Atomic Systems

I have been doing some tuning of the adhesion parameter with the idea of having the loco bebin slipping at a certain tractive effort as measured from the HUD steam chest pressure.

This is shown in the thread

http://forums.uktrainsim.com/viewtopic.php?t=62074

I have not yet got a fully stable, universal, calculation, however during the tests some odd displays appeared on the HUD.

The HUD shows slipping at the expected psi in the steam chest and using the sanders traction is regained up to the full available steam chest pressure.

As train speed increases and cut off is reduced the psi in the steam chest drops below the original slipping value the sanders are shut off and it would be expected that slipping would not occur.

This is not the case and after closing the regulator and reopening slowly the slipping should start at about the same psi in the steam chest. In fact slipping starts at a much lower steam chest pressure about 50% of the original slipping value.

Graphically looking at the wheels and observing the HUD steam usage it does not appear to be slipping until it gets to the predicted psi in the steamchest.

If the sanding is started the cycle repeats with an even lower value the next time.

A second phenomena is the situation after a save is reloaded where the sand was in use at the time of the original save.

Here the HUD does not show sanding, but the loco does not slip where it should and appears to have sanding in operation regardless of the operation of the sanders.

I don't believe this can be adressed in the .eng file and is probably a glitch in the hard programming.

Any thoughts and experiences on this would be welcome.

I have been using Conn McCarty's and Bill Hobbs' thoughts on this with some detail modification to the traffic class function.

Ian

I have finally got something that makes sense.

Unfortunatly not the "Saving With Sanding On" issue which is hard wired and the only way to avoid it is not to save while sanding.

There are four entries in the Adheasion (misspelled) parameter.

The second controls the steamchest pressure at which slipping will start.

The third increases the pressure at which slipping will start if sanding is used.

The first shows the pressure at which slipping will stop if sand is not used

The fourth is apparatly a dummy which can safely be set at 0.

The programming seems to allow for slipping to start anywhere between the pressures for the first and second values, especially once slipping has ocurred.

Thus if the calculated value for the second number is for example 0.337, then the first value should be as close as possible and using 0.336 here works.

I probably will allow a little more between the first and second values to gain some limited randomness, but generally the range of the start of slipping is now controlled. The third value should be set so that the loco can utilise full steam chest pressure on dry rail.

Thus the parameter for the 4-8-4 reads as follows:

0.336 0.337 0.471 0

This is loco specific and I am trying to find a constant for calculating between different locos.

Ian

Hi Ian
AFAIK....and I think others will post similar.

the first is on normal rails, the second on wet rails, the third is a multiplier of the second to imitate the use of sand, and the fourth is a dummy or not used.

Adheasion ( 0.2 0.4 2 0 )

Personally I would have thought these fiqures should be pretty much standard for most locos....as steel agaisnt steel must be fairly constant.

If you look through the forums, I'm sure its been discussed at length before.

Cheers
Jon

Jon,

I have looked through the forums and what you say is probably correct for diesels.

The problem is that whoever wrote the hard programming for steam was probably a different person from whoever wrote the one for diesel.

What I have done is not to take at face value anything that was stated in the tech docs or by the early self proclaimed experts.

Most of what they say was correct, however occaisionally you get a result different from what you expect and this is one of those cases.

Another one was the way the air brake on a late steam loco works in the sim compared to the way it works on a early diesel, where, apart from the method of powering the compressor, every notch on the brakes should have worked identically.

They did not!

If we accept conventional wisdom without analyzing what is happening and why, then we will be doomed to mediocrity in this sim.

I work in an industry where part of my job involves investigating incidents.

The last answer I wan't to hear when questioning "Why did you do this?" is "Because we always did it that way" or the alternative "Because the procedure said so and it was issued by the office so it must be followed without question".

I look forward to more input specific to steam on this.

Ian

Some further observations while using the modified parameters:

Adheasion (0.336 0.337 0.471 0)

and

NumWheels (1)

The NumWheels is a strait divisor for the Adhesion parameter

Thus using NumWheels (2) the parameter becomes

Adheasion ( 0.672 0.674 0.942)

The end result is the slipping commences at the same place.

Further observations show that this particular locomotive should slip on dry rail at a steamchest pressure of 175psi, or 50838lbf tractive effort and this is what happens.

The intersting thing is that on entering a tunnel slipping starts at 136psi or 39988lbf tractive effort.

This happens despite the closeness of the first and second numbers in the adhesion parameter.

The steamchest pressure for the onset of slipping cahanges when the value in MaxForce changes so it would appear that there is some hard programming in this where the sim calculates slipping partly using the Adheasion parameter and partly usin a percentage of MaxForce.

The Mass Parameter does not appear to be a component for steam although it is for diesel.

More thoughts later

Ian

Hi folks,

I've only tested adhesion for steam, I found the following (which I think pretty much agrees with Ian's findings):

First value- Co-efficient of friction used during slipping

In real life, the amount of friction between a slipping wheel and the rail is less than that between a non-slipping wheel and the rail. This is why in order to stop slipping, you have to reduce the regulator opening to a lower figure than the setting that started the sllipping in the first place.

Second value - Co-efficient of friction used during normal running (i.e. when not slipping)

Thrid value - Fraction of loco weight that is available for adhesion.

Fourth value - Not used

Numwheels - Just as Ian says, straight divisor for the adhesion.

Using sand seems to set the second value to '1'.

For a 2 cylinder steam loco will all wheels driven, I use something like this:

Adheasion ( 0.07 0.19 1 0 )

and always set the numwheels to 1. The third adhesion value can be used to set the proportion of adhesive weight, so numwheels seems redundant.

You can calculate the steamchest pressure that will start a slip using the maxforce, mass, max boiler pressure and adhesion paramaters. More later once I've woken up a bit.

http://en.wikipedia.org/wiki/Friction

may be of interest as well

cheers

Rich

Still trying to work out what the sim is doing with the data here

Latest test is to set the boiler pressure to a very high level (500psi)

Retire to the bunker and gently open the throttle to see where slipping commenced on dry rail.

The actual adhesion parameter was unchanged and the intention was just to get a more useable range of slipping to establish a base for tractive effort (Max Force) changes which I already knew changed the start of slipping.

The surprise was that the slipping started much later than previously.

It appears that slipping, with these particular parameters occurs at 80% of boiler pressure regardless of what the boiler pressure is.

More Later

Ian

As far as I can tell, the sim calculates the tractive effort being applied to the wheels at any one moment by dividing the steam chest pressure by the max boiler pressure and then multiplying this by the max force. So:

Tractive effort at any one moment = (Steam chest pressure/Max boiler pressure)*Max force

This would agree with your finding of the slipping always occuring at about 80% boiler pressure.

If you keep the mass, boiler pressure and adhesion parameters the same and vary the max force, slipping should take place at a different boiler pressure.

The frictional force between 2 surfaces (wheel and rail in our case) is normally calculated by the formula:

F = μ*N

Where F is the frictional force, μ (the Greek letter 'Mu') is the co-efficient of friction and N is the force pushing the two surfaces together at 90 degrees (called the Normal force). In MSTS terms, the Normal force is the force created by the mass of the vehicle being acted on by gravity. The first 2 terms of the MSTS adhesion parameter are different values of μ.

The two force terms need to be in the same units, so both in lbs or Newtons. To convert Kilograms to Newtons, multiply by 9.81. I think MSTS uses metric tonnes for its 't' mass units, so 1t =

1000kg
2204.6 lbs
9810 Newtons
9.81kNewtons

One thing to be aware of is that MSTS provides a way of altering the value of the force of gravity for a route. It's the 'Gravity compensation scale' term in the route properties. Altering route gravity does change slipping characteristics and also accelerations downhill as it should, as far as I can tell anyway

Cheers

Rich

Examples of Adhesion Effect Changes:

Original Settings:

Boiler Pressure 220 psi
Tractive Effort 63749 lbf
Steam Chest 175.5 psi
2nd Adhesion 0.470
Effective Tractive Effort 50854lbf

Test 1

Boiler Pressure 500 psi
Boiler Pressure % Change 127.27%
Tractive Effort 63749 lbf
Tractive Effort % Change 0%
Steam Chest 398 psi
Steam Chest % Change 126.78%
2nd Adhesion 0.470
2nd Adhesion % Change 0%
Effective Tractive Effort 50744lbf
Effective Tractive Effort % Change -0.22%


Test 2

Boiler Pressure 220 psi
Boiler Pressure % Change 0.00%
Tractive Effort 75000 lbf
Tractive Effort % Change +17.65%
Steam Chest 149 psi
Steam Chest % Change -15.10%
2nd Adhesion 0.470
2nd Adhesion % Change 0%
Effective Tractive Effort 50795lbf
Effective Tractive Effort % Change -0.12%


Test 3

Boiler Pressure 220 psi
Boiler Pressure % Change 0.00%
Tractive Effort 100000 lbf
Tractive Effort % Change +56.87%
Steam Chest 112 psi
Steam Chest % Change -36.180%
2nd Adhesion 0.470
2nd Adhesion % Change 0%
Effective Tractive Effort 50909lbf
Effective Tractive Effort % Change -0.12%


Test 4

Boiler Pressure 220 psi
Boiler Pressure % Change 0.00%
Tractive Effort 63749 lbf
Tractive Effort % Change 0%
Steam Chest 186 psi
Steam Chest % Change +5.98%
2nd Adhesion 0.482
2nd Adhesion % Change +2.55%
Effective Tractive Effort 53897lbf
Effective Tractive Effort % Change +5.98%


Test 5

Boiler Pressure 220 psi
Boiler Pressure % Change 0.00%
Tractive Effort 63749 lbf
Tractive Effort % Change 0%
Steam Chest 157 psi
Steam Chest % Change -10.54%
2nd Adhesion 0.447
2nd Adhesion % Change -5.00%
Effective Tractive Effort 45494lbf
Effective Tractive Effort % Change -10.54%

So varying parameters have the following effect

Increasing Boiler Pressure: Steam Chest Pressure increases in direct proportion but tractive effort for onset of slipping remains the same


Changing Tractive Effort:

This has a variable effect on steam chest pressure, but tractive effort for onset of slipping remains the same

Changing 2nd Adhesion Parameter:

A change in adhesion parameter results in a 2x change of both steam chest pressure and tractive effort for onset of slipping.

Now the challenge is to get this to calculate the 2nd Adhesion parameter consistently.

Ian

cheers for those tests, I'm a bit hampered by being 90 miles away from my PC, so no MSTS access for a bit .

Just wondering if you made a note of the mass and the third adhesion term perhaps? The 1st 3 tests behave as I'd expect, but the last two puzzled me a little.

Cheers

Rich

Rich,

So far I've been trying to get a basis for calculations.

Mass was ruled out earlier as a factor in steam adhesion, but is a factor in diesel.

I have done Boiler Pressure and 2nd Friction Parameters and both at least have a strait line relationship to change in steamchest pressure.

I'm still working through change in tractive effort which seems to have a curve in it.

However a mound of leaves awaits my services outside, so more on this later.

Ian

supergoods wrote:Rich,

So far I've been trying to get a basis for calculations.

Mass was ruled out earlier as a factor in steam adhesion, but is a factor in diesel.

That's a bit bizzare, I wonder if it's one of those time MSTS behaves differently on different PCs. Have know this happen with other things! My tests have shown the mass to have an effect on steam loco slipping on my set up.

I've discovered I do have a default MSTS installation on a PC here, I'll have a play with the default scotsman and get back to you

Cheers

Rich

Ian,

had a bit of a play now. I've done a test with the default Flying Scotsman (completly unmodified) to see what steam chest pressure makes it slip on my set-up here. The answer is 142 psi on a dry rail, on the S and C route. I then changed the loco mass to 66.25t (instead of 96.25) and slipping started at 98psi. I've knocked up a quick spreadsheet to calculate where slipping should start, maybe you could see if it works for your pc? Perhaps if you could PM me your email address again please (I can't access my normal email account at the moment) and I'll send the spreadsheet over?

BTW, I found the easiest way to do slipping tests is to use the Scotsman and 8 coaches consist, but to set the Scotsman's MaxBrakeForce to zero. That way you can leave the brakes on the train and stop it moving, but still get the driving wheels to slip.

HTH

Cheers

Rich

I suspect we may have backed in to something else important for steam.

The default Scotsman parameter is:

Adheasion ( 0.15 0.3 2 0.5 )

Conventional wisdom has the 4th value at 0 as it was stated that it doesn't work. This may be true for a diesel, but for steam it may represent the adhesive weight as a proportion of the total weight.

If this is correct, and it will need some testing later to prove, the whole parameter may work correctly.

Ian

"Adheasion ( 0.2 0.4 2 0 )
Personally I would have thought these fiqures should be pretty much standard for most locos....as steel agaisnt steel must be fairly constant."

This is not true, especially for modern diesel electric locomotives (steam will actually be a slight less in many cases, due to the lack of a smooth power curve, this is why tapping an object with a hammer will break it loose).
For diesel electric locomotives with modern traction monitoring systems, the adhesion is actually able to get up to 36% of weight on clean dry rail, and AC units approach 50% adhesion under optimum conditions. This results because the adhesion actually increases if the wheelslip is allowed to take place, but carefully controlled to not exceed 11%. Computers monitor the wheel rotation speed, and compare with actual travel speed determined by radar. This is called creep. AC units have better adhesion, because they have a much more precise control of torque application.
Bob