

For one person it is quite an achievement as steam physics is so much more complicated than that needed for diesel or electric locomotives.

I am hoping someone with better knowledge and information can assist with how to adjust the advanced parameters to achieve desired performance (is 1crick14a still around?)

For boiler efficiency, I have some very basic idea of how to set this. OR has used the grate area by default to calculate maximum evaporation. With the LMS pacifics the grates were so big that maximum evaporation would be determined by the evaporative surface rather than by the grate area. Reducing the boiler efficiency factor allows for this.

For locos with shorter valve travel or "strangled steam circuits that lead to wire drawing" it seems logical to change the port opening time and size and possibly back pressure or cylinder efficiency. The problem is that short of trial and error I do not know which way to go (or how far) with these.

It might be possible to get some information from engineering drawings about port size and set that correctly for a few examples to begin with.

Drawings give figures for "lead" and "lap", I presume they could be used to calculate the port opening time but not sure how.
Back pressure is probably a part of this, but I am not sure that I yet understand how to determine the figure I need from indicator diagrams (even if they are available from testing!)

Equally I am not sure how to determine cylinder efficiency (though I happily used this in MSTS to adjust performance :oops: )
Regards

Darwin
copperpen
Been on the forums for a while
Posts: 120
Joined: Fri Feb 07, 2014 8:21 pm

Re: Steam Physics

Post by copperpen »

Open Rails is getting more realistic with physics as time moves along. There are a number of factors that will now affect train handling and locomotive performance.
The Curtius-Kniffler and Davis figures affect adhesion and resistance respectively.
The advanced options in the steam eng file are later additions, and thus far have not received attention in detail from myself to see how they affect performance.
The CoalstoNewcastle site does have a UK Hall class locomotive for download which has 4 different configurations to allow comparison between different styles of eng file. They are vanilla MSTS, MSTS with basic OR additions, basic OR style and advanced OR. Using these you can see how the different eng files perform in OR, and get to see how the OR steam eng file is structured. Like MSTS there is a bit of "wriggle room" to get a steam locomotive to perform properly, but not much in the way of adjustments to get a poorly performing one.

I think that trial and error is very much the way to go right now, much like the early days of MSTS, with the main difference being that we now have a good code base to work from.
User avatar
DarwinS
Very Active Forum Member
Posts: 1355
Joined: Fri May 28, 2004 10:08 am
Location: York

Re: Steam Physics

Post by DarwinS »

Having spent the day testing vacuum brakes, I can report good news, bad news and peculiar news....

With a little tweaking it is possible to get the braking performance needed.

Having started from something like this for a bogie coach:

Comment ( *** Brakes *** )
BrakeEquipmentType ( Vacuum_brake )
BrakeSystemType ( Vacuum_single_pipe )
MaxBrakeForce( 23kN )
MaxReleaseRate( 10 )
MaxApplicationRate( 10 )
BrakeCylinderPressureForMaxBrakeBrakeForce( 21 )
Brakepipevolume ( 1.33ft^3 )

I have found that surprisingly few things have an influence up this. For example neither BrakeCylinderPressureForMaxBrakeBrakeForce( 21inHg ) nor
BrakeCylinderPressureForMaxBrakeBrakeForce( 0inHg ) changed things... so that line is out of use for the present.

Looking at the original Kuju Flying Scotsman cars the parameters seemed very much what ought to work in real life.

I tried adding an Auxilary_reservoir and a Triple_valve, but again there was no change in the end result. Looking at the extended HuD it assumes there is an Auxillary reservoir whether or not this is specified. Triple_valves do not show up specified or not.

This brings me back almost to the beginning.

Comment ( *** Brakes *** )
BrakeEquipmentType ( Vacuum_brake )
BrakeSystemType ( Vacuum_single_pipe )
MaxBrakeForce( 23kN )
MaxReleaseRate( 20 )
MaxApplicationRate( 20 )
Brakepipevolume ( 1.33ft^3 )

Now found that making changes to MaxReleaseRate and MaxApplicationRate make little or no difference.

The brake force of 23kN is taken from this document. http://lmsca.org.uk/wp-content/uploads/ ... MSRAVB.pdf which states on page 13
The system of levers between the main pull rod and the brake block beams then multiplies this force even further, until the total of the forces on the sixteen brake blocks is equal to 22 to 24 tons


But using that number it takes nearly 2 miles to stop a passenger train from 75 mph... about double the distance it ought to.
For the present moment, until vacuum brakes are properly sorted and coded, it would seem the only way to deal with this is to double brake force so that at the end of the day my statement reads:

Comment ( *** Brakes *** )
BrakeEquipmentType ( Vacuum_brake )
BrakeSystemType ( Vacuum_single_pipe )
MaxBrakeForce( 46kN )
Brakepipevolume ( 1.33ft^3 )

8) :(
Regards

Darwin
ozziedriver
Been on the forums for a while
Posts: 296
Joined: Sat Jul 02, 2005 11:52 pm
Location: Traralgon,victoria, Australia

Re: Steam Physics

Post by ozziedriver »

How much force is 47Kn in relation to stopping power in MSTS
From memory one of the locos had a braking force of something like 300 % of its weight......impossible
60 % of weight would be about the maximum before wheels would begin to skid

A long time ago while working on MSTS brakes I tried a formula that MSTS code excepted
Same thing for handbrake only the last number % of braking was a lot lower

MaxBrakeForce( 75.85*1016.05/100*30.5 )
MaxHandbrakeForce( 75.85*1016.05/100*15.7 )

75.85 ........is the weight of the loaded truck in tons
*1016.05..... convert tons to kgs which MSTS seems to like
/100...........divides the kg weight by 100
*30.5 ..........% of braking force from builders drawings

This then made it very easy to work on braking forces and was exactly what was designed by the manufacturer
first number is weight of truck in tons and last number is % of braking
Give this formula a try and see what you come up with
Ozziedriver
User avatar
DarwinS
Very Active Forum Member
Posts: 1355
Joined: Fri May 28, 2004 10:08 am
Location: York

Re: Steam Physics

Post by DarwinS »

That is a great idea. Thanks Ozziedriver.

I can simplify the formula and put this in for the moment for passenger carriages:

Comment ( *** Brakes *** )
BrakeEquipmentType ( Handbrake, Vacuum_brake )
BrakeSystemType ( Vacuum_single_pipe )
MaxBrakeForce( 31*1.5kN ) Comment ( first number is tare in uk tons )
MaxHandbrakeForce( 31*0.3kN )
Brakepipevolume ( 1.33ft^3 )

Then when the physics of vacuum brakes have been corrected in ORTS I can use ConText to do a global replace of *1.5kN with *0.75kN so that everything works with the correct brake force specified. :) :)
Regards

Darwin
User avatar
DarwinS
Very Active Forum Member
Posts: 1355
Joined: Fri May 28, 2004 10:08 am
Location: York

Re: Steam Physics

Post by DarwinS »

So best deal for the present time seems to be for passenger carriages:

Comment ( *** Brakes *** )
BrakeEquipmentType ( Vacuum_brake )
BrakeSystemType ( Vacuum_single_pipe )
MaxBrakeForce( 31*1.5kN ) Comment ( first number is tare in uk tons )
Brakepipevolume ( 1.33ft^3 )

For goods wagons and locomotive tenders:

Comment ( *** Brakes *** )
BrakeEquipmentType ( Handbrake, Vacuum_brake )
BrakeSystemType ( Vacuum_single_pipe )
MaxBrakeForce( 26.75*1.2kN ) Comment ( first number is tare in uk tons )
MaxHandbrakeForce( 26.75*0.3kN )
Brakepipevolume ( 0.59ft^3 )

For locomotives:
Comment ( *** Brakes - Wagon Section - General *** )
BrakeEquipmentType( Vacuum_brake )
BrakeSystemType( Vacuum_single_pipe )
MaxBrakeForce( 67.5*1.5kN ) Comment ( first number is DriveWheelWeight in uk tons )
Brakepipevolume ( 1.15ft^3 )

The next question seems to be about propogation time. I will go hunting but perhaps some one knows the answer. For the set up that I have at the moment with a train of loco tender and 14 bogies with a total length of 946ft it takes around 45 seconds to fully create a vacuum from 0 to 21 in throughout the whole train. To fully destroy the same vacuum takes 70 seconds. I am not sure if these are realistic figures or not. I know there is a time delay for simple single pipe brakes to work down the length of a train. In the eng file section I have tried adjusting TrainBrakesControllerMaxReleaseRate ( ), TrainBrakesControllerMaxApplicationRate ( ) and ORTSBrakePipeTimeFactor ( ) but have not yet been able to reduce these times.

The other issue with such a long train is that it seems impossible trying to hold a particular vacuum by putting the brake handle in the running position. Again I don't know if this is realistic or if things are being a little too unresponsive.
Regards

Darwin
ozziedriver
Been on the forums for a while
Posts: 296
Joined: Sat Jul 02, 2005 11:52 pm
Location: Traralgon,victoria, Australia

Re: Steam Physics

Post by ozziedriver »

Wouldn't you still have to give your first numbers (tons) and convert them to KGs *1016.05
I don't think MSTS understands the first number is tons or KGs
It would just see a number you gave it but it is looking for the answer in metric KNs
Hope this makes sense
Ozziedriver
copperpen
Been on the forums for a while
Posts: 120
Joined: Fri Feb 07, 2014 8:21 pm

Re: Steam Physics

Post by copperpen »

It is currently not possible to have proper working vacuum systems as the whole thing is not complete. I have sent Peter everything I have on the subject and he is working on development as time allows. Currently we have code that allows for an ejector, but not a lot else. There are three methods. One used separate ejector handles, one uses the brake handle position to operate the ejectors and then there is the GWR system of the small ejector being a pump driven by the crosshead. Additionally we have the diesel and electric system of exhausters to develop.

For the sake of simplicity I think that brake force figures should be expressed as a complete figure, no multiplication or other operands in there.
User avatar
DarwinS
Very Active Forum Member
Posts: 1355
Joined: Fri May 28, 2004 10:08 am
Location: York

Re: Steam Physics - Vacuum brakes

Post by DarwinS »

Not only are multipliers untidy, but ORTS does not seem to understand them anyway.
All in all best to wait until there is a working model available.

Best option in the meantime is something like the following, but it is only a temporary fix....

Comment ( *** Brakes *** )
BrakeEquipmentType ( Vacuum_brake )
BrakeSystemType ( Vacuum_single_pipe )
MaxBrakeForce( 46kN ) Comment ( 2x actual value - temporary fix )
Brakepipevolume ( 1.33ft^3 )
Regards

Darwin
User avatar
DarwinS
Very Active Forum Member
Posts: 1355
Joined: Fri May 28, 2004 10:08 am
Location: York

Re: Steam Physics - settings for older locomotives

Post by DarwinS »

Finding this old paper on the internet has been a great help with this task: http://users.fini.net/~bersano/english- ... l%20-O.pdf

If I understand correctly, which I am not at all sure about then for a Midland Railway 2P 4-4-0:
ORTSCylinderPortOpening ( 0.048 )

...and for a rebuilt Royal Scot
ORTSCylinderPortOpening ( 0.125 )

Testing those numbers with all other factors kept the same did not seem to make the significant difference to performance that I had hoped!

The curves on p110 of the paper show that in the older design Mean Effective Pressure decreases as rotational speed increases. This effect can be achieved in ORTS by editing the ORTSCylinderInitialPressureDrop

ORTSCylinderInitialPressureDrop( 0 0.94 40 0.77 80 0.63 120 0.52 160 0.44 200.0 0.38 240 0.36 280 0.35 )

Gives the required curve and results in a very significant drop in MEP and IHP at higher speeds, typical of older locomotives with short travel valves. I would highly recommend including this line (or a similar one) into the eng file of the vast majority of older British locomotives. [roughly: GWR before 1902, LBSCR and most LSWR designs, SECR before 1914, LNER and constituents before 1928, LMS and constituents before 1933 except for Royal Scot, Patriot, Crab, 4P 2-6-4T and surprisingly to me the 7F 0-8-0] The default OR values seem to correspond to the performance of more modern designs with long travel valves and well proportioned steam circuits.

Adjusting ORTSCylinderBackPressure also has a significant effect, but as this is a power related curve rather than a speed related curve I am not sure that it is having the appropriate effect that I am looking for. (I was hoping that back pressure might be the factor that would prevent older locos racing away like high speed trains on the downhill bits... more testing is needed but this factor does not seem to have that effect.) I have yet to experiment with the ORTSCylinderExhaustOpening which might be the final factor needed.
Regards

Darwin
User avatar
a214sli
Established Forum Member
Posts: 426
Joined: Thu May 29, 2003 2:08 pm
Location: 'Route-building' North West of England, plus starting Waverley Route join up from Carlisle

Re: Steam Physics

Post by a214sli »

:o at this thread.

I look forward to seeing the fruits of the labours of those who, unlike me, know what they're doing on this.

Including vacuum brakes on diesels and even (whispers) electrics.
Locked

Return to “[OR] General OpenRails Discussion”