[RTP]

The height of the plume depends only of the exhaust temperature and the ambient air temperatura.
In x1950 the height variation with regulator opening is irreal.

Regards.

[Matej Pacha]

James, I think your shape is OK, the problem could be the height (for steam locos). It could be variable - like train speed dependent - the faster the lower - caused by higher horizontal speed difference (definitely with some limit several feets above the train). Great job anyway - I think we can see the Smoke station ahead :)
I've found this video, but there are plenty of such examples on youtube (class 498.0, 2MW, 160 km/h).

[James Ross]

Matej Pacha, on 19 January 2014 - 01:08 PM, said:

James, I think your shape is OK, the problem could be the height (for steam locos). It could be variable - like train speed dependent - the faster the lower - caused by higher horizontal speed difference (definitely with some limit several feets above the train). Great job anyway - I think we can see the Smoke station ahead :)
I've found this video, but there are plenty of such examples on youtube (class 498.0, 2MW, 160 km/h).

That's a pretty good video, thanks!

If it really is only the temperatures that affect the height of the curve as RTP says, it'll have to wait as we don't have air temperature nor steam temperature I don't think.

I've reduced the number of particles in X1951 but will continue to look at the emitters.

[RTP]

In x1951, with the UP9000 running 40mph the plume grows almost vertically.
That is impossible.
The height of the plume is not so variable.
If the air is cold the steam cools faster.
I think that a compromise can be about 50-100 meters.
The steam, or smoke, loses all horizontal velocity in about two meters.
This is the height of the vertical section of the plume.
In any case, steam-smoke are improving.

Regards.

[rdamurphy]

James Ross, on 19 January 2014 - 04:00 AM, said:

What I really need right now is not a new texture (although that's appreciated) but some hard data. :) I'm sure there are people with some data on this so, for a single steam locomotive:

• What is the smoke stack diameter?
  
• What is the typical volume of exhaust coming out of the smoke stack?
  
• What's the curve of smoke like when operating? (Pictures appreciated.)

The problem I've got is that the inputs to the emitter code have been tuned for the old behaviours, but since all the inputs are data (sizes, volumes, etc.) I want the emitter to be as close to reality as possible - and then have the MSTSLocomotive code be tuned to correct for any weirdness that produces due to MSTS data.

E.g. the MSTS in-box Flying Scotsman locomotive has a smoke stack area of 0.031416m^2 and in current OR full-throttle emissions of 12.322m^3/s. That's 392m/s = 1411kph. That seems a bit fast to me.

Those were all set to be read from the MSTS .eng files. Changing those values should be left to the user, rather than hard-coded. Otherwise, there's no customization possible, or perhaps I'm not understanding your question?

Or are you trying to adjust the OR values to use the (wrong) values in the MSTS defaults?

Please accept my apologies if I'm not understanding the question, and I can try to come up with better data for you...

The amount of steam is based on an awful lot of factors, one of the most important is simply ambient temperature. For instance, D&RGW 346, which still runs on a monthly basis, will hardly show steam at all in July at 80+F, but in the Winter, when the temps are 32-F, she can obscure the nearby highway...

Robert

[copperpen]

James Ross, on 19 January 2014 - 04:00 AM, said:

• What is the smoke stack diameter?
  
• What is the typical volume of exhaust coming out of the smoke stack?
  
• What's the curve of smoke like when operating? (Pictures appreciated.)

The problem I've got is that the inputs to the emitter code have been tuned for the old behaviours, but since all the inputs are data (sizes, volumes, etc.) I want the emitter to be as close to reality as possible - and then have the MSTSLocomotive code be tuned to correct for any weirdness that produces due to MSTS data.

E.g. the MSTS in-box Flying Scotsman locomotive has a smoke stack area of 0.031416m^2 and in current OR full-throttle emissions of 12.322m^3/s. That's 392m/s = 1411kph. That seems a bit fast to me.

Smoke stack diameter will vary by locomotive type and exhaust system in use. Unfortunately, unless you have access to the relevant drawing that is an unknown, although it is possible to do a rough calculation based on known dimensions and comparison.

The volume of exhaust emitted is dependent on throttle, cutoff and engine speed.

The curve of the smoke/steam is dependent on both of the above.

That exhaust speed for the Scotsman does indeed look to be high. However, as the state of the fire is determined by the draught caused by the exhaust of steam from the cylinders entraining the flue gases out of the stack this speed will be reasonably high. Gas speeds through the boiler tubes and flues can reach over 200mph. How that translates into exhaust emission speeds I am not sure because I cannot so far locate any data.

Far as the actual curve of the stack emission goes, that is dependent on escape velocity, volume, train and wind speed. As Robert states, that curve and volume is handled in MSTS by the second set of data in each effects data block. The emission point of origin is set in the first line. Edit: The third data line is the diameter of the emission in meters.

[James Ross]

rdamurphy, on 19 January 2014 - 03:10 PM, said:

Those were all set to be read from the MSTS .eng files. Changing those values should be left to the user, rather than hard-coded. Otherwise, there's no customization possible, or perhaps I'm not understanding your question?

Or are you trying to adjust the OR values to use the (wrong) values in the MSTS defaults?

Please accept my apologies if I'm not understanding the question, and I can try to come up with better data for you...

The particle emitter gets given two key values from the simulation: volume (m^3/s) and hole size (m^2). I found that emitting particles at the seemingly-logical speed of volume/hole size and decelerating over a couple of seconds resulted in them rising a km or more and other obviously wrong situations. So, I would like to find out whether these inputs are correct and it is the particle simulation that's off, or if the inputs are bogus.

FWIW, it currently looks like a bit of both, but I'll be constructing a new particle simulation with much faster deceleration which should help.

copperpen, on 20 January 2014 - 02:29 AM, said:

That exhaust speed for the Scotsman does indeed look to be high. However, as the state of the fire is determined by the draught caused by the exhaust of steam from the cylinders entraining the flue gases out of the stack this speed will be reasonably high. Gas speeds through the boiler tubes and flues can reach over 200mph. How that translates into exhaust emission speeds I am not sure because I cannot so far locate any data.

It looks like the smoke decelerates much faster than I was anticipating (although I still think we're getting a higher velocity than might actually be the case in reality), so hopefully when I create the new model of that we'll get lower smoke plumes. I'll be adding in the wind too, though that seems less useful than more randomness in particle direction (which I'll also be adding).

[RTP]

The cinetic energy of the smoke particles is low and the aerodinamic caracteristics poor.
The particles deccelerates in thenths of a second, as much.
The only vertical velocity remaining is due at the diferent density against the ambient air.
The horizontal velocity is, of course zero.
A good shape for the plume will be a parabole, perhaps with some noise added.
Regards.

[markus_GE]

RTP, on 20 January 2014 - 10:08 AM, said:

The cinetic energy of the smoke particles is low and the aerodinamic caracteristics poor.
The particles deccelerates in thenths of a second, as much.
The only vertical velocity remaining is due at the diferent density against the ambient air.
The horizontal velocity is, of course zero.
A good shape for the plume will be a parabole, perhaps with some noise added.
Regards.

Well, horizontal velocity initially is not zero, too. But it will decrease similarly fast to the upwards movement.

No, it wouldn´t be a parabole, it would rather be similar to the graph of a mathematical root function.

Cheers, Markus

[BB25187]

Hi all,

Just one thing I noticed while testing the latest smoke implementation. In some cases, the deceleration has a undesirable side effect: the density of the plume increases with the distance (from the stack). It looks a bit strange, because the dispersion should result in the opposite.
I also noticed some weird gaps and discontinuities in the shape of the plume.

James Ross, on 20 January 2014 - 06:51 AM, said:

I'll be adding in the wind too, though that seems less useful than more randomness in particle direction

I agree: a plausible randomness is probably key (probably even more than any accurate technical shaping parameters). The movement of the air around the machines is partly turbulent. Wind plays its game. With the current implementation, the plume looks very regular and "clean". More irregularity would be fine. Easy to tell, probably harder to render...

The task may be easier for diesel engine which emit a sparse smoke, than for steam locomotives. But in this case a fine tuning of the parameters, an additional randomness in the particle emitter can give a result which looks more "living" (based on a slightly modified rev #1947).

Good luck in this interresting work!

Regards