Page 1 of 5 123 ... LastLast
Results 1 to 10 of 49

Thread: Physics of Spirit of Le Mans DLC

  1. #1
    Vehicle Technical Lead Casey Ringley's Avatar
    Join Date
    Sep 2011
    Location
    Cigar City
    Posts
    308
    Platform
    PC

    Physics of Spirit of Le Mans DLC

    It's June. Time for endurance racing.

    Audi R18 e-tron quattro (2016): Being that the car was retired after 2016, Audi went the extra mile and provided some impressive reference material for this one; something of a rarity for top-level, high-tech cars. It was possible to fill in the few remaining gaps with data from the 2014 car and make a model which I feel is a very good representation of the real thing in our systems.

    Engine is the same 4.0L turbo diesel they have been using for 6 years with constant, little updates for the regulations. By 2016, being in the 6MJ hybrid that meant reduced fuel flow so it would only make about 520hp peak and it holds that from 3500-4250rpm, which is the entire rev range you use. The ICE drives the rear wheels through a 6-speed sequential (down from 7 in previous years) with ratios that, best I can tell, didn't change through the entire season. We have added in a couple alternate final drive ratios in ours to cater to the wider variety of tracks.

    Up front, the hybrid system had a huge upgrade since 2014. Flywheel is gone and replaced with batteries, with the motor upgraded for 350kW peak power (limited by rule to 300kW at Le Mans). Analyzing on-board telemetry shows the energy store works out to a nice, round 1kWh and the system software does some fancy tricks to send an average of 260-270kW to the wheels during each energy burn. The hybrid system normally cuts off around 260km/h even if there is plenty of energy left in storage; getting up to that speed is what makes lap time, not excessive burn for higher top speed. Our hybrid system can't do all the same tricks as the real thing (steering sensors, GPS, etc.) but it was possible to calibrate for a similar result to the real thing while using the full 6MJ over each lap at Le Mans.

    Aero model is similar to our 2014 car data with more efficiency and less drag to hit the right top speed reference points: ~320km/h at Le Mans, 295 @ CotA. There were some rules changes meant to cut LMP1H downforce in 2016, but the teams surely clawed it all back and cornering performance looks close to the mark on the 2014+ downforce levels. We do have the R18 artwork split to Le Mans and high-downforce variants for liveries. Jussi had a great idea to split our aero model between the two, so the low-drag LM version gets steps 0-5 and high-downforce model has steps 4-10. Choose your variant wisely to suit a track and your own driving style.

    Neatest thing about this one was the detail sent over by Audi in the CAD model. First LMP1 car where we could do an exact model of the suspension geometry to see how they handle sending so much power to the front wheels without harming steering feel. There is an almost absurd amount of caster angle built in, and the front would be better described as multi-link than double wishbone. The front and rear bars of what look like wishbones aren't actually connected at all at the upright, and allow the wheel to move fore-aft slightly when steered. It looks like it shouldn't work at all, but it actually drives great and gives better FFB than approximating a pure double wishbone setup from the same links.


    Porsche 919 Hybrid: This and the Toyota were still active during dev time, so information was a bit more secretive on them. It uses a chunk of its base from the 2016 Audi R18 and the differences are largely due to the power unit.

    Engine here is a 2.0L turbo V4 which Porsche claim generates 'just under' 500hp with the 8MJ fuel flow limits. Mapping out fuel flow, estimated thermal efficiency, boost pressure curves, and all that, we can be pretty certain that it holds at that 500hp (or slightly above) from 5,500rpm right up to the shift point we see in on-board telemetry at 7,800rpm. Augmenting this are electric motors on the front axle for an additional 300kW where the typical strategy is to burn hybrid power up to around 280kph and then let the internal combustion take it from there to top speed in the 313-320kph range depending on traffic/drafting effect. Used like this at the beginning of Le Mans' six long straights, it eats up almost exactly the 8MJ energy allowance per lap while shaving upwards of 10s from the average lap time. Impressive stuff considering that top speed on each of those sections doesn't really change whether using hybrid power or not; just that with hybrid power you are at top speed for almost the entire time. Charging happens with a regenerative braking system like the Audi and also adds a generator on the turbocharger which takes extra energy which would normally dump out the wastegate and directs it into the battery at a rate of up to about 40kW. You will see the battery charge increase during full throttle driving thanks to this.


    Toyota TS050 Hybrid: As is usually the case, three years of rules stability saw the main competitors converge on similar approaches in how to get best performance from the energy they were allowed to use. Where 2014 saw a Toyota take pole position with a lap where its top speed was nearly 40km/h slower than that of Audi's on their fastest lap (336 vs 298), by 2016 the three big teams had all honed in on similar aero and hybrid strategies which saw performance over all segments of a track become quite similar.

    Engine is a 2.4L V6 twin turbo compared to Porsche's 2.0L V4, and both make around the same 500hp in a wide band from 5,000-8,000rpm due to fuel flow restrictions of the 8MJ energy class. Main point where they differ is the hybrid system. Porsche run regenerative braking on the front axle plus an MGH unit taking excess exhaust energy from the turbo to charge a battery which powers a motor on the front axle alone. Toyota don't do turbo-compounding, but do take regenerative braking from both axles and then deploy hybrid energy to both axles as well; the only one of the three 2016 cars to send extra power to the rears via the hybrid system. I like very much what this does to the carís handling. In the Audi or Porsche, you can sometimes catch yourself out using the hybrid too early when exiting a turn, spin the front wheels a bit and waste time+energy+cornering power. The Toyota, on the other hand, is only pushing 150kW to the front tires instead of 300kW, so it is a little easier to get on the hybrid boost early without upsetting balance of the car. Does a better job at equalizing tire temperatures and wear (at least for my driving) even if there might be instances where it gives up a little performance when the rears can't handle the full 700hp and it kicks into the traction control. Takes a slightly different driving style than the point & shoot Porsche or Audi, but all three are super close in performance over a lap by the end.


    LM P1H 2016 Hybrid use and strategy: One big difference for these 2016 cars is that we've changed the hybrid system to activate on button press rather than throttle input. Rules for the real car don't allow it to be on a button like this, but the real systems are becoming so complex that we can't really copy their action from throttle input alone; it worked fine for the simpler cars of 2014, but not now that everyone is in the high energy categories and working out better energy deployment strategies. Manually controlling the hybrid to match closer what we see done in the real thing has a significant effect on lap times. Avoiding any wasted energy on unimportant parts of the track and saving it for big burns made me about 4s per lap faster in the Porsche and 2s in the Audi; plus it gives a fun push-to-pass feature if you manage to save a little energy while racing closely with anyone.

    The most effective hybrid strategy in these three cars tends to center around finding the slowest corners of a track and burn energy on the exit up to about 260-275km/h. Using more of the battery than that yields diminishing returns and is better saved for a slow spot elsewhere on the track. Run some practice laps at each track to see how the car recharges over a lap and where you might want to save hybrid energy to have a full boost out of the slowest corners.

    Qualifying can change your strategy a bit too; consider Fuji as an example. In a race at Fuji, you would typically save up charge from the second half of a lap and boost out of the final corner to reach a 290km/h top speed early on the long, Start:Finish straight, doing this consistently lap after lap. In qualifying, however, you donít care about the lap before or after and can use this to your advantage. Exit that last corner and accelerate to about 240km/h, only then using the hybrid for extra top speed of 310+ from the start-finish line into braking for the first corner. Then finish the lap with a long boost right out of the final corner for the best lap time. It doesnít work for multiple laps in a row, but stringing together a fast second half of the straight to start a lap plus fast first half of the straight to end it shaved over 1s from my typical race pace lap time.

    A similar quirk comes into play at Spa-Francorchamps. Rather than burn all of your energy our of La Source before Eau Rouge, save 50% charge or so for the Kemmel straight afterwards. It is a long, uphill run, and boosting to top speed immediately after Eau Rouge can be a huge win for your lap time. Every track will have unique strategy plays like this, so experiment and keep an eye out for what works best in both qualifying and race situations.


    Porsche 924 Carrera GTP: This one is a funny little piece of history thanks to those three letters at the end of the name - GTP. Porsche had introduced a 924 Turbo model but there was not enough time to meet homologation requirements for it to run in the production GT category as intended. Simple solution: Just run it in the GTP class against other full-on prototypes where the rules were effectively 'have a roof and a minimum weight for your engine size'...300hp disadvantage and road car aerodynamics be damned, the 924 Turbo would be racing!

    There is a fantastic documentary video of the car's restoration over HERE where the drivers all talk about how great 924 GTP was; having some of the best handling they ever experienced...high praise when the group includes five-time Le Mans winner Derek Bell. And that great handling paid off as the race was *very* wet. Despite being a good 50mph down in top speed and lapping 30s slower in qualifying, the 924 GTPs used their strengths to pull off a fantastic result of 6th overall (3rd in class) with the other two cars in 12th and 13th.

    Design-wise, the car is both very similar and very different to the 911 Carrera RSR 2.8 we have in game from the Porsche Legends DLC. The differences, obviously, are that this one is front-engine with a 2.0L turbo 4-cylinder sourced from Audi. Doubling boost pressure from the street 924 Turbo to 2.5bar, it cranks out 320hp with a smooth torque curve and power is fed through a 5-speed manual Getrag G31 much like the road car but with a wide range of ratios to choose from for race use. A clutch-type limited-slip differential with symmetrical 40% lock is standard issue, much like in the 911 Carrera RSR and even 908/03.

    Engine position, roll cage addition, and general lightening of the chassis make for a total car weight of 945kg with 52% rear balance. Suspension uses struts at the front and trailing arms at the rear, like the 911 Carrera RSR; spring and damper settings for that car also work here to great effect. It even used brakes from the 917 parts catalog, another 911 Carrera RSR similarity. For tires, they fit the widest things possible. This car is around 100kg lighter than the BMW M1 Procar and with 160hp less, but it has more rubber at both ends; thing has a surplus of grip in the dry and wet. While it can struggle on top speed tracks, it is so easy to chuck around that time can be made up quickly in twisty sections of tracks and it has come out ahead in testing here at the shorter tracks.


    Porsche 961: A real unicorn here; they only made the one. While Group B and the 959 are largely remembered for rally racing, circuit racing was always part of the plan and the 961 was to be Porsche's customer car in that regard. The end of the Group B era and cost of a 961 compared to competitive Group C machinery cut those plans short and left us with only one example of what was, for nearly 30 years, the only AWD car to race at Le Mans.

    Chassis construction was not all that different than the road-going 959S model from our Porsche Legends DLC. The usual strengthening and lightening measures were undertaken; the highly-computerized AWD system was simplified to a fixed 20:80 power balance front to rear; the 'zero-lift' aerodynamics were bolstered to produce some downforce for the necessary cornering grip at expense of a higher drag coefficient; and the 959 engine, which was in turn a de-tuned 956/962C race engine, was swapped back out for a Group C racing version of the 2.85L twin-turbo flat-6.

    The race-spec engine brings power up to 680hp @ 7,800rpm and drives through a 6-speed gearbox just like from the 959S but with a range of ratios as setup options. For Le Mans, they ran the rear end with a spool axle as a reliability measure (fewer moving parts to break) but you can be sure it would have run a clutch-type LSD at the rear had it raced in a full WSC season; ours does that with a typical 40/60%-lock differential at the rear.

    Wider fenders, mild diffusers on the flat floor, and an aggressive rear spoiler cost about 20% in drag increase over the 959S for a gain of roughly 500lb downforce at 150mph. Not a huge amount, but very welcome to help stabilize the 1150kg car in cornering and under braking. The extra drag limits top speed to 205mph on the Mulsanne straight; about the same as the 959S road car despite the extra power and some way off the top Group C cars it raced against, but still a respectable number.

    Tuning the suspension proved to be not a difficult job at all. The 959S had already forgone the fancy, adaptive, computerized suspension of the 959 Komfort model in favor of fixed-rate dampers and ride height. Turns out that, being some 400kg lighter, the 961 races extremely well on the same springs and dampers of the 959S. Drop the ride height, stiffen the anti-roll bars to account for increased cornering grip from the racing slicks and it is ready to go. Probably my favorite handling car in its class. Maybe not as exciting as the F40LM's turbo lag or Mustang's pure torque, but it inspires huge confidence on corner exits where you can plant the throttle to the floor and let the AWD system pull the car out with just the right amount of slip angle. The Audi 90 IMSA GTO finally has something that can compete with it in the rain.


    Porsche 917LH: The 917LH models are very similar to the 917K at fundamental levels. The main evolution here is in the aerodynamics with a heavy focus on top speed for Le Mans. Early iterations of the LH bodywork were great for drag reduction, but also generated significant aero lift at the rear. That deficit had largely been removed by the 1971 model weíve simulated here; wind tunnel tests showed about 300lb downforce @ 150mph with a stable balance of 20-25% front. Thatís enough to hit a solid 390km/h and feel perfectly comfortable taking the Mulsanne kink flat out.

    The longer bodywork plus other detail changes of the LH model pushed them up above 820kg and closer to 66% rear weight.

    See our Physics of Porsche Legends DLC thread for more notes about the common engine and gearbox used by these cars.

    http://forum.projectcarsgame.com/sho...he-Legends-DLC


    Ferrari 512M & 512S Coda Lunga: In researching this one, the word most often used for it by drivers, owners, and Ferrari's engineering team was 'under-developed'. I suppose that is a fair criticism considering it only took a single race win to the Porsche 917's 17 in the 1970 and 1971 seasons, but you also have to consider that the 917 began life a year earlier in 1969 and benefits from an extra year's worth of development at a time all aspects of car design, aerodynamics chiefly, were moving at a extraordinary pace. That Ferrari were able to win Sebring 1970 in a 512S variant right at the start of its life shows how strong a platform they had.

    The technical directive for 512 was essentially to use as much existing material Ferrari had to build a 917 beater. For this, the engine took castings and design cues from the recent 6.9L 712 Can-Am engine with bore and stroke reduced for a capacity of 4,994cc. It had a slight power advantage over the Porsche flat-12 (610 vs 600hp) and, being a known and tested architecture, was generally reliable and less sensitive to over-revving than the Porsche. Peak power comes between 8,700-9,000rpm and it is happy spinning up to 9,600 if you get favorable winds or a nice draft down the Mulsanne. Drive goes to the rear through a simple, 5-speed gearbox designed in-house and a ramp & plate clutch differential; the similar layout and weight balance means a 75% locking factor works well here as it does in the 917 (some drivers in period preferred using a spool axle). Gear ratio setup was distilled to four crown wheel & pinion options to adjust final drive for top speed of the circuit; the longest, 11/35 (3.182:1), being the choice for Le Mans and getting the Coda Lunga variant up to 230mph.

    Tube-frame chassis for the 512 was essentially carryover from the 612 P Can-Am car. The basic concept and design was standard Ferrari construction of the time (much like how the 908 and 917 variants evolved from the same base) and it all works well. Suspension geometry produces benign handling with nice steering feel, letting you focus on placing the car and putting power to the road. Reading old race reports, it seems the main issue they had was one of quality control and inconsistency where not all 512 chassis were build exactly to the same standard. Teams which had the most success, namely Penske, put a lot of work into extra prep of their cars to make sure the design worked as intended.

    The chassis did have a significant weight disadvantage to the Porsche; tube steel fabrication making for a frame maybe 100kg more than the 917's aluminum chassis. Later 512M examples have a claimed weight of 815kg, but this appears to be a dry weight while true, Le Mans scrutineering weights from 1971 put the cars more in the range of 880kg with fluids included. The 512 S Coda Lunga models weighed in at an average of 940kg.

    Aerodynamics is a big area where the 512 suffered from lack of development. It never saw a wind tunnel (early versions of the 512S even had little road testing because of an unfavorable winter in Sicily) and was drawn largely from past experience about aero ideas which should work in theory. Huge difference to the extensive wind tunnel program of the 917 models, and it showed at Le Mans where the Ferraris typically had between a 6-10mph deficit on similar, long-tail bodywork. Short-tail 512M models will top out around 217mph where the 917LH reaches 240. The 512 S Coda Lunga is good for another 10-15mph, but setting up for that comes with a handling penalty of losing maybe 50% of rear downforce compared to the 512M, which itself is some way behind the 917 for downforce numbers. The Coda Lunga gets very light going over the Mulsanne hump at speed, so be careful with your steering inputs when taking that crest at 230mph.

    That all sounds like it adds up to a car which will just be smoked by the Porsche at every track in game, but truth is they proved quite well balanced in testing during our development. For one thing, the Ferrari ran wider tires at both ends, especially the front (285mm wide vs. 235mm), and that extra grip helps counter the aero deficiencies. Then there is that engine. The air-cooled Porsche flat-12 was a tremendous machine, but it was also a complex design and Porsche's own data shows that it was quite peaky in power production. Its peak of 600hp is only held for a few hundred rpm; go outside of that range and it drops significantly, change up a gear and it might dip below 500hp for an instant. That Ferrari V12, though, was a known design with which they had lots of experience. It's torque curve is much smoother and holds it over 550hp (90% of the peak 610hp) all the way from 7,500 to the limit at 9,600rpm, letting it pull much harder after each gear change. There may be a lower top speed, but it gets there more quickly and that is where you can find a lot of performance.
    Haste Is Slow

  2. #2
    Superkart Pilot
    Join Date
    May 2015
    Posts
    369
    Platform
    PS4
    Excellent notes as always and thanks for explaining the best way to use the hybrid systems in the 2016 LMP1 cars.
    ...You've seen how I react to people, make them feel good, make them think that anything's possible. If I make them laugh along the way, sue me. And I don't do it so they turn round and go 'Thank you David for the opportunity, thank you for the wisdom, thank you for the laughs.' I do it so, one day, someone will go 'There goes David Brent. I must remember to thank him.'
    The following 6 users likes this Post: AbeWoz, APR193, blakflag, davekojo, ShneebnaMRR108, Tar Heel


  3. #3
    Superkart Pilot
    Join Date
    Oct 2017
    Posts
    106
    Platform
    PC
    I'd like to take this moment and go ahead and initiate a slow clap for the devs on this one 👏👏👏👏. Amazing work guys.
    The following 4 users likes this Post: AbeWoz, Plato99, r18etron, Tar Heel


  4. #4
    WMD Member Christiaan van Beilen's Avatar
    Join Date
    Oct 2011
    Location
    Groningen
    Posts
    463
    @Casey:

    I just drove the 924 GTP and I noticed that, at least for me as a Vive user, the outside mirrors are way off in their adjustment (left mirror having to go +100 horizontal). I also had this with the Lotus 72D.

    Could this be an issue on my end or could it be that in the entire car catalog (not just DLC) there are cars with poorly setup mirrors? Some even having to be adjusted by a crazy amount like the above left mirror (which was the right mirror on the 72D if I'm not mistaken).


    That said... I love your walls of text like this. I've only just read the parts I was most interested in for now but I know I'll be back for a proper read later tonight.
    System: Intel i7 2600k @ 4.2GHz - Noctua NH-D14 - MSI Z68A-GD65 (G3) - Vengeance LP 16GB 1600mhz - MSI GTX 970 Gaming 4G @ 1114Mhz - AverMedia Live Gamer HD (1080p@30hz capture card) - OCZ Vertex 4 256GB (FW: 1.5) - Creative Soundblaster X-Fi titanium champ - Logitech Z-560 4.1 - Philips 32PFL9604H/12 32inch TV

    Input: Logitech Wingman Formula Force - Logitech G25 - Thrustmaster T500 RS - Thrustmaster TH8 RS - Fanatec ClubSport Wheel - Fanatec ClubSport Pedals V1 & V2 - Fanatec ClubSport Shifter SQ - Fanatec ClubSport BMW M3 GT2 rim - Fanatec ClubSport Formula rim (*italic = in use)

  5. #5
    GT5 Pilot Jezza819's Avatar
    Join Date
    May 2015
    Location
    Alabama
    Posts
    1,039
    Platform
    XBOX1 X
    I'm definitely going to have to learn how to use the hybrid boost if I'm going to drive these cars a lot. Last night I finished dead last in both the Porsche and Audi Fuji at Silverstone GP against a 23 AI car field at 65 strength in a 10 lap race. Both cars finished 40+ seconds behind the winner and my fastest lap was 3 seconds slower than the AI. I boosted on the two straights but nowhere else but I still got smoked.
    Xbox 1 X - Fanatec CSW V2 wheel - Fanatec V3 pedals

  6. #6
    GT5 Pilot RoccoTTS's Avatar
    Join Date
    May 2015
    Location
    Circuit-Zolder
    Posts
    1,284
    Platform
    PC
    I only tried 1 car so far, the 924 GTP and i have to say it's already one of my favorite cars. It's a blast to drive.

    And a big thanks for this post, always nice to read.
    MSI TRIDENT-014EU specs : Intel Core i7-6700 3.4GHz | NVIDIA GeForce GTX1060 3GB | RAM 8GB DDR4 | Windows 10 64 bit
    Thrustmaster TX base | F599XX EVO 30 Alcantara rim | T3PA pedals | Playseat Evolution Alcantara | Panasonic 50" 4K TV

  7. #7
    Kart Driver TX3 Slawek's Avatar
    Join Date
    Jan 2018
    Location
    Clavier (Huy)
    Posts
    29
    Platform
    PC
    thank you for the information
    Manager Virtualdrivers by TX3 Project CARS PC
    www.virtualdrivers.org

  8. #8
    WMD Member Sessionerror's Avatar
    Join Date
    May 2015
    Location
    Stuttgart, Germany
    Posts
    393
    Platform
    PS4
    Excellent read, as always Casey. Thank you!

    Regarding the hybrid system of the LMP1 (2016), I just discovered something odd, though: Pushing the KERS button without throttle input won't result in acceleration. However, the battery load is reducing while the KERS button is being pressed. Since the shift times of the Toyota TS050 seem to be significantly longer than the shift times of the TS040, this results in wasting quite an amount of electric energy: While I accelerate and push the KERS button, I do have to shift of course - during shifts though, the throttle is being cut. In this time, I still press the KERS button of course, but the energy is just being wasted.

    Is there a possibility to do something about that? Either by cutting the KERS input as well, or - even better - getting some acceleration from the electric drivetrain alone?
    a.k.a. "Chief fast sock" // German Project CARS Community: www.pcars-forum.de

  9. #9
    Superkart Pilot
    Join Date
    Sep 2017
    Posts
    286
    Platform
    PC
    You're write up just made my work day seem a little longer. This has been the DLC I've been most excited for. Cant wait to hit the track when I get home!
    Ryzen 1700 @3.8ghz|16GBs RAM @2933mhz|EVGA SC2 Hybrid 1080ti|1TB Samsung 960 EVO|250GB Samsung SSD|480GB SanDisk SSD
    1TB WD HDD|750w 80+ gold EVGA G3 PSU|Dell 27" 1440p 144hz G-Sync monitor|Thrustmaster TX| 599XX Evo Alcantara|TP3A Pedals|TH8A Shifter|Fanatec Handbrake

  10. #10
    Superkart Pilot
    Join Date
    May 2016
    Location
    Tilburg
    Posts
    147
    Platform
    PC
    Thanks but I will read later on. I have a very good excuse: first try out the (classic) cars on LeMans myself. Did I already tell you that I'm in love again? That Ferrari 512 is so beautiful and so great to drive. Tears in my eyes of joy. In VR it is stunning with a lot of subtle details (like the rear view mirror in VR)
    PC: Omen i7-1080ti, OculusCV1, 43" 4K screen and Optoma 1080p beamer
    NextLevelRacing GT Ultimate v2 & V3 motion platform + Fanatec CSL Elite & LC pedals, TH8A shifter

    PS4Pro, T300GTE, T3PA, 599xx wheel / F1 rim

Similar Threads

  1. Spirit of Le Mans
    By g.stew in forum General Discussion
    Replies: 127
    Last Post: 16-06-2018, 00:02
  2. When's The Steam Release of Spirit of LeMans?
    By slider916 in forum General Discussion
    Replies: 47
    Last Post: 05-06-2018, 17:51
  3. Replies: 2
    Last Post: 29-09-2017, 16:14
  4. Virtual Le Mans vs. real Le Mans
    By JLM_1988 in forum eSports
    Replies: 1
    Last Post: 14-02-2016, 12:23

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •