Engine simulation

Hi all,
Just wanted to know few things about engine simulation in rfactor 2.
I was doing MIN vs MAX radiators test (both water and oil) to see what difference it makes on torque and HP.
to my surprise it makes no difference at all, which i find strange. espcially when i was conducting the test on fpro....from what i heard F1 style engine are quite sensitive to temperature, so they should have narrow window of operation.
here is a picture of my test. (White one is MAX radiators i.e highest size) >



as you can see temp diff is decently big for both water and oil.
so does water and oil temperature have any effect on engine performance ?

 

60 degrees VS 0 degrees > (white one is 0 degrees run).

 

@Simulation_Player have a good look at one of the engine.ini files, and see what you can find. You'll see nothing defining power/torque vs temperature.

I've seen speculation that there is some hardcoded effect, but I haven't seen it personally.

Remember what I said about losing the magic of games? This is the sort of thing I'm talking about. People will tell you, with conviction, that something is happening, but sometimes (often?) when you dig in and investigate, actually there's nothing.

There is an optimum temperature, and in the past I think that has helped to define the 'thermostat' operation, but that's all changing a bit now as well with recent additions.

 

I understand your magic point better now , but honestly this the kind of basic effect you would expect to be simulated.(even if it is simplified)
so there was nothing "magical" about this specific effect because it is a very common and basic engine behaviour that even a amateur would know.
I will take a look at engine.ini file , thanks for suggestion.
But honestly i'm already quite disappointed with the above results.

 

Well, I would guess that engine engineers would never allows a radiator too small to grant optimal performance or too big for the intended operating window. So this could explain why the engine still work ok.
It's just like when people believe that running on minimum pressure would make the car run on rims.

 

That is fair point for my first post only but what about massive temperature difference in my 2nd post having zero effect on HP & torque ?
That is just rf2 lacking imho.

 

For sure it is, the "problem" with RF2 is that when it was a ISI creature, the developer made it clear: feature will be implemented only if can be accurately simulated. They put in place almost everything that was needed along the road, but failed to implement it all, I think both because it was a small team, and because this aproach was a bit too much draconian. Now Studio 397 showed us that they have the means and the will to fill those gaps, but you have to understand that to correctly simulate the performance loss you talk about, you need specific data from manufacturers and teams, a data that is probably quite hard to achieve or to share. So, in the end, there is a choice to do: ignore it because 99% of the times it does not apply, or implement it, somehow, guesstimating how to apply it to all cars in game.
Rf2 is full of shortcomings if you search them close enough, but you have to remember that is just a videogame you can buy for few bucks and run on a medium performance PC with decent results, it will never be 100% simulated or 100% accurated, otherwise we would not call it simulation anymore.

 

I understand that no sim is 100%, but this feature seems fundamental and didn't really took me any efforts to find it , i only hope that a sim cannot be picked apart so easily by someone like me with no real education of this field (apart from reading stuff on internet).
Surely real data will be hard to find....but honestly i will take a good estimation everyday instead of nothing at all. (atleast in this case).

New RC has really shown what s397 is capable of , also because of this new RC has expanded bit on engine physics this is why it is bit let down to find this results.

"

• Implemented enhanced cooling system for oil & water with new parameters in the .hdv file The new model includes:
  • Direct heat transfer between water and oil
  • Oil radiator
  • Switchover Thermostat "

After reading this i decided to test the above, because in previous builds i saw no difference in temp VS performance.....so i thought maybe they added this in RC.

again it don't need to be 100% accurate....a good appox will do just fine, that is ofcourse if S397 is confident that they can do it much more hardcore....i'm willing to wait for it.

 

Uhm just for the sake of funny and constructive talk, would you please list all the factors that have a weight in this calculation?
What I mean, is that once you touch something there is not a simple and easy way to approximate something without a consistent pool of data. I understand nothing on the subject, but I guess you need a consistent number of variables di simulate this effect even vaguely realistically: from pump flow, heat transfer rate of oil water air and several other solid materials, air temperature, moisture, all the masses involved : fluids, engine, radiators, and once we are at it, we need to differentiate from fresh still air and hot turbolent gases from the car ahead, that we know have devastating effects on tires, aero, engine and brakes. Now stuff are already getting a "bit" complicated and maybe it would be worth investigating it in a deep way, but the question is, in % how much power can be loss running the engine outside his project values? If we are talking 5% will it be worth the effort? I would really like to have more and more features modeled in RF2, but I'm the kind that take what I have and do not worry much with what I haven't.

 

So by what mechanism in rF2 do engines overheat? Because overheat = loss of power. I assume that it's not based on data but just somekind of calculation that only takes in a limited amount of factors. Not sure if what I'm asking is relevant to the conversation tbh.

 

That is probably easy: the engine produce a X amount of heat depending the fuel flow, efficiency of combustion engines is pretty known, so it's easy to assign a credible value, then radiator dissipate Y amount of heat multiplied by air speed. Old F1 mods overheat if you stand still, and I even believe that maximum power is affected in those cars. Crude but effective to portray those cars that cannot stand still with running engines for long times.
Probably a subpar solution for a simulation, but the lore was there.

 

similarly why can't we have a simple map,
like this >



No need to consider all the elements that make this effect, we just need end result NOT actual engine simulation.(altho that will be cool but probably useless for rf2 case since it is not a engine design sim).

im sure a good enough data is out there or some manufacturer might give rough estimations or devs can reverse engineer etc. After all Assetto corsa competizione managed to simulated air temp, humidity and altitude affects on engine for massive collection of real and latest cars......so doesn't seem impossible.
I just hate it when i have to add yet another thing in rf2 car setup which basically doesn't work.....so yet another value best set to minimum.
ARBs , camber , tyre pressures , rear wing angle (on some cars) and now radiator cooling........... the list keeps piling up.

 

When we were working on the A1GP mod for rF1, to simulate the Mexico City round we built an entire collection of duplicate engine and aero parts, with the engine detuned something like 5% to account for the high altitude. We also reduced the aero efficiency, again because of high altitude. The calendar in rF1 made it rather easy to only call those upgrades at Mexico City. Another missing feature in rF2.

 

Then, you'll brake your engine.
Take a look to this text from ISI:

Code:

LifetimeEngineRPM=(17730.0,344.0) // (base engine speed for lifetime, range where lifetime is halved)
LifetimeOilTemp=(125.4,2.90) // (base oil temp for lifetime, range where lifetime is halved)
LifetimeAvg=8105 // average lifetime in seconds
LifetimeVar=2940 // lifetime random variance

The LifetimeAvg and LifetimeVar are ... the mean and variance of a normal distribution.

The lifetime of the engine is continually reduced based on the current RPM and temperature. If you were able to drive around perfectly maintaining an RPM of 17730.0 and oil temp of 125.4 deg C, the lifetime would reduce by 1 every second. So on average, you'd have 8105 seconds before your engine exploded.

Now obviously you can't maintain that exact RPM & temp, so the lifetime reduction happens faster if you're above the RPM & temp and slower if you're below. It's an exponential curve - at 17730.0 + 344.0 RPM and 125.4 + 2.90 deg C, the lifetime reduces by 2 every second. At 17730.0 + (2 * 344) and 125.4 + ( 2 * 2.90) deg C, the lifetime reduces by 4 every second, etc. The RPM and temperature each contribute half to the calculated lifetime reduction

 

oh yes aero gets affected by altitude and air temp (density) , don't think rf2 has that too.
tbh seems like the feature u talk about in RF1 gets job done but will probably get too tedious to manage for vast selection of official cars and tracks.

 

yes im aware of the engine damage itself but for the short duration races that won't matter, similarly for qualifying....since no power penalty from running engine hotter.
I don't want car setup to follow realistic approach ONLY for specific conditions (like long races in this radiator's case) but the car setup should follow realistic route in every condition.

Edit : in my theory , the drag reduction advantage by smaller radiator is so small that on most circuits you would rather have more power and torque (and long lasting) engine for that acceleration advantage on shorter acceleration zones/sections of circuit.

 

The problem with alot of the rF1 "features" was that it was a hack job and not really fleshed out and not working as it should. There were workarounds used for turbo engines, because turbo support wasn't there for rF1, yet some of the most popular mods featured cars with turbo engines. There was rain hacked into rF1, yet it didn't really work. The thing with stuff that - as you call it - get's the job done is that it needs to work out of the box with no further interaction of the user and there is actually a quite long list with stuff that seems easy on paper but requires a bucketload of work. I think people need to adjust their expectations a bit when it comes to features and how they work in our favorite racing sims. Whenever there is a discussion about rF2 on RD there is a bunch of guys telling how advanced AC is as they have hacked the crap out of it and they pull out a chart that shows a correlation between their custom tire and a real tire. The question is: do those new physic features work across the board without requiring the end user to do anything and are they networkable or do they have any understanding of what correlation means and how statistics work? And this is where the discussion usualy stops.

With todays racing sims it's fundamental, that all the new features work flawless based on a number of tests, that you don't have to use notepad or any other 3rd party app and that they are networkable. And there are tons of features that are existent in some games that aren't there in some others, that look very simple and basic on paper. AC doesn't even have any form of radiator cooling, nor proper brake simulation and noone gives a flinging f*ck. And yet, it's the most popular sim game on Steam. It's the least suitable sim for endurance racing, yet people keep releasing enducance content for it. Sometimes the world is wierd. rF2 has quite a few features that are still more advanced compared to ACC and it seems that this stuff somehow get's lost quite often.

 

Yep you right about AC1 , it is quite lackluster in many areas. ACC too has some holes.
my point of stating ACC in comment is just to show that the effect regarding radiator can be done.
my philosophy is simple, looking at competitors you pick their strong points for betterment of your product.
while using their shortcoming relative to yours as a selling point for your product.
funnily enough i mention ACC simulation qualities more yet here i'm playing RF2, probably because rf2 IMO doesn't "boast" their qualities as much as kunos do so i don't know what rf2 is capable of under surface without doing some research, its not "easily" available as kunos's blogs. I could go into moddding sections (which ACC lacks completely) to learn tho, so thats lazy and clumsy on my part.
At the end of the day my whining is only because i think rf2 is "correct" feeling to drive, hence I would love to have some missing features to truly utilize its potential.

 

I get what you mean and I think it's a good attitude in life in general to follow that logic. The big problem with this is that what seems as a good benchmark for one reason or another might be not the right benchmark. ACC and rF2, I think I am repeating myself here, are completely different products with completely different scales, development cycles and concepts. One is a product developed for a very specific type of car with a small range of customizability and options following very specific regulations - there are two types of tires for the GT3 cars for example in ACC, while that type of car offers four different compounds within rF2. Now incase, S397 starts tweaking the tire model or making changes to the physics model when it comes to environmental effects like track temps, makes the workload two times bigger. The recent RC for rF2 brought quite a few nice updates, but when I checked Imola from Reiza for example, I noticed that there isn't something quite right - again the issue with backwards compatibility. Same goes for the sound engine. An open product like rF2 is allways harder to maintain and develop than a closed platform like ACC. And that's an issue that you won't overcome unless you close the paltform.

At this point I would rather see S397 flesh out the content they have instead of allways pouring in new features, as nice as they are. The P2P system is very nice, but it needs to be ported to the cars that feature it, like the DW12 and the Class 1 BMW. And by adding new features like in the changelog the pile of WIP stuff get's bigger and bigger. Anyway, this isn't meant as a rant or the advice to not add new features, but much more of an observation.

 

I disagree little bit on feature point.
P2P and weight jackers are nice features but imo problem is that they don't apply to majority of content.
I would like to have a base features in future that benefits majority of cars.
this will obviously take more time to update all content, so imo S397 can safely avoid updating older content or just content that are not as popular and ofcourse start with paid content first.
But new content is always welcome
man i think i'm greedy .
They definitely need to strike a nice balance between working on features or new content.
I personally think new features add much more value BUT new F1 car ! that would probably get more hype.