Atomic Systems

For me, it would seem logical to follow a historical progression of sorts. Starting off with very basic early Victorian era locomotives which basically have a boiler and cylinders and a mechanical feed pump, and then slowly progress into more modern locomotives in a logical historical fashion, introduction of hand brakes, introduction of injectors, introduction of steam, vacuum or air brakes, introduction of superheating, feedwater heaters, compounding, condensers, electrical lighting etc, heck maybe even steam turbines. And you keep up this trend of gradually getting more complex until you eventually reach more modern locomotives like a BR standard or Spam can, or heck maybe something fancy and German and by that point you have all the building blocks you need right from the very start of steam railways.

After all there's very little point starting off at the deep end and getting a hodge podge of different parts which barely work or are just tailor made for one locomotive... The advantage of doing it in a logical fashion is with each generation you can refine and improve upon that of the previous say the boiler simulation for example (rather like how boilers improved over time).

Anyway thats my thoughts

kindest regards

Edward

I would think that would be better done in-house, Edward, than in public. I can see your point of making progress in small steps but wonder how well people would take to what would, inevitably, be a flawed machine. Would there be a demand for an early type of loco with grasshopper valve gear, no brakes and a Lancashire-type, single-flue boiler? That would be a dead-end development in many ways for DTG because it would be of its own time. Multi-tube boilers and blast-pipes came in with Rocket but which version of that would you produce? The original, with steeply-inclined cylinders, which lasted little time after Rainhill, or the modified one where the cylinders were placed much nearer the horizontal that worked better but which nobody would recognise?

Can we sensibly simulate the difference between slide and piston valves? With very few people with the experience of driving locos of either type who would know whether the simulation was right or wrong? Should we model early piston valves with one or two thick rings as well as later types with many thinner ones, where the differences would become only gradually more apparent with mileage after shopping? Where do you draw the line? In my book the crucial question for any developer is quite simple: does this aspect enhance, have no effect or detract from the player's enjoyment? If it's either of the last two then it should be binned.

Surely the fundamentals of steam simulation are less complex than you suggest. We need a boiler with a known evaporation rate under different conditions (should there be variants that need a wash-out or have a variable number of blocked tubes?), maximum pressure and an idea of how that translates to power at the rail-head at different regulator openings and cut-offs. Saturated or superheated don't seem to come into it, though it might well affect coal consumption but beyond that?

Keith

I probably wouldn't go as far as the thickness of piston rings, but certain the various types of valves can be differentiated. A balanced slide valve, D valve, piston valve, poppet valves etc will all behave quite different with difference in resistances and resulting drag, along with the ability to let steam into and out of the cylinders effectively with their resultant effective on the cylinders performance and hence driving performance. An engine with slide valves for example can't really be driven at less than say 30% cutoff, while one with piston valves can be driven down to about 10-15%, poppet valves even lower.

The reason I suggest this method is it makes like much easier for developers and players alike since you get a consistent set of tools as it were, instead of the current steam simulation which is a complete and utter mess and relies upon both developer knowledge to overcome its short comings and his patience to bend the scripting to make something akin to real life performance. The resulting effect also being that no two developers have engines which quite drive the same, if given the same engine to work on all the devs would produce quite wildly different engines at the end. There is also the aspect that technical data on steam locomotives, especially specific and less well known (heck even some well known ones) are very difficult if impossible to come by in detail, as such the ability for the tools to for example to estimate evaporation based on physical properties such as tube diameter, length pitch, pattern etc would be important for ensuring consistency and preventing the previous disaster. In other words new system should be designed to allow consistency and realism with ease and the key to that is having as many components as possible.

Developing in the fashion I mentioned is more of a logical approach, you shouldn't suddenly start with the most complex thing possible, you slowly build up to it, otherwise problems are bound to occur. I never said you had to make every engine from Locomotion up to Evening Star needs to be made to get the full set of component, only that the components should generally be developed in a logical chronological fashion with the obvious possibility content being generated as you go along.

regards

Edward

metrobus wrote:I probably wouldn't go as far as the thickness of piston rings, but certain the various types of valves can be differentiated. A balanced slide valve, D valve, piston valve, poppet valves etc will all behave quite different with difference in resistances and resulting drag, along with the ability to let steam into and out of the cylinders effectively with their resultant effective on the cylinders performance and hence driving performance. An engine with slide valves for example can't really be driven at less than say 30% cutoff, while one with piston valves can be driven down to about 10-15%, poppet valves even lower

The type of valve has a definite influence on performance, so it's a very good idea to give devs the option to differentiate. This shouldn't be set in stone (i.e. no drop-down list somewhere where I can pick a valve), but more in the way of a method of setting the induced friction of the valve and the obstruction to steam flow (i.e. a discharge coefficient). Because in the end a steam engine valve is just that, a black box that produces friction and a opening for steam to flow.

Apart from that, why not keep track of the valve motion. Formulas are available for (what I know of) Stephensons' and Walschaerts' , DTG might want to google for Zeuners' approximation. This gives the oppertunity to take the numerical derivative (speed -> valve friction), openings, and different valve events so you have varying pre-admission and the lot. I've done some work in this area; the key issue is that this sort of approach requires short physics ticks, shorter than what low framerates allow for.

Don't like the above? Fortunately the implementation should be simple enough that using two .csv's to produce a tractive effort numbers is a possibility.

As far as the sentiment of starting easy, the expanding to more difficult steamers, I'd agree, but then I consider something among the lines of an A4 to be actually a remarkably simple machine. And rather sellable. Don't start of with the de Glehns, do them second .

Because conjugated valve gear always behaves itself and doesn't at all produce very weird valve events on the middle cylinder De Glen compounds would be amusing if not a headache to drive

regards

Edward

The conjugated gear is solvable given the valve displacements in the outer cylinder I would think. Add math.random to simulate the bushings gone bad . At least de Glehns are drivable (and startable...), unlike the brain children of a certain LNWR chief engineer .

The nerdy point here is to facilitate nerds like me to simulate these things without resorting to ugly hacks.

It is relatively easy to model the engine as a (black) box containing other boxes. For each of these boxes (e.g., evaporation, valves), you can come up with a set of boxes to fill in the details, plus you can/must come up with a set of formulas or CSV tables to describe its behaviour in black box style. You need the latter while you do not have the detailed model. And you can derive it from the detailed model if you are worried about performance. Of course, that assumes CSV tables with many inputs aka n-dimensional function plots. In practice, those could be what we resort to for the actual game run.

The difficult part is to find & agree a complete list of inputs and outputs of each of these boxes. A good technique is to look at the most simple implementation as a dummy and at the same time at the worst case in complexity just to get the maximum information requirement of this box. Inputs like time since last shopping would get on the list but the initial model would not use them.

We all need to remember that we are simulating what is gone for good. "It does not work like the real thing" provokes the answer "the real thing was gone when you were born" for most cases. Now we can drive preserved engines when we are lucky and this is much better than nothing. But from there to what similar engines did in the hands of professionals day in day out around the year at full speed and under full load it quite a different thing.

It is often mentioned that identical engines performed quite different in the hands of different staff. And trouble when feeding different coal, water, grease to these beasts is also well documented on many occasions. Now this can be very interesting to simulate, but only for a very small minority.

Re. engines from different developers behaving different, I'm afraid we need to be a bit honest to ourselves here. There will be two things.

On the one hand, we can dream up the best possible steam engine simulation, not just in theory but coming up with concise SimuGraph setups (noodles). It will be open source and it will be great fun.

On the other hand, DTG and maybe others will have their own ideas of what makes a good, enjoyable, entertaining, challenging, fun-to-drive steamer. This will not deal with fat books and impressive formulas. It will deal with customer feedback. If there is too much whining that a certain train is too slow or too short, going up some serious grade, then that will be changed, no matter what the books say.

Sorry if I am unfair against DTG, but they are a business and people find steamers too complicated anyway and whether the new AI fireman will be 10 times more clever than the RW one will be seen. And many rely on him and still want top performance.

"Top performance" takes us to the final observation: Real train operation is not about extremes. This means that the everyday performance of engines has little to do with the maxima you can derive from a detailed calculation or simulation. More precisely, it is a percentage of the maximum performance but you don't know which. I.e., you have a big unknown factor between the performance found in working time tables and similar sources telling you what the engine was supposed to do and the actual limits of the hardware.

Differences in the behaviour of many a part of the system may be small against this unknown ratio. In other words, much of whole thing can be stabbing in the dark. At the same time, getting to terms with all the parts should rule out some folly that could be observed in the past. Whether this will stop everyone from repeating old mistakes will be seen. The aim of the project should be "finding the Truth", not evangelising the world.

This has become a very interesting discussion.

I think Steam engines illicit more feeling than any other type of engine due to their uniqueness, even among classes. Even now in preservation drivers can often say that engines have "off" days when they just don't want to steam well and can definitely vary in their performance and operation in a single class. More so when they were running full time, at full load and there were dozens of them. I think the "developer chaos" that we currently have is actually a boon to steam locos as it creates the effect that each class does drive differently - and as long as the overall performance is close to reality then I think that difference actually gives a better flavour of driving a steam engine than if they were all drove the same apart from speed/power differences.
This is the joy of steam engines to me: the fact they live and breathe, they have good days and bad days, they reward knowledgable crew and they punish the ignorant. I'm sure some players just want a cookie cutter box that looks like their favourite engine and just goes but I like this uniqueness to be reflected in each DLC, educating the driver on how the engines are really driven and giving a learning curve for the player to achieve harmony with their engine.
This was why I made the 56xx.. because in the early days of Railworks every single steam engine ran exactly the same and had exactly the same controls and operation, they just had different appearances, power and speed. I thought this was very boring and in no way reflected everything I had heard and read about when driving steam engines. We're not creating plastic lumps with a stop/go lever

Boxes in boxes as Andi describes is a good solution to my mind, where you can use default boxes with appropriate input values or you can write your own replacements which you can vary over classes, engines, weather conditions, etc that can introduce or remove procedures to each stage of the simulation and create something that gets as close to replicating the crew experience as we can.

watched a stream by MP last night,he was asked about steam for TSW. he said that nothing was planned and they are a complex beast to do. so I think it will be a fair old time before any are produced. they are (DTG) ironing out any bugs in TSW first and improving on them.

And I absolutely love you work ethos Pete, and what you have achieved! Love to see some of the older ones like the 56 given features of your later releases. But I realise that's a mammoth task and not worth the time invested. Can't wait for the Large prarie's and it's mainly your quality of work that keeps my interest in TS2017.

And will be great to see, as undoubtedly there is so much more scope with the new platform.

Some intersting discussion...

Leaving aside the complications of how to 'do' steam in TSW for a moment, I've been thinking about what would be a good 'steam' route for TSW.

Arguably, I think heritage railways suit TSW best out of all types of route: people have (somewhat rightfully) complained about the short length of the TSW:GWE route, and how we still have the 'end of shift' moment at Reading. The longest heritage railways are the West Somerset and Welsh Highland/Ffestiniog, the former at 22 (or is it 23?) miles, and the latter at 40 miles. All the others are shorter. This means that an entire route can be modelled. Also, the nature of heritage railways mean that a complete (without the missing bits of GWE) timetable can be modelled. Personally, I would suggest the following:

Mid Hants Railway
Length: 10 Miles
Stations: 4
Locos to be included:
Ex-LMS 2MT 2-6-2T (41312)
SR West Country Class 4-6-2 (Previously 'Wadebridge', soon to be 'Swanage')
LSWR/SR S15 4-6-0 (Soon to be 30499, 30506 & 30828)
Coaches to be included:
BR MK1 (Majority of fleet)
SR Bulleid (A couple in service or under overhaul)
Wagons to be included:
Various, including an SR 'Queen Mary' brake van.

Bluebell Railway
Length: 11 Miles
Stations 4 (Excluding possible West Hoathly, Ardingly & Haywards Heath in the future
Locos to be included:
SR S15 4-6-0 (847)
SECR H 0-4-4T (263)
SECR P 0-6-0T (178 & 323 'Bluebell')
Coaches to be included:
Metropolitan Railway 'Chesham' set
SR Maunsell & Bulleid's (Various diagrams
BR MK1 (Various)
Wagons to be included:
Many!

Any other ideas?

All the best,

sem34090

This thread's a bit quiet at the moment...

Any other news relating to steam in TSW? Or has GWE been such a great flop that they've abandoned it? Shame. I knew that if they'd started with steam, and released it in say another two years it would be better...

All the best,

sem34090

sem34090 wrote:This thread's a bit quiet at the moment...

Any other news relating to steam in TSW? Or has GWE been such a great flop that they've abandoned it? Shame. I knew that if they'd started with steam, and released it in say another two years it would be better...

All the best,

sem34090

You should have given them you expert advice!

I doubt they have abandoned it, it is still early days yet. Looking at the difference in detail and the fact that since it was released in February we have had two routes should tell anyone that we are not going to see the same pace of new dlc that we have had with TS!

Peter

Absolutely not a flop (not sure why people keep jumping to that conclusion all the time!) - but why say something if there's nothing concrete to talk about? Talking about half things and possibles, maybes and things that end up not being the same in the final product just creates disappointment, anger and frustration. Best to just stay quiet until it's ready to talk about

Think of it like a duck swimming... calm and serene above the surface, but masses of activity under it.

Matt.

You will know you little special expert board here.

I for one predict total gloom and absolute end-of-the-world because your pour friends did not even model the broad gauge that is so iconic of GWR. How can one claim it is GWR and then come up with that disgusting narrow gauge that was imposed on the work of the only genius by the underlings who could never come up with anything proper, or at least broad?

I mean, if you would have convinced those who run trains under the holy name now that their gauge is wrong, then I would have been impressed. I would even consider buying the sim for the full price then, instead of telling everyone of the importance of waiting for the next sale like some do.

No one wants to drive plastic. Brunel beware! Give me a steamer with open cab where I can watch the fireman break up coal as required. That is what matters. All the uber ladies on the platforms are just for the kids. Make it a real sim about working man (and ladies during war time) or else ......... some non-anorak might buy it in the end.