Leading-Edge Motorsport Technology Since 1990

Formula 1 Technical review 2018

F1 Supp_Cover-ac.indd 1 17/12/2018 11:52 30 & 31 October 2019 | NEC, Birmingham

The UK’s largest annual gathering of OEMs and engineering supply chain professionals

91% 88% 98% of exhibitors met their of exhibitors have either of exhibitors were satisfied objectives of networking booked or will be booking with their experience at with new and existing clients again for 2019 Advanced Engineering 2018

Secure your stand today

“We have done this for the 5th “I found the event a great year in a row and it is a very networking opportunity to meet good event for us to meet new industrial professionals from clients.” different backgrounds with Thomas Evans, Business different products.” Development Manager, Kat Clarke, Wing Manufacturing EBS Automation Ltd Engineer, Airbus

Secure your stand now www.advancedengineeringuk.com

AE_2019Advert_versions2.indd 1 13/12/2018 11:19:51 CONTENTS

ercedes dominated the 2018 This year also saw the introduction of the 4 Unintended consequences Formula 1 season, wrapping up Halo head protection system, which has changed How an attempt to curb aero the constructors’ title as well as the the ethos of the world’s premier single seat development has had a dramatic drivers’ title for Lewis Hamilton. But, series. Open cockpits have always been at the eff ect on CFD in Formula 1 in 2018 Mthis was not a walkover by any stretch. The team heart of open wheel racing, but the Halo has 14 Pressure sensors believed, as did many observers, that Ferrari had fundamentally changed that. Simon McBeath and the latest the better car in the middle of the season, and The new mounting hardware, and meeting advances in scanning technology was unable to convert that advantage into a title. some incredible crash testing targets, means 20 Halo protection We are now coming to the end of a rule-set weight. This has led to a technical challenge for Gemma Hatton on what makes the and fi nally we have the 2019 regulations to the Formula 1 teams, one that no doubt will be Halo such a capable safety device hand. There have been steps taken to improve refi ned in the 2019 racecars. 28 Safety drive overtaking, to increase the life of engines, and Further safety features will come next year, in New helmet design for 2019 revealed to reduce the array of tyre choices that are the form of new helmet design. Head protection as the FIA steps up safety campaign available for each race, while retaining the spirit has been a focal point of recent FIA investigations 34 Big wheels keep turning of competition. But this is not the end of the and only now are we starting to see the results Why Formula 1 has fi nally taken the story. As we discovered this year, the regulations that will benefi t drivers in the near future. On a plunge and gone for 18-inch wheels surrounding the computing restrictions need lighter note, the cars will be fi tted with 18-inch 42 Cool runnings work, and since we ran the article (see page 4) the wheels in future, which will have an extraordinary Sam Collins uses Caterham’s 2014 FIA has announced a change, but not one that we impact on suspension and chassis design. data to explain F1 cooling tech expected. While the computing power limitations One shadow on the horizon is the news meant a waste of money, the FIA has opened up that Porsche will not be coming into Formula 1. those restrictions to ensure more data comes Produced by Racecar Engineering team This lack of new blood is worrying; unless new its way, and therefore it will be in a position to manufacturers join, Formula 1 must survive do a more competent job on the next set of with the power unit suppliers currently in play regulations, due out in 2021. (Renault, Ferrari and Mercedes). This puts them in It was a smart move, but the costs are within a very strong negotiating position, as the loss of the reach of only the wealthiest of teams. In this one of them would have far-reaching impact on supplement, we also look at pressure sensors, and the remaining manufacturers. how they can impact the rate of development and ANDREW COTTON the lessons learned in testing. Editor

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 3

F1 Supp_Contents_MBGHAC.indd 3 17/12/2018 15:25 TECHNOLOGY – AERODYNAMICS Know your limits

This year Formula 1’s CFD restriction regime has been shaken up big time as the FIA looks to cut the costs of aerodynamic development. But has it worked, and how has it changed both the tools and the process? Racecar investigates By GEMMA HATTON

n January 2018, the FIA introduced the latest evolution of aerodynamic testing restrictions for Formula 1, and with them came the biggest change in CFD restrictions since they were first introduced back in 2009. Racecar Iwent behind the scenes with HPC specialist, Boston Ltd, to discover the impact of these changes and how Formula 1 teams have not only benchmarked new solutions, but also upgraded their CFD supercomputers. But to put these latest changes into context we need to understand the history of the restrictions, both for CFD and the wind tunnel. In 2008, aerodynamic testing was at its peak. BMW Sauber, Honda, Williams and Toyota had all invested huge sums of money in new state of the art full size wind tunnels, each costing tens of millions of pounds. All the top teams were operating in two wind tunnels simultaneously, while Toyota was not only using two wind tunnels 24/7, but each of these was full size. However, the vast majority of this wind tunnel testing utilised scale models, and over the years the scale of these models increased from 40 per cent to 50 per cent and then 60 per cent. Operating two wind tunnels full time allowed these teams to complete around 500 wind tunnel simulations per week, with each simulation incorporating approximately

20 different car attitudes. Full size wind tunnel testing was LTD UNIFI commonplace, with teams either using their own facility or a Boston worked together with UniFi and CE to benchmark the performance of new CFD technologies in customer facility such as Windshear in the USA. accordance with the 2018 regulations to see if F1 teams would be forced to upgrade their CFD capability

4 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 4 14/12/2018 11:13 It quickly became clear that something had to be done to curb the growth of aerodynamic testing in F1, and its associated costs

NOVEMBER 2018 www.racecar-engineering.com 5

RCE F1 digi mag.indb 5 14/12/2018 11:13 TECHNOLOGY – AERODYNAMICS

In 2008 teams were already using Between 2009 and 2017 the regulations evolved wind tunnel. For example, let’s assume that CFD CFD routinely as part of the aerodynamic and generally served to reduce the aerodynamic capacity allows a maximum of 12.5 TeraFLOPs. development process, and as the software resources available to the Formula 1 teams, Using the equation with the 2013 limits results and correlation improved while hardware particularly in the wind tunnel. This was done in 41.3 hours of wind on time, as shown by the costs reduced, teams began to use it more, through introducing a ‘limit line’ which is green square. In 2015, however, 12.5 TeraFLOPs integrating it further into the design cycle. defi ned by the following equation. would only give you 12.5 hours in the wind At that time, BMW Sauber was leading tunnel (blue square) – that’s 70 per cent less the way in CFD hardware with the Albert 3 than 2013. The exact balance between CFD and supercomputer and over 4000 Intel cores, but wind tunnel resources varies from team to team, other leading teams were not far behind. It Where: and sometimes from year to year, depending quickly became clear that something had to be WT = Wind on time on the strategic approach and technology done to curb the growth of aerodynamic testing WT_limit = 25 hours advances adopted by each team. in Formula 1, and its associated costs. CFD = TeraFLOPs usage Of course, every restriction that is The fi rst step came into force in January CFD_limit = 25 TeraFLOPs introduced simply triggers the teams to 2009 as part of the FOTA Resource Restriction exploit the loopholes and optimise their Agreement (RRA). This controlled the Therefore, the amount of time a team chose to designs and working practices to maximise aerodynamic resources the Formula 1 teams run its CFD directly dictated how much time it their performance from the regulations. For could deploy via restrictions on the wind tunnel could utilise the wind tunnel. Equally, if a team the TFLOPS CFD restrictions, this became ‘wind on time’ (WON) and the CFD compute could complete its maximum allocation of wind an arms race as teams pushed to develop capacity, measured in TeraFLOPS (TFLOPS). tunnel runs using less wind on time then it their supercomputers to run the most CFD Wind on time was simply a measure of would have more capacity for CFD simulations. simulations per given TFLOP allowance. This led the amount of time the fan was turned on in teams to operate CFD hardware in ways which the wind tunnel with the wind speed in the Working area were quite diff erent from the wider industry, test section above 15m/s. For CFD, TFLOPS Looking at the WT_limit and CFD_limit data with a clear focus on regulatory effi ciency was eff ectively the number of fl oating point from the last few years, Figure 1 can be created. rather than fi nancial effi ciency. For example, operations completed within the designated Essentially, by plotting the maximum of each of the TFLOPS calculation naturally includes a chip eight week Aerodynamic Testing Period (ATP) these limits, you can establish the ‘working area’ clock speed term which is reported either as the and was defi ned by the following equation: that the teams could operate in. For example, maximum turbo clock frequency stated on the in 2013, when the maximum WT_limit was 60 CPU specifi cation (if the turbo mode is used), hours and the maximum CFD_limit was 40 or the base clock frequency if the turbo mode TeraFLOPs, the team could operate anywhere is not used. Teams quickly established that the Where: within the green shaded area. In 2014, the turbo mode was not an effi cient way to run TotFLOPs = Total number of TeraFLOPs used per second limits were 30 hours WT and 30 CFD TeraFLOPs, CFD simulations, in terms of the number of CFD MFPPC = Peak double precision fl oating point operations illustrated by the red shaded area, whilst simulations completed per TFLOP. This was also per cycle per core of the processing unit 2015 was limited to 25 hours WT and 25 CFD true for many higher clock speed chips. CCF = Peak processing unit clock frequency in GigaHertz TeraFLOPs, represented by the blue shaded area, Eff ectively, running supercomputers with NCU = Number of processing unit cores used for the run which remained the same until 2018. slow clock speed was giving teams more NSS = Number of solver wall clock seconds Since 2013, you can see that overall testing effi ciency under the regulations but with the elapsed during the run has dramatically reduced, but particularly for the obvious penalty in terms of CFD simulation

Figure 1: The FIA has restricted aerodynamic testing over recent years for both CFD and the wind tunnel, but particularly the latter. This graph shows the ‘working area’ that the teams have been able to operate in. Assuming a maximum CFD capacity of 12.5 TeraFLOPS you can see that wind on time has dropped by 70 per cent between 2013 and 2015

6 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 6 14/12/2018 11:13 Untitled-61 1 30/06/2016 13:22 TECHNOLOGY – AERODYNAMICS

turnaround time. Therefore, teams then had to regulation, but it was very ineffi cient fi nancially, the FIA agreed to consider the rival Intel chips balance the speed with which they receive their with as much as half of the purchased HPC (Sandybridge and Ivybridge) as four dp fl ops/ CFD results against the total number of CFD compute cores being left idle. cycle for the purposes of the regulations rather simulations they were able to complete within The biggest issue came when one of than their rated eight dp fl ops/cycle. the regulatory framework. This is quite diff erent the teams developed the Fangio chip in By 2012 AMD had been persuaded by many to the wider CFD industry, where the turbo collaboration with AMD, a chip specifi cally teams to produce a second limited run of Fangio mode was ‘free’ performance and quicker designed to optimise the balance between CFD chips, allowing more of the grid to upgrade clock speeds were performance gains if your case turnaround time and throughput which their supercomputers to this specifi cation, main criteria was fi nancial effi ciency, and so gave that team a huge initial advantage. with most of the remaining teams running an the divide between the two environments was This exploited the fact that the modern Intel Ivybridge system. With the FIA unwilling underway from 2009 onwards. HPC chips were then rated at eight double to extend the fl ops/cycle exemption to more precision fl ops/cycle but commercial CFD modern Intel chips, such as the V3 Haswell Cores and effect codes were only capable of delivering CPUs which were rated at 16 dp fl ops/cycle, and Core under-population also became approximately one dp fl op/cycle. The Fangio AMD not producing any more Fangio chips, the commonplace in Formula 1 as it delivered chip was designed to operate at two dp fl ops/ teams were now locked into these older systems further regulatory effi ciency gains for the cycle giving a big effi ciency improvement in purely by virtue of the regulations. Newer chips teams. It was effi cient for the FIA TFLOPS MFPPC. Following lobbying from various teams, were simply not viable because of their high

Fully correlated and complex CFD models, such as this by Simscale, are becoming an ever-increasing asset to F1 teams, with some full car models now exceeding one billion cells

Every restriction simply triggers teams to exploit the loopholes, and optimise designs and working practices to maximise performance

8 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 8 14/12/2018 11:13

TECHNOLOGY – AERODYNAMICS

A typical HPC cluster from Boston. With each new generation of BOSTON LTD BOSTON compute chip delivering up to 20 per cent effi ciency improvement the increased capacity of modern CFD clusters means that teams can now have an extra 200 runs, as opposed to 20 back in 2009

fl ops/cycle rating. These older systems were WON was retained and a parallel regulation was node system based on the Intel Skylake 8176 coming to the end of their life and were no introduced with the aim of allowing teams to Platinum 28 core chip. The group also had longer supported by Intel or AMD. continue using their old systems if they wished, access to a smaller four node Intel Ivybridge Clearly the FIA had to do something, and the without too large a performance penalty – at HPC which was used to provide a baseline of the target was to introduce a new regulation which least that was the intention. performance gains that teams could achieve by aligned the Formula 1 aero departments more Boston Ltd has been specialising in high upgrading from their older systems to the new closely with the wider CFD industry as well as performance computing (HPC) in a wide range hardware. This allowed Boston to benchmark its allowing teams to upgrade to more modern, of sectors for over 25 years. In 2017 it formed own internal CFD model across a range of CFD supported technology. This resulted in the 2018 a new partnership with Tim Milne of UniFi codes with a wide variety of hardware set-ups. CFD restrictions and a move from TFLOPS to Engineering Services Ltd (UniFi) and Dr Lee The systems were all set up with the very latest Mega Allocation Unit hours (MAUh) as defi ned Axon of Computational Engineering Ltd (CE). in networking fabric, up to date operating by the following equation: Milne and Axon have extensive Formula 1 systems and storage solutions, ensuring that experience, most recently at Manor F1 where the results obtained would be aligned to the they were head of aerodynamics and head of expectations of the F1 teams. CFD correlation respectively. This group combined Boston’s extensive Hot chips HPC technical knowledge with UniFi’s and Following the benchmarking of the older Where: CE’s F1 aerodynamics and CFD experience to Ivybridge system, a number of options within AUh = Allocation unit hour provide the F1 teams with a comprehensive the AMD EPYC range as well as the Skylake NCU = Number of processing unit cores benchmarking of the new AMD EPYC and Intel 8176 chip were evaluated as single node tests NSS = Number of solver wall clock seconds Skylake Platforms. They were able to use all the to gain an initial assessment of the various elapsed during the run main F1 CFD codes with models aligned to F1 diff erent chips available in each family, as CCF = Peak processing unit clock frequency in GigaHertz methodologies and HPC hardware set-ups to well as some insight into the time/iteration extract the maximum possible performance performance benefi ts of diff erent options such Eff ectively this a very similar measure to TFLOPS from the new regulatory environment. as the turbo mode. This also ensured that a clear but without the reliance on fl ops per cycle, understanding of the raw performance of the hence removing the barrier to upgrading to Node to joy compute chip was gained and that the results newer, better supported, technology. The FIA The project began in August 2017, by which were not clouded by any networking issues commissioned an independent study to be time Boston Ltd was one of the fi rst companies which could be useful later in the process when carried out in order to set the regulation limit worldwide to have invested in its own eight trying to understand the results on the larger with the intention of giving parity between the node dual socket AMD EPYC system based on scale multi-node systems. The performance old regulations and the new ones. The link to the EPYC7601 32 core chips and a similar eight gains over the older Ivybridge system were The new method that was introduced at the start of this year is a very similar measure to TFLOPS but without the reliance on fl ops per cycle

10 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 10 14/12/2018 11:13 INTRODUCING... THE BOSTON INTEL® SELECT SOLUTION

BUILD YOUR SUPERCOMPUTING INFRASTRUCTURE ON BOSTON'S EXTENSIVE INDUSTRY AND DESIGN EXPERTISE FOR HIGH PERFORMANCE COMPUTING APPLICATIONS.

BOSTON’S CFD SERVICES

THE DIFFERENCE BETWEEN WINNING AND LOSING

Our team of motorsport experts are on hand to support with code profiling and optimisation, can aid in tuning meshing and solver performance, and more...

VISIT: WWW.BOSTON.CO.UK/AUTOMOTIVE/CFD.ASPX

WEB: WWW.BOSTON.CO.UK EMAIL: [email protected] PHONE: +44 (0) 1727 876 100 TECHNOLOGY – AERODYNAMICS SIEMENS Despite the efforts of the FIA to restrict the costs of CFD work, Formula 1 teams can now complete up to 1500 simulations each week on a typical model of around 200 million cells

very quickly evident and it soon became For example, the CFD case will be repeated clear that the teams would all be forced to on the same HPC system but testing the It soon became clear that upgrade their HPC systems in order to remain simulation on 48, 96 and 192 cores. It was competitive, which is the nature of Formula 1. accepted that the case being run on 96 cores the teams would all be But this upgrade was extremely expensive. This will take slightly longer than half the time of is not what the FIA had been aiming for, but the case on 48 cores and slightly less than forced to upgrade their reflects how quickly the HPC industry moves half the case being run on 192 cores – so forward with the Formula 1 environment forced there is an element of inefficiency by running HPC systems in order to to follow suit to remain competitive. on an increasing number of cores. However, Once testing migrated onto the full, multi- it is in the teams’ interest to complete their remain competitive node systems the full optimisation process CFD simulations quickly in order to allow could begin. This involved running the same their iterative aerodynamic development model over a wide range of different set-ups, programmes to continue as quickly as possible delivered performance gains which would including options for memory bandwidth – so it’s a trade off and one which was vitally enable the F1 teams to run approximately twice per core used and process bindings. The key important for the Boston group to understand. as many CFD simulations per week in 2018 than at this stage was for the group to develop an they had been able to in 2017 (for the same understanding of the efficiency vs performance Core values wind tunnel operation). Furthermore, the teams of each compute system – ideally each compute The next step was to understand the impact of would complete each of these simulations in chip in each family from Intel and AMD. leaving some of the compute cores dormant, approximately half the time that was required In reality UniFi and CE were able to use their as previously mentioned. This is an approach under the 2017 regulations. experience in the industry to limit the testing quite alien to most of the CFD industry (why to the most likely candidates for Formula 1 would you buy compute cores and then not Formula 1 specific operations and Boston used its extensive links use them?) but something that was already well Much of this optimisation is not relevant to in the HPC industry to gain access to relevant known to deliver regulatory efficiency in the F1 the wider CFD industry, but is now considered systems for benchmark testing. Once a small environment, if you could afford it. basic within the Formula 1 teams. The next step range of AMD and Intel compute chips had Tests were completed leaving a range was for the Boston group to really exploit the been selected, the focus was on understanding of the cores dormant in order to give less expertise available from the UniFi/CE group. how they performed against the Formula 1 operational cores per memory channel, and The details of this remain confidential, but it regulations. This required repeating the CFD thus increasingly improving the memory enabled the group to develop solutions which simulation of their Formula 1 car on a range of bandwidth available to the CFD simulation. delivered even more performance for the F1 different HPC sizes and set-ups. The conclusion of this benchmarking study teams, and a further 20 per cent reduction Much of this optimisation is not relevant to the wider CFD industry, but it is now considered basic within the Formula 1 teams

12 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 12 14/12/2018 11:13 The FIA focus remains on reducing wind tunnel reliance and delivering greater CFD capacity in exchange, and the current regulations deliver that

The benchmarking study concluded that teams would gain a huge performance advantage if they purchased a new multi- million pound system because they would have twice the CFD capacity of 2017 – this was not the aim of the regulations

in solve times was extracted from the same the huge performance advantage available by replacing their CFD clusters. In 2018, with the CFD set-up, which also increased the CFD purchasing a new multi-million pound system, massive increase in capacity, the same 10 to throughput by the same 20 per cent. which was not the aim of the new regulations. 20 per cent improvement available from each Finally, as the benchmarking study neared Furthermore, the impact of the increase evolution of compute chip technology is 100 its conclusion Boston worked with AMD to in CFD capacity available to the teams under to 200 runs – that is the same as the total further optimise for the requirements of F1 these new regulations only serves to increase capacity of the systems in 2009. by increasing the memory bandwidth whilst the financial pressure on the teams and in Is this a bad thing? Arguably not. HPC retaining a relatively low base clock speed. particular the pressure to increase headcount systems are much cheaper now than they ‘AMD EPYC delivers exceptional levels within the aerodynamics departments as the were back in 2009. The FIA focus remains on of performance in a number of workloads, CFD capacity available increases. Not only reducing wind tunnel reliance and delivering including high performance computing CFD have they effectively been required to invest greater CFD capacity in exchange, and the applications,’ explains Roger Benson, the senior in new HPC architecture in order to remain current regulations deliver that. director of the Datacenter Group, EMEA, AMD. competitive, but the incentive to adopt future However, does it help to level the playing ‘We are excited to be working with Boston on improvements in chip technology has now only field between the high budget teams and their automotive engineering focused platforms increased. How so? The benchmarking work the low budget teams? Does it help to and improving the efficiency of aerodynamic completed by Boston suggests that teams are encourage new teams into the sport? And testing for their customers.’ now able to complete between 1000 and 1500 does it make the working practices within the CFD simulations per week based on a typical Formula 1 aero departments more aligned The results CFD model of around 200 million cells. Teams to the wider CFD industry? The stated targets of the FIA for this change may elect to ‘trade’ some of this capacity for With AMD releasing its second generation in regulations was to enable the F1 teams larger models (some teams run CFD models of EPYC chip in 2019, the reaction of the to upgrade from their Fangio and Ivybridge approaching one billion cells) or better quality teams will be interesting. Will they all upgrade systems to the latest technology available, but models (transient simulations rather than steady immediately? Or will the well-funded teams without a clear performance pressure to do state). But the key point is that the F1 HPC take the opportunity to get a performance so, and with the aim of better aligning the F1 regulations have now given the teams twice as advantage from the new technology that industry with the wider CFD industry. much capacity to play with than in 2017. the smaller teams cannot afford? Firstly, it is clear that all the F1 teams have upgraded to a new system, with most teams Step change Boston, UniFi and CE continue to develop having done so ahead of the regulatory change Typically each generation of compute chip their partnership with a focus on the F1, date of 1 January 2018. So, the first aim has that is released by AMD/Intel delivers around motorsport and automotive industries across been achieved – the Fangio and Ivybridge 10 to 20 per cent improvement in efficiency. all CFD codes and working practices. For systems that the teams were operating are now Back in 2009 this would give the teams an extra more informaition visit the websites at: www. obsolete. However, the benchmarking work 10 to 20 CFD runs per week, and therefore boston.co.uk; www.unifimotorsport.com; completed by Boston clearly demonstrates would not easily justify the large cost in www.computationalengineering.co.uk

When AMD releases its second generation of EYPC chip in 2019 the reaction of the Formula 1 teams is going to be very interesting

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 13

RCE F1 digi mag.indb 13 14/12/2018 11:13 TECHNOLOGY – PRESSURE SENSORS Pressure relief Thanks to a tiny new device, measuring aerodynamic pressures on and around cars may have just become a whole lot easier, and in some areas actually possible for the first time. Racecar investigates By SIMON McBEATH

hotographs of pre-season themselves, via tiny pressure port tunnel as well, to accurately monitor calibration of pressure sensors and pre-race testing often tappings over and under the major and control flow conditions. for many years, and have worked show top race teams, downforce-generating surfaces. But how are these measurements closely with a concern that might especially in F1, measuring Using these techniques enables made? And what are the practicalities? accurately be described as the Paerodynamic pressures around their aerodynamicists to gather pressure We visited Evolution Measurement sector founder, Scanivalve (see cars. The most obvious manifestation data around and on the car’s surfaces Ltd (see sidebar), based in Andover in sidebar) since 2001. With an intimate of this is the pressure sensor array, like that can be used to correlate with southern England, to learn more about knowledge of the available products a two-dimensional rake with pressure CFD and wind tunnel data, as well as the challenges involved, and to see an and, importantly, a firm grasp of sensors arranged over the area of the to calculate the real forces acting on ultra-compact new pressure sensor the customers’ needs, Evolution rake, positioned on the raceccar in key individual components. the company has come up with, that Measurement has come up with the areas where what might be described Pressure measurement is an will open up exciting new possibilities. EvoScann P Series pressure scanner as partial plane pressure plots can be everyday critical part of wind tunnel Although Evolution Measurement which, they confidently assert, is logged and recorded. testing too, with the wind tunnel is itself a new company, managing the smallest, lightest such pressure Less apparent and no less models equipped with surface director Paul Crowhurst and export sensor currently available – no surprise important is the measurement pressure ports, and measurements sales manager Iain Gordon have then that it has been generating of surface pressures on the cars also being made around the wind been involved in the distribution and interest among F1, DTM, LMP and

14 www.racecar-engineering.com FORMULA 1 · WINTER 2018

F1 Supp_EM_Pressure_MBAC3.indd 14 14/12/2018 15:19 Renault RS18 with pressure sensor array. This is highly visible but there might also be small sensors measuring air pressures that are fitted to the car itself The EvoScann P Series miniature pressure scanner is so small (36mm x 33m x 8mm) it can be located in areas of a racecar that have previously been impossible to access

MotoGP teams. This compactness – applications. Traditionally pressure other controlled environments’. It is the 8-channel launch version is just sensor array and surface pressure a 16-channel device typically used in These enable 36mm (1.42in) x 33mm (1.30in) x 8mm measurement has often been done wind tunnels where installation space (0.31in) and weighs only 16g (0.56oz) – with one of the Scanivalve multi- is not an issue, overall length being aerodynamicists will enable the sensors to be located in port scanners such as the ZOC (Zero, around 200mm (8in). areas previously difficult or impossible Operate Calibrate) range of analogue The newest and most advanced to gather to access, such as within front wing devices in 32- or 64-channel form, and Scanivalve product is the MPS elements or other important small these measure 105mm (4.1in) x 36mm (miniature pressure scanner) range, pressure data aerodynamic parts. 1.4in) x 14mm (0.55in), finding use in a 64-channel high-end device wind tunnel models and other space- measuring roughly 89mm (3.52in) around and Pressure change limited applications. x 40mm (1.56in) x 22mm (0.87in), To see where and how this new Another oft-used device is the several of which might typically on the car’s compact pressure scanner fits into popular DSA range, described by be seen on a wind tunnel model, the overall scheme of things, let’s Crowhurst as ‘a good workhorse’ and and which is also used in on-car surfaces first take a brief look at some of the by Scanivalve themselves as ‘intended applications, too. ‘It’s a fantastic complimentary products and their for most laboratory, educational or product,’ says Crowhurst, ‘and the

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 15

RCE F1 digi mag.indb 15 14/12/2018 11:13 TECHNOLOGY – PRESSURE SENSORS

Scanivalve MPS Scanivalve DSA3217 4264 pressure pressure scanner scanner, with pen for scale

Figure 1: Schematic layout of tube runs in a conventional pressure scanning system Tubing lengths should be as short as possible to avoid

frequency response issues The schematic in Figure 1 might not look so complicated but in reality on a wind tunnel model (or a racecar) a conventional pressure scanning set-up can look more like this

latest version is a big technical so forth has therefore inevitably usual on-vehicle communication on accuracy, performance, sensitivity advance over its predecessor as well as been highly compromised by the protocols (CANBus). The scanners and noise insensitivity was great.’ being the most technically advanced, long tubing runs that have been pick up their power supply from the The P8 launch model offers a fully digital instrument available.’ intrinsically required. communication cable, and feature choice of pressure ranges; at the low However, on the strength of integral signal conditioning. end of the scale +/-20kPa range is Tube lengths feedback from customers, Evolution offered with an accuracy of 0.1per Among the criteria for obtaining Measurements’ concept for the Small and light cent full scale claimed, providing +/- accurate and repeatable pressure EvoScann P Series was to reduce Gordon says: ‘Our Evoscann P series 20Pa resolution. However, a number measurements is that tubing lengths the size (and channel count) of the is certainly the smallest and lightest of developments are underway, should be as short as possible to scanner so that they could be located pressure scanner available and it Crowhurst says: ‘The launch version avoid frequency response issues, and much closer to where the pressure can be fitted [to measure pressures] was primarily aimed at absolute should ideally all be of equal length measurements needed to be taken. in places that were previously not surface pressure measurements. to maintain consistency between The concept of a ‘distributed system’, measurable. It’s been interesting But we will shortly be releasing a channels. The diagram in Figure 1 shown in Figure 2, was to switch from that the door has been held open differential pressure version which generically illustrates this challenge. a small number of centrally located wherever we have introduced it, so will measure the difference between With tightly packaged racecars, multi-channel scanners with complex it’s clearly meeting an un-met need. atmospheric pressure and the [local] available space for pressure scanners, tubing runs to more scanners with We have presented it to the F1 teams surface pressures creating lift or associated hardware including power lower channel counts, and to locate and a few others outside F1 and downforce. This will have a choice of supplies and signal conditioners, the scanners around the car close many have either bought it or asked ranges from +/-7kPa upwards.’ not to mention bundles of tubing, is to areas of interest. for customisations, or want to have it With a similar accuracy to the clearly at a premium, and options are This vastly reduces the length when we reach the next stages in our launch model this would enable +/- severely limited within the chassis of a and complexity of the tubing runs roadmap development.’ 7Pa resolution, which would be well racecar and perhaps other areas within and associated installation, with Gordon also revealed that able to resolve the small variations the body of a scale wind tunnel model. just a single cable emerging from EvoScann sensors were run on three in surface pressures seen on many So getting reliable data from ‘outposts’ each scanner to connect with cars in free practice at the last F1 race areas of racecars. However, Crowhurst such as front wings, endplates and data acquisition systems using the of 2017, and reported that ‘feedback added that ‘sensitivity and resolution

16 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 16 14/12/2018 11:14

TECHNOLOGY – PRESSURE SENSORS

will also improve, we will likely increase resolution by a factor of 10.’ Further variants with more channels will also be available; a 16-channel version that fits the same dimensional envelope will come, and a 32-channel version that will only be slightly longer will follow that. Embedded device One of the advantages of this slim pressure scanner is that in some instances it can be inset into the surface of the device it is intended to measure. For example, one F1 Figure 2: A distributed pressure sensor system vastly reduces the length and complexity of the tubing and installation. It puts local client wanted to attach one to a scanners at the point of measurement, with the tubing installation then connecting via a single CANBus cable to the nearest node bargeboard. The slender dimensions of the EvoScann would enable it to scanner inside the small chord front ‘We can even supply the EvoScann and mount the sensor nearby where it be embedded in the surface of the wing flap elements on an F1 car. And without its carbon composite casing, can be accommodated.’ bargeboard itself and cause negligible in some instances wind tunnel teams which reduces the thickness to just One F1 insider who is familiar interference with the flows and have used additive manufacturing 4.5mm (0.18in) if the customer wanted with these new sensors says: ‘The test pressures it was there to measure. methods (3D printing) to provide not an even more compact installation,’ teams are interested in mapping ever Clearly, the small size of the device only a snug location for the pressure Gordon says. ‘But if there really is an more areas of the car with these small will enable it to be mounted inside scanner within the test part but also inaccessible location that is too tight sensors. Getting a feel for downforce, many aerodynamic components to print the tubing runs that connect even for the EvoScann P8 then it is especially at low speed, is better done to allow the collection and local to the surface pressure tapping ports. possible to run the tubing through it with pressure mapping than with the conversion of pressure data into four [suspension] pushrod load cells. electronic signals to be communicated The pushrods have to measure car to the data acquisition system Its slender dimensions would weight and take impacts, whereas through just a single cable, rather than pressure sensors can be scaled for the through many metres of delicate and enable it to be embedded in the pressures they have to measure.’ vulnerable small bore plastic tubing. A further interesting feature is It is even feasible to fit the P Series surface of the bargeboard itself that because the EvoScann P Series Pressure scanners

imply put, a pressure scanner is data points could be obtained. Nobody Corporation in 1955, and the products including connections, small bore a device that converts data from else was doing this at the time and it were quickly adopted by the wind tubing, steel tubulations and so Spressure tappings, for example became a huge advantage. tunnel industry. The company forth. Subsequently the company over a car’s surfaces or from a sensor Pemberton subsequently left developed a large line of products had to adapt to – and exploit – some array, into electrical signals that can be Boeing and founded Scanivalve to support the use of the scanners, world-changing technological logged by a data acquisition system. The concept was devised by one JC Pemberton, who worked at Boeing In Seattle, Washington, in the 1950s and was endeavouring to measure pressures over aircraft surfaces using lots of liquid- filled U-tube manometers. Scanning valve Not surprisingly it was incredibly difficult to zero and stabilise all these devices and to obtain synchronous data. So Pemberton invented and developed a motorised, mechanical scanning valve that multiplexed many pressure signals into one transducer. The device was called a scanning valve because, in essence, the motor drove the transducer to sequentially scan the 48 ports connected to the valve. It did this in about 90 seconds, which vastly The original Scanivalve improved the rate at which pressure

18 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 18 14/12/2018 11:14 than most people need’. Pressure ranges from +/-20kPa to +/-120kPa (200-1200mbar or 2.9-17.4psig) are currently available, with lower ranges set to become available. Homologation As this article was being written Evolution Measurement was notified that EvoScann had been homologated for use with the FIA standard ECU. This process included, among other things, being able to demonstrate that, as a microprocessor-equipped sensor, it was not possible to re-programme the device for ‘alternative purposes’, something it is demonstrably not possible to do with the single CANbus Figure 3: Screenshot of EvoScann GUI. Device has an onboard integral microprocessor so the output is in engineering units of pressure communication cable that provides the power to and data from the devices have an onboard integral fits most known requirements’ Gordon environmental conditions to be sensor. So prospective customers microprocessor, the output is directly says. ‘Our aim is to not only fit the handled, although for applications now have the added confidence in engineering units of pressure, as niche applications where the standard that don’t need it the outer casing can that this compact, innovative new Figure 3 illustrates. Here the pressure device is most suited but also, where be omitted. The size of the internal pressure sensor has FIA homologation scale is in mbar but this is configurable required, to customise for bespoke printed circuit board is a key factor for use on their racecars. via the GUI to the units of choice. requirements. And, for example, the in the dimensions of the sensor; This also means that it will be The design of the EvoScann price per channel is comparable to the however there is the capability to permissible in FIA-sanctioned events incorporates temperature correction Scanivalve MPS scanner [mentioned reduce the size and thickness still to run the sensors during qualifying for each pressure channel. It can be earlier] so it’s a good fit in terms of the further. The in-plane alignment of the and in the races, and not just in test configured to measure differential applications it can satisfy.’ output tubulations helps to maintain sessions or free practice. Before such pressures by setting one channel The carbon fibre outer casing, a compact installation. Scanning rates a compact sensor was available, to measure static pressure. And it combined with the resin potting up to 1kHz per channel are possible, though, this probably wasn’t even corrects for ambient pressure too. ‘It process enable a wide range of which Crowhurst says ‘is typically faster considered by the race teams.

Pressure scanners Evolution Measurement

advances. During the 1970s Scanivalve In 1982 Pemberton sold the volution Measurement The company offers sales of the developed a combination of valving and company to his sons Addison and was founded in July 2016 devices including spares supply, calibration that was applied to miniature Jim, and around this time PCs started Eand is located in Andover, consultancy, support, installation, silicon sensors and used computer landing on all our desks, so the Hampshire, UK. It is staffed by a team service, calibration, repair and correction of temperature errors. And company developed PC-based data of engineers highly experienced bespoke solutions, as well as now measurement of individual sensors was acquisition systems which allowed in measurement, instrumentation designing and manufacturing its now multiplexed electronically, enabling ever faster data sampling rates. and calibration. The team has own new products. much faster sampling rates. actually been in its current premises, Managing director Crowhurst Digital age in a previous guise, since 2006 says: ‘we work in various sectors, In the 1990s Ethernet-based and is now fully focussed on the niche areas especially where, as communications boosted things, and high-end fluid temperature and a small, responsive company we in the early 2000s UBS connectivity pressure measurement market. The can provide turn-key solutions for speeded everything up further, and relationship with Scanivalve in fact multiple applications in building all the while miniaturisation was goes back to 2001, via the companies design, wind engineering, aircraft continuing. Originally signal outputs that the current staff used to work at. design, automotive, wind tunnel were analogue so the voltage signals assessment and models, and of required conditioning, but now Highly evolved course in motorsport.’ The latter everything is digital and processed Evolution Measurement now handles includes MotoGP, from where interest to output in pressure units, and northern Europe-wide factory- is emerging in what is an increasingly it is possible to take thousands of level calibration, repair and service aerodynamics savvy sport. readings per second per channel from support for Scanivalve while special There are applications in cycling, increasingly compact devices. relationships also exist with Guildline too. And the company is also working Such has been the impact of (as its exclusive UK distributor for on a package for Formula Student. Scanivalve Corporation that it has precision measurement solutions) Please form an orderly queue … become the Hoover of the pressure as well as with Meggitt (as its UK Contact: Evolution Measurement Ltd scanning industry, the generic name application-specific distributors for Tel +44 (0)1264 316470, web: An advert for the first Scanivalve device most folk reach for when a need arises. dynamic pressure sensors). www.evolutionmeasurement.com

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 19

RCE F1 digi mag.indb 19 14/12/2018 11:14 FORMULA 1 2018 – TESTING

The reigning champion Mercedes team has taken a conservative approach to Halo, where design and implementation regulations allow some freedom

20 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 20 14/12/2018 11:14 The Halo effect Formula 1 2018 hit the ground running in Barcelona with all new Pirelli compounds, aero and the controversial head protection Halo system By GEMMA HATTON and SAM COLLINS

ormula 1 has ushered in a host of is confirmed we’re confident that the parts we’ll changes for the 2018 season. The new bring to the car will sort out those losses.’ ‘In testing we will head protection system, known as It is something being worked on right Halo, is the most obvious from a visual up and down the pitlane with lots of airflow make sure we understand Fpoint of view, and has already attracted a lot of sensors fitted to cars around the Halo structure negative feedback from the teams. It has also and downstream of it. ‘Aerodynamically that the losses coming had a significant effect on the rest of the car, speaking, Halo is certainly not penalty free in terms of weight and design thanks to a late and I think there is a challenge there to either off Halo are where we introduction of the regulation leading in some cope with it in the first instance, let’s call cases to an all-new chassis design. With new it damage limitation, and thereafter think think they are’ tyres from Pirelli, offering teams a new challenge about opportunity and exploitation,’ Peter of working them at different circuits, and longer Prodromou, McLaren’s Chief Technical Officer ‘It has effects on the cockpit because it is local life power units for this season, teams have had for aerodynamics adds. ‘It does open up some to that opening. You have got the driver in anything but an easy preparation for the season. avenues which are possibly interesting to look there and so you’ve got to make sure you don’t The Additional Frontal Protection-Halo (AFP- at. I am sure there will be a variety of different have the negative effects there,’ Toro Rosso Halo, or just Halo) is without doubt the biggest solutions out there but the scope is quite Technical Director James Key adds. ‘You’ve visual change between the 2018 Grand Prix cars limited to the allowance around the basic got effects on the engine air intake and effects and those used in 2017. In design terms the shape, but there is opportunity.’ after that towards the back, so there are a Halo is governed by its own specific appendix to number of different things you have to think the FIA technical regulations. Everything from Aesthetic gain about. None of them are massive effects but the shape and dimensions of the device to the The rules allow a 20mm area of freedom around they all require some level of attention.’ material it is made from (titanium alloy Ti6Al4V the titanium structure, introduced partly for Fitting the Halo is no easy challenge either; Grade 5) is defined. However there is still scope aesthetic reasons but predictably these fairings not only does the Halo have to be homologated for different manufacturers to supply their are being used for aerodynamic gain, as some independently, it also has to pass crash tests as own products into the category, though each teams have added winglets and in one case part of the chassis homologation procedure. must be homologated independently at the airliner style vortex generators to their Halos. This has proved to be a major issue for teams. Cranfield Impact Centre. At the time of writing three companies had homologated Halos; CP Autosport of Germany, SS Tube Technology in England and a third company, V System, from which each team must purchase their Halos. Airflow impact As can be imagined for such a visually obvious addition to the car, the aerodynamic impact of Halo is noteworthy, and the teams are doing what they can to deal with its impact, particularly on the airflow over the whole car. ‘It has a significant downstream effect, especially round the rear wing area,’ highlights Andy Green, Technical Director of the team known as Force India at time of writing (the team name is likely to change by the first race in Australia in March). ‘It is not designed to be an aerodynamic device, so it doesn’t do us any favours in that department and it requires a lot of work to mitigate the issues that it causes. In testing we will make sure we understand that the losses coming off the halo are where we think they are from our modelling tools. If that Toro Rosso is one of several teams to try to increase aero efficiency with its Halo design, one of many to choose this option

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 21

F1 Supp_F1_2018-acHGAC3.indd 21 14/12/2018 15:20 FORMULA 1 2018 – TESTING

The 20mm area at the top of the Halo has been exploited differently by the teams. The Haas team has adopted this toothy solution while others have mounted a wing

everything that moves around – which it is designed to do – it is a bit scary.’ During the chassis homologation tests the Halo has to withstand various loads without it or the monocoque failing. The biggest load applied to the structure is 116kN from above, which has to be endured for five seconds. Longitudinal forces of 46kN and 83kN are applied from the front as well as a lateral load of 93kN from the side. For comparison, the roll structure on top of the car has to withstand 50kN laterally, 60kN longitudinally and 90kN from above. Weighty issue To survive these severe loads, the Halo itself has become quite a substantial structure, weighing by regulation 7kg (+0.05kg, -0.15kg). In addition, the monocoque has also had to increase in strength significantly to cope with these tests. This has further increased the weight of the New rubber from Pirelli is designed to help drivers and teams at ‘We always knew it was going to be a chassis by approximately 12-13kg. Whereas, the particular tracks. Pressure sensors were all the rage in Barcelona challenge so have invested time and money 2018 technical regulations have only allowed as teams completed their aero maps during pre-season testing up front to do a lot of test pieces,’ McLaren a minimum weight increase of 5kg to 733kg, Chief Technical Officer Matt Morris admits. forcing teams to save weight in other areas of ‘Obviously, you don’t want to build a complete the car. Interestingly, now at the start of a race a ‘It takes the weight of a chassis but we built a few test pieces with 2018 Formula 1 car will weigh roughly the same dummy Halos and parts of Halos to test how as a non-hybrid LMP1 in qualifying trim. London bus and when the interfaces would behave and we found ‘From a design perspective, weight is a big some issues. It was close, we didn’t breeze part of it. The weight limit did go up, but not by you see that test going through and there were some heart-stopping nearly as much as the installation weight of the moments with particular static tests coming halo so it put additional stress on all the other with that amount of load, in from an oblique angle. It takes the weight parts of the car,’ Green continues. ‘We had to try of a London bus and when you see that to optimise the weight in those areas to try and it is a bit scary’ test going on with that amount of load and keep the weight limit below the minimum so

22 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 22 14/12/2018 11:14 Short sidepod concept

n 2017 Ferrari introduced a new short sidepod concept, relocating the upper side impact structure (a single Ispecification design shared by all teams) and moving the main cooling aperture rearward. A set of box shape aerodynamic elements forward of the duct ensure rules compliance. Ferrari took this approach for aerodynamic reasons rather than those of cooling. In 2018, half the grid featured the same solution, but not all teams agree that it is the right route, with Mercedes, Renault, Force India and others all opting against adopting the concept.

Conservative approach Sidepod design seems to be led by Ferrari, with impact structure relocation for efficient aerodynamic effect ‘Everything you do in aerodynamics has an opportunity cost; there is much more opportunity to make the car worse than better,’ claims Mercedes Technical Director, James Allison. ‘If you want to pursue a new and different concept, you will expect to find a fair amount of loss before you get back into positive territory. We looked at that concept and felt it would spend too much time being in negative territory before it would perhaps offer any gain at all. If you are a [team] that is a long way down the grid the situation is different it is worth taking that gamble, as you have less to lose and you know that the path you are on is not right.’ It is likely that the relocation of the side impact structure would require a substantial change to the monocoque design, while getting adequate airflow into the cooling system with such a complex arrangement of aerodynamic elements around the leading edge of the sidepod duct is also likely to be a major challenge. Mercedes has not adopted this same approach, believing that too much time would be lost in development

that we can run ballast because the other area that we have to bear in mind is we have to hit a weight distribution target as well.’ Although it was originally introduced as a temporary measure to help Pirelli develop tyres when it became the sole tyre supplier in 2011, the technical regulations still limit every car in terms of weight distribution, with just a 7kg window of freedom. This means that while some teams may be able to build a car under the minimum weight, they cannot get it fully within the distribution window. Halo kitty ‘You only have a very small window of weight distribution so the actual architecture of the car needs to be correct to start with, otherwise you’re adding ballast to a car that doesn’t need ballast just to get the weight distribution right,’ Green says. ‘We would have loved to have added a huge safety margin to the whole The loss of the T-wing is not total; some teams are trying to recover some of the effect with lower mounted winglets design so that we would happily sail through the crash and load tests without any issues but rear end of its car as a result, abandoning its that wasn’t possible because the weight limit of cast titanium gearbox casing (something it ‘You have a small window the car didn’t go up enough. We couldn’t afford has evolved over many seasons) in favour of to increase the base weight of the car more a composite transmission. of weight distribution so than a few kg because we knew we only had a While the price of the Halo itself is relatively few kg that wecould take out of the car. It was, modest, the cost of developing a chassis to fit it the architecture of the car structurally, incredibly challenging.’ is higher than some of the smaller teams would This weight challenge has seen at least like. This cost was worsened by the late decision needs to be correct’ one team, Renault, substantially rework the to adopt the Halo as the 2018 AFP solution, with

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 23

RCE F1 digi mag.indb 23 14/12/2018 11:14 FORMULA 1 2018 – TESTING

With only a 7kg weight distribution, teams have struggled to get the weight down and remain in the window; Renault adopted a composite gearbox casing to reduce weight

teams only informed of this final decision in as it is much lighter than Halo with lower The significant aerodynamic changes of September, 2017 after a long discussion process. requirements on the chassis structure. the 2017 regulations resulted in an increase in ‘Expense-wise it’s huge because we had to The weight increase as a result of the Halo loads of over 20%, demanding the tyres to be do a new chassis. We wouldn’t have anticipated also places an additional demand on the four extremely robust, leading Pirelli to ramp up the doing a new chassis this year given the number power unit suppliers, which have also had to stiffness of their entire compound range. Pirelli of changes we made last year. For a team like increase the life of their power units. Teams can also had to develop tyres with little knowledge us we would look to try and get two years out now only use three combustion engines (ICE), of the potential performance that the teams of the chassis if possible. So in that respect three MGU-H’s and three turbochargers (TC) could achieve in 2017. Despite 12,000km of it cost us a huge amount to redevelop and during the season, compared to four last year. testing, the 2014 adapted ‘mule’ cars that Pirelli redesign the new chassis. It is in the hundreds That’s 2,100km of racing mileage not including used to develop the 2017 compounds only of thousands, if not million dollar mark, to put practice sessions or qualifying. Whereas the achieved a 10% increase in downforce and the Halo on the car for us,’ says Green. energy stores (ES), control electronics (CE) and therefore the results were unrepresentative MGU-K’s are all limited to two per season, or and inconclusive. To cope with this, Pirelli went Screening process 3,150km of racing. This demand for increased for a conservative approach last year, and The Halo has had a largely negative reception reliability will no doubt have forced the having tried and tested their designs for an from drivers, teams, the media and fans. This suppliers to manufacture more robust units, yet entire season, the 2018 range is a slightly more has lead to work continuing on alternative they have had to minimise weight to help teams aggressive evolution of 2017. additional frontal protection systems. In comply with the minimum weight regulations 2017 a brief test run was conducted with which have been challenging to achieve with Compounding issues a clear windscreen fitted to a Ferrari, but the consequences of Halo. It remains to be seen ‘The 2018 compounds are from the same family while this solution solved the frontal impact how successful they have been. of compounds as 2017,’ explains Mario Isola, requirements, the driver complained of Sporting Director of Pirelli. ‘The reason why visual distortion. However, Indycar is now Tyre dilemma degradation was so low last year was because experimenting with a similar aeroscreen The other major changes for this year come these compounds have less surface overheating solution (see p16). Teams prefer the windscreen from the tyres. To encourage overtaking and pit and in general behave in a different way. In option not only for aesthetic reasons but also stops, Pirelli have added two more colours, and particular we used the 2017 Soft as a baseline therefore compounds, to their tyre compound [for 2018] because last year the Soft had a rainbow, the Superhard and the Hypersoft, as wider working range compared to the other ‘We used the 2017 Soft as well as making the entire range a step softer, compounds. Last year’s Soft is now the Medium.’ and introducing new allocation rules. The From there, the 2017 Soft ‘baseline’ was a baseline because last Superhard is now the hardest compound, then developed and used to create this year’s adopting the conventional orange colour of softer compounds (Soft, Supersoft, Ultrasoft year the Soft had a wider the Hard, which has now become the light and Hypersoft), each decreasing in stiffness in blue, and the Hypersoft is the softest compound relatively consecutive steps. Although Pirelli, working range compared and is light pink in colour. However, to gain a along with some drivers, have commented full understanding of these additional that the softer compounds of the 2018 range, with the other compounds’ compounds we need to reflect on 2017. tested at Abu Dhabi last year were ‘much closer

24 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 24 14/12/2018 11:14 MADE IN THE USA MADE FOR THE WORLD

Learn more about Hastings Piston Rings at www.hastingspistonrings.com.

ATHENA RACE HEAD GASKETS

The Science of Speed

Performance Engine Bearings

Race Series Bearings Provide Exceptional Performance and Outstanding 2nd place is for the other guy Durability

XP BEARINGS - Tri-metal, Copper-lead HP BEARINGS - Bi-metal, Silicon Alloy ENGINE PARTS (UK) LTD

Purfleet: 01708 890494 Hinckley: 01455 250222 [email protected] www.enginepartsuk.net

King Combined A4.indd 1 28/07/2017 15:55 FORMULA 1 2018 – TESTING

➤ HARD SUPERHARD

MEDIUM HARD

SOFT MEDIUM SOFT SUPERSOFT SUPERSOFT ULTRASOFT ULTRASOFT

Compound stiffness HYPERSOFT

2017 2018 ➤ Working range temperature

Above: A simplified diagram illustrating the compound changes from 2017 to 2018. This year’s compounds are all a step softer, with the 2017 Soft and its wider working range becoming the 2018 Medium. The delta between the Soft, Supersoft and Ultrasoft are much closer, and the Hypersoft is an aggressive step, based on running at the Abu Dhabi test last year. Right: Pirelli’s new tyres on display in Barcelona – the colours were chosen by the marketing department

the compounds more predictable,’ says Isola. ‘Teams complained that they would set up the car for the Soft and it was difficult to manage when they put the Supersoft on. Now, with this change in working range it will be much better.’ With regard to the Superhard; ‘Forget it,’ laughs Isola. ‘We’re not going to use it. The Superhard compound is an insurance for us in case we have underestimated the development of this year’s cars. It’s much better to homologate an additional compound to keep in our pocket, rather than introduce a new one. From our simulations we are quite confident that we are not going to use this compound.’ Joy division This year’s softer tyres are not only going to make the drivers happier, but hopefully the fans as well. Softer compounds lead to higher Traditional testing methods like flow-vis are still a primary aero tool for F1 teams in testing; here a Haas in the pitlane degradation, resulting in larger performance differences between drivers out on track, so together’ in terms of the performance delta, The aggressive nature of the softer promoting more overtaking. To encourage this the Hypersoft is much more aggressive. compounds has also led Pirelli to modify the further, Pirelli have changed their tyre allocation ‘The Hypersoft is quite a step softer front tyre construction. Not only do this year’s rules. Rather than teams choosing their compared to the Ultrasoft,’ highlights Isola. ‘We tyres feature a rounder profile, incorporating allocation from three consecutive compounds don’t have a lot of data but at Abu Dhabi, which new materials, but the distribution of forces specified by Pirelli, teams can pick a double step is a low severity circuit and not that far from over the contact patch have also improved. in compound. For example, instead of running a street circuit, the Hypersoft was behaving the Medium, Soft and Supersoft, teams can use like a very soft compound. It was about 0.9-1.0 Shooting range the Medium, Soft and Ultrasoft, as is the case for seconds per lap quicker than the Ultrasoft and ‘The other difference for this year is that the this year’s Chinese Grand Prix. This opens up the it was able to run for eight laps on average.’ working range now decreases consecutively options for some interesting strategic decisions, However, Valtteri Bottas at this year’s Mercedes from the Medium to the Hypersoft,’ says Isola. which again could result in more exciting racing. launch highlighted how the Hypersoft was only ‘We don’t have this alternating between Although 2018 is an evolutionary year in suitable for 2-3 laps during Abu Dhabi testing. low working range and high working range terms of regulation, once the effects of Halo Similar to when Pirelli introduced the compounds. The harder compounds are high have been validated on track, teams will be Ultrasoft in 2016, the pink Hypersoft has been working range and the softer compounds are bringing plenty of performance upgrades predominantly designed to give drivers that low working range.’ Previously, the high working throughout the season. This, together with the extra level of grip at street circuits. Depending range compounds were the Hard and Soft with unknown performance of the new tyres and the on the results from Monaco, however, teams the low working range compounds the Medium increased pressure on power units, gives 2018 might just see the pink tyres at other low and Supersoft. The Ultrasoft was Medium to all the ingredients for an exciting season – severity tracks towards the end of the season. High working range. ‘This is important to make let’s hope that this is the right recipe. ‘The Superhard compound is an insurance for us in case we have underestimated the development of this year’s cars’

26 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 26 14/12/2018 11:14 HIGH SPEED CONNECTIVITY SOLUTIONS

MINIATURE IP68 SEALED RUGGED RELIABLE

IP68 SEALED UNMATED High Tech | High Speed | High Quality Connect with us Call +44 23 9245 9600 Email Sales@fischerconnectors.co.uk

NEW USB 3.0

MiniMax: ultra-miniature USB 3.0 connector, 12 mm diameter

 Ultra-compact and lightweight  Signal & power  Up to 10 Gb/s data transmission Pankl Systems Austria GmbH Engine Systems

A-8600 Bruck/Mur, Kaltschmidstraße 2-6 Phone: +43(0)3862 33 999 0 Fax: +43(0)3862 33 999 290 e-mail: [email protected] USB 3.0 Rugged Flash Drive for the safe storage of sensitive data

 Up to 100 MB/s read speed www.pankl.com  Up to 128 GB memory size  Withstands corrosion, vibration, shock and extreme temperatures

www.fischerconnectors.com THE RELIABLE EXPERT TECHNOLOGY – HELMETS Head first The FIA’s updated helmet standard, which will be mandatory in Formula 1 from 2019, promises to take driver safety to ‘the next level’ – but how will this improve what are already the very best racing helmets on the planet? By PETER WRIGHT

28 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 28 14/12/2018 11:14 n motor racing’s early days drivers made do helmet standard, working with the industry The 8860-2018 with leather helmets and goggles to protect and test laboratories to develop test standards, them from stones, dust, wind and rain – but protocols and helmet construction methods helmet standard not head impacts. Then, in 1954, Bell Sports to achieve the latest performance standard. Ideveloped the first mass produced motorsport Mellor worked at TRL (Transport Research extends the test helmet, and five years later the Snell Foundation Laboratory), performing the R&D and writing developed the first motorsport helmet standard. the specifications for the UK DOT’s Advanced areas at the sides of Nearly 60 years on the FIA has now Motorcycle Helmet. Since working with the updated its racing helmet standard, 8860-2018, FIA Institute (now the Global Institute) he the helmet to match launching it at the FIA Sports Conference Week has authored a series of helmet standards: in Manila in June. This standard provides the 8858-2002, Auto Racing Helmet; 8860-2004, the car components most sophisticated and protective performance Advanced Racing Helmet; Visor reinforcement in a motorsport helmet. The helmet industry for 8860 in 2011; 8860-2010, update; 8859-2015, – represented by Stilo, Bell Racing Helmets, Premium Helmet and now 8860-2018 update. Schuberth, and Arai – has worked with the FIA With so much experience in the R&D of throughout and developed their prototypes for helmets, there is no one better than Mellor to this standard. The first homologated products discuss the physics and engineering behind the will be available for Formula 1 in 2019, quickly latest motorsport helmet standard. followed by other top FIA championships (see box out on page 86 for the full spec). Bone dome Since composite and polystyrene foam replaced Head start leather and cork in the 1950s, the concept A modern motorsport helmet may look of a strong, semi-rigid outer shell and an simple, while also being stylish, but it provides energy absorbing liner has dominated helmet protection to the driver’s most valuable asset, construction. Carbon fibre has replaced glass their head, under extreme and emergency fibre for the shell, whose task it is to prevent conditions when involved in a crash. It sets out penetration or fracture that would leave the to restrain the head, via the built-in FHR (frontal skull vulnerable to injury, and to spread the head restraint) anchors; to prevent skull fracture impact load into the foam liner and limit the and limit deceleration to below 300g whatever deceleration of the brain and its connections the head hits in his racecar, at the highest likely experienced during an impact. head velocity relative to the car. It must also Unlike motorcycle accidents on the public protect against a loose object at an impact roads, what the helmet actually strikes can be velocity of over 250km/h without inhibiting tightly controlled in most motorsports these vision. All this with a structure that is just 50mm days. In the past, the driver’s head could hit the thick and at a weight of under 2kg. road, barriers, trees, and other cars, but today Andrew Mellor is the person who has, over the helmet and the car are regulated as a single the last 20 years, steadily progressed the FIA’s protective system. The development of FHR,

The FIA worked closely with 8860-2018 standard specifies that the helmet must withstand Stilo, Bell Racing Helmets, this 225gm metal curved disc fired at 250km/h into the top of Schuberth, and Arai while the visor area while not subjecting the head to more than 275g formulating its new 8860- NEWTON LABS FIA

2018 helmet standard FIA

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 29

RCE F1 digi mag.indb 29 14/12/2018 11:14 TECHNOLOGY – HELMETS NEWTON LABS The projectile is loaded into the barrel of a pneumatic gun powered by a pre-charged reservoir of compressed air (this test rig was developed by the FIA and Newton Laboratories)

One issue that has evolved with the very high average g necessary is that the energy absorbing materials, whether in the helmet or headrest/seat foam, have to be very stiff . This stiff ness tends to lead to quite high peak g at lower head impact velocities, potentially leading to concussion. Balancing the trade-off between concussion and severe head trauma has led to the use of Confor foam in headrests and seats. This is relatively soft at low velocities and stiff ens up at high velocities due to its inherent viscous damping properties. Visor panels Then, in 2009, Felipe Massa was struck on the head by a rear suspension third spring at the Hungarian GP. The spring penetrated his visor and helmet and caused skull fracture. After extensive simulations of the accident at Aermacchi’s ballistic impact test facility in Italy,

NEWTON LABS Mellor developed a visor reinforcement panel The velocity of the projectile is measured at the muzzle of the gun. The target is instrumented with a tri-axial accelerometer for the 8860 helmet. This consists of a 50mm wide strip of Zylon composite, bonded to the headrests, steering columns, racing nets, roll top of the visor and covering 25mm of the The concept of a strong, cages, and seats are all specifi ed to interact with helmet just above the eye port. The 25mm of the helmet. The 8860-2018 helmet standard visor covered by this strip is generally used for a semi-rigid outer shell takes this compatibility to the next level by sponsor banner, so does not reduce vision. Tests extending the test areas at the sides of the showed that this would have prevented helmet and an energy absorbing helmet to match the car components. penetration in Massa’s accident. To limit head acceleration to less than 300g Since being used in Formula 1 and in IndyCar, liner has dominated – a level at which there should not be long-term this anti-penetration strip has prevented serious medical consequences – the new drop test is or fatal injuries on at least three occasions. helmet construction conducted at 9.5m/sec, the equivalent of a head However, fi tment and maintenance of the impact at a head velocity of 35km/h! To pass protective strip requires signifi cant servicing, since the 1950s the test requires careful design and use of so it could only be mandated in open-cockpit materials characteristics for the shell and helmet championships where helmet manufacturers liner that spread the load into the head below were able to provide this service. the skull fracture load, and ride down the head The 8860-2018 standard, as applied to at a deceleration that does not exceed the 300g helmets used in open-cockpit championships, limit. The average g over the 30mm of eff ective incorporates this protection directly into the liner crush will be at least 150g. helmet, with the eye port upper edge lowered

30 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 30 14/12/2018 11:14 Figure 1: Development of FIA standard 8860-2018 ANDREW MELLOR

ANDREW MELLOR Table 1 Table 2 Being suspended, the helmet and headform are free to move Projectile Helmet shell Head Projectile + helmet shell Head longitudinally and laterally to achieve the right body dynamics Mass: 0.225kg 1.3kg 4.7kg Mass: 1.525kg 4.7kg Velocity: 250km/h 0km/h 0km/h Velocity: 37km/h 0km/h Energy: 542J 0J 0J Energy: 80J 0J

10mm, and does so using a simple but clever the panel fails to prevent penetration of the physical principle: momentum transfer. shell, the projectile easily passes through the The standard specifi es that the helmet must liner and strikes the headform, resulting in withstand a 225gm metal curved disc fi red at head gs of over 330g. When the panel prevents 250km/h into the top of the visor area, and not penetration, g-levels are kept below 50g – subject the head to more than 275g. Such a probably a headache, but not the appalling projectile has energy of 542 Joules, exceeding injuries that Massa suff ered. the muzzle energy of most pistols … except for This protection, built into the 8860-2018 Dirty Harry’s .44 Magnum. helmets, will deal with small loose objects such The physical conditions just before the as the spring that struck Massa in Hungary in projectile hits the helmet are shown in Table 1 2009, then. But it will not deal with large, heavy while Figure 1 shows just after impact, during objects such as a wheel and tyre (Henry Surtees, which momentum is conserved. 2009) or a nosecone (Justin Wilson, 2015). It is The loss of energy is accounted for by the for protection against these objects in particular plastic deformation of the Zylon panel, which that the Halo was developed. dissipates the energy by the crushing of the composite in the same way as the composite Barrier brief crash structures on racing cars absorb impact Momentum transfer is not just used in ballistic energy. Now the helmet shell is travelling at only protection for helmets. It is also employed in the 37km/h relative to the head. These conditions design of high-speed barriers. To bring a car to a are such that there should not be signifi cant halt from 200+km/h without hurting the driver injury to the head (Table 2). requires deceleration at around 60-70g over

ANDREW MELLOR around 3m. The FIA high-speed barrier achieves This shows the actual hit of the projectile against the helmet and Head shot this partially by momentum transfer. Segmented visor. The test helped prove that using momentum transfer worked The test rig and procedure was developed with barriers of a prescribed mass (110kg) are set up Newton Laboratories, in Milan. The projectile in layers, with spaces between the rows. Built consists of an aluminium piston and the curved into the sections are anti penetration layers, This protection, built into steel impactor, weighing 225gm in total. just like the helmet’s Zylon panel, but in this This is loaded into the barrel of a pneumatic case they are steel to resist the pointed nose the 8860-2018 helmets, gun, powered by a pre-charged reservoir of the car. The car connects with the barrier of compressed air. On fi ring, the velocity of at the point of impact and draws connected will deal with small loose the projectile is measured at the muzzle. The sections of the barrier inwards and forwards, target is an inverted, suspended headform, progressively adding mass to the car. The initial objects such as the spring instrumented with a tri-axial accelerometer, deceleration as the car hits the fi rst barrier is on to which the test helmet is fi tted. Being attenuated, and energy is dissipated by the that struck Felipe Massa suspended, the helmet and headform are free crushable nose cone and the friction between to move longitudinally and laterally to achieve the barrier segments and the ground. Once the in Hungary in 2009 the representative body dynamics. car has coupled with suffi cient barrier mass, Development of the specifi cations for raising the total by a factor of around three Mellor’s approach to using momentum transfer times its running mass and totalling around 2.5 theory showed that it worked in practice. If tonnes, it will have slowed to 60km/h, which

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 31

RCE F1 digi mag.indb 31 14/12/2018 11:14 TECHNOLOGY – HELMETS

is dealt with by crushing the barrier sections against the Armco or concrete final barrier. Carlos Sainz’s impact at Sochi in 2015 at over 150km/h, when he hit three rows of Tecpro barriers spaced to give an overall depth of 3m, illustrates this effect very well. The barriers slowed him to under 60km/h at an average of 20g, and the final phase, as the car crushed the Tecpro blocks against Armco, had a peak of only 40g. This progressive slowing enabled both the car and the driver to race the next day. Sudden stops Momentum transfer, or the conservation of momentum, is a principle that forms a valuable tool in the motorsport safety toolbox. Whoever it was who said: ‘It’s not speed that kills, it is the sudden loss of it,’ was right. Excessive deceleration or dissipation of energy into a human is injurious. Firing the 225gm helmet test projectile at 250km/h into a stationary 2-tonne car would stop the projectile and accelerate the car to under 3km/h. Collision with a 3km/h car would not injure a human if it Carlos Sainz’s accident at Sochi in 2015 was a good example of how momentum transfer works in the Tecpro crash barriers hit them. Matching up the masses of a helmet and projectile and managing the energy of the Momentum transfer, system is a neat trick of physics. Other changes to the helmet standard have set out to generally increase the helmet’s or the conservation performance characteristics. Tests to check the overall crush resistance, and resistance to of momentum, is penetration of the shell are uprated. To meet these tests the shell structural properties are a principle that tuned along with the impact requirements. Because helmets are produced in a range forms a valuable of sizes to suit the full spectrum of adult head dimensions, the variation in head mass tool in the motorsport that goes hand-in-hand with size means that each size of helmet/head mass needs safety toolbox These days crash helmets reflect a driver’s personality tuning to the homologation standards to and allegiances and also carry advertising. But this provide the same level of protection. should not detract from their main purpose; saving lives Visor performance, chinstrap strength, and flammability have also all been improved. Helmet spec 8860-2018 Evolving knowledge The head is the most vulnerable part of a • Standard impact. Helmet impact at 9.5m/s. • Visor coating. Transmitter test to ensure colour and racecar driver, particularly in an open-cockpit Peak deceleration on driver’s head shall vision is not significantly changed or distorted. car. Injuries to the head are among the most not exceed 275g. • Retention system. Roll-off test and dynamic test to serious, being either life-threatening or leading • Low velocity impact. Helmet impact at 6m/s. ensure strength of chin strap and its attachments. to long-term impairment. Although for many Peak deceleration shall not exceed 200g with a • Chin guard linear impact. Impact test with full years drivers relied on a leather helmet or maximum average of 180g. headform at 5.5m/s. The peak deceleration reversed cloth cap, science has provided a • Low lateral impact. Helmet impact at 8.5m/s. Peak shall not exceed 275g. very high level of protection in the event that deceleration shall not exceed 275g. • Chin guard crush. A hammer hits the chin the head is either struck by an object or strikes • Advanced ballistic protection. A 225gm metal guard and measures its ability to keep impact some part of the car’s structure. Research projectile fired at 250km/h. The peak deceleration away from the head. continues in line with evolving medical shall not exceed 275g. • FHR mechanical strength. A test to ensure knowledge about brain trauma and the physical • Crush. A 10kg weight falling 5.1 metres on high strength of attachment points for the forces and accelerations that cause it. to helmet. Lateral and longitudinal tests. The frontal head restraints. Not only does the racing helmet protect transmitted force should not exceed 10kN. • Projection and surface friction. Test to ensure the driver but it also provides a surface upon • Shell penetration. A 4kg impactor dropped helmet surface uniformity and that friction is which he or she can express aspects of his or her on to helmet at 7.7m/s. minimised. Shell surface also subjected to BARCOL personality and allegiances to the world. In that • Visor penetration. An air rifle fires a 1.2gm pellet hardness test for resistance to penetration. respect it is so much more than a safety device at the visor. The pellet must not penetrate the • Flammability: Helmet exposed to 790degC flame; it these days – but as a safety device it is still a interior of the crash helmet. must self-extinguish once flame is removed. seriously effective piece of kit.

32 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 32 14/12/2018 11:14

FORMULA 1 – 2021 REGULATIONS Reinventing the wheel

With Formula 1 announcing a shift to 18-inch wheels in the first of its 2021 regulations to be revealed, we examine the technical implications for F1 teams, brake suppliers and its future tyre manufacturer By SAM COLLINS

t is no secret that Formula 1 will introduce announcement took teams completely by change wheel size had been taken and it was an entirely new set of technical and sporting surprise. The revelation came in the form of an clear that they were somewhat taken aback that regulations for the 2021 season. However, official invitation to tender posted on the FIA the media were aware of a major change to the until now the rules themselves have largely website. The tender is for the Formula 1 tyre technical regulations before the teams were. Ibeen a closely guarded secret. But the veil was supply contract from 2020-2023 (inclusive) and ‘We were not aware of that,’ Paddy Lowe of partially lifted at the German Grand Prix where in the documents uploaded to the website there Williams said after the tender document was the first solid elements of the new regulations is an unprecedented level of detail about what shown to him. ‘I know the idea of 18in wheels were revealed. These were that in 2021 F1 will the FIA wants from the new F1 tyres. has been debated many, many times over the adopt 18in wheels, low profile tyres, and it will At the pre-race press conference in Germany last 10 or more years. So, it’s an interesting thing also ban the use of tyre warmers. none of the four technical directors present to make a commitment to that because it’s This change has substantial implications (from Williams, Renault, Red Bull and Force not absolutely clear that’s a great way forward, for the overall car design in 2021 and the India) were actually aware that the decision to and I think we need to analyse the implications

34 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 34 14/12/2018 11:14 Formula 1 tyres are to go low profile avoid, despite Isola’s claims of lighter front tyres. from 2021 to fit new 18in wheels, while ‘I think what will make a huge difference to the F1 will also be banning tyre warmers packaging inboard of the front of the car will be mainly what we do with uprights, brake drums and all the outboard kit, understanding what the tyres need, in terms of suspension kinematics, all of that,’ Bob Bell of Renault F1 says. One detail of the new regulations which is also included in these tender documents is that the mandated weight distribution will shift rearwards by three per cent, though the reason for this is not provided. Wonder wall A low profile tyre also gives teams a much smaller sidewall to work with, something which both complicates and simplifies some car development tasks. Currently the sidewalls of F1 tyres play a major role in the suspension system and that role will change substantially in 2021. ‘The influence on the inboard suspension will be higher in terms of tyre deflection,’ Red Bull Racing technical director Pierre Wache says. ‘For sure, the spring will take more load. I think that it might give more control for the chassis people of the ride height of the car. So it is not so bad from that aspect, and the tyre deflection will have less of an influence on the aerodynamics of the car.’ Paddock rumours But what degree of influence the sidewall change will really have on the design and layout suggest that there of the suspension system on the 2021 cars is uncertain, as it is not yet clear if a conventional is interest from system will be used at all, as there are ongoing discussions around the reintroduction of both Michelin active suspension for 2021. If this happens then the system will likely be a common set of and Hankook components shared by all teams. Black art technically before going in that direction. 18in wheels and the front tyres will be sized at As Wache mentions, the reduction in sidewall Certainly, it makes a very different tyre. A much 270/D1-18, the rears 405/D2-18 (D1 and D2 are deformation will have a major impact on the heavier package as well, and quite challenging two uncertain dimensions, left up to the tyre aero development of the cars. Currently every to design and manufacture.’ makers but which must be ranged between team is struggling to accurately model the Switching to 18in wheels is by no means a 700mm and 720mm). Notably, this means that behaviour of the Pirelli product, especially in the new discussion in F1, as Lowe suggested it is the front tyres will be 35mm narrower than wind tunnel, and the lower profile tyres could something which has been under consideration present but have a slightly larger diameter. The well make that task a lot easier. for years, with tests of prototype wheels and rear tyres will also have a larger diameter than ‘Wind tunnel tyres are a real black art, to try tyres taking place on both F2 (GP2) and F1 cars. at present but will remain the same width. This and replicate reality is very hard,’ Force India means in terms of tyre weight alone that the technical director Andrew Green says. ‘You end Tyre change rears will be heavier, while the fronts may be up in a situation where you can try and replicate Developing the new low profile tyre for 18in slightly lighter than the 2018 rubber. a condition on the track, say a high speed wheels will fall to the nominated tyre supplier, ‘The thing which is really important is the corner, you can have a wind tunnel tyre that currently Pirelli, but this could well change for volume of air inside the tyre because [of] the replicates that, but it is no good for replicating 2020. At the time of writing the Italian company variance in the loading capacity of the tyre,’ a low speed corner. You are always stuck with had not decided if it would put in a tender or Pirelli racing manager Mario Isola says. ‘The the compromise of which tyre do you develop not, while paddock rumours suggest that there front tyre will be narrower than now, while the around. So we have to be quite smart about that is also interest from both Michelin and Hankook. rear is the same width. Don’t forget we do 18in and know where the limits of the tunnel tyres The change of tyre shape and size is something tyres for sportscar racing, and we did it for F2 in are at, where it is reporting good correlation, that the rule makers feel will increase road the past as well as doing the prototype tyre for and where it is reporting bad correlation, and relevance, in theory, but in reality this has been show and testing with a Lotus F1 car, so we have make sure we are not developing around done more for aesthetic reasons, and it’s known investigated this to some extent.’ something which does not exist. Pirelli give us that this is something that Michelin favours. The changes are of particular interest to car updates through the year to mitigate that but The tender document states that from the designers, with an increase in unsprung weight it’s a real challenge, especially as they are busy start of the 2021 season all cars will use larger inevitable, something they always strive to developing the full size tyres too.’

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 35

RCE F1 digi mag.indb 35 14/12/2018 11:14 FORMULA 1 – 2021 REGULATIONS

‘We don’t know what the rules around the brake disc size will be yet’

Modelling the real tyre is a major job for all teams not just in terms of physical shape but also thermal performance, and it is something that may be eased in 2021 simply because with a lower profile there will be less deflection. ‘We have an FEA model provided to us by Pirelli, and we use that as a basis then modify and tweak it to match the data that we see on the real car,’ Green says. ‘We do on-car measurements of the tyre behaviour, feed that to the tyre modeller, then that into the FE model, and finally those shapes go into CFD and that is when we start looking at where the wind tunnel tyres are deficient and when we have to turn the attention to CFD. You start getting creases and all sorts as the tyre changes shape. Pirelli tested 18in wheels fitted with low profile tyres on a GP2 car during a demonstration run at Monaco three years ago ‘We have to look at sidewall deformation, the way the contact patch changes shape at high speed, low speed, medium speed, and what it is doing on the inside wheel and outside wheel, front and rear,’ Green adds. ‘At times some of the tyres are barely touching the ground as they go round corners, other times they are planted so hard you think that they are about to pop. Steer, slip, roll, all of that factors in as well and it’s an incredibly complex equation and you can get yourself lost very quickly if you are not incredibly clear about what you are trying to achieve.’ Big brake However, before Pirelli or any other organisation can even start to consider the design of the new tyres it needs to understand more about other areas of the car; not only the weight distribution but also aerodynamic performance as well as some mechanical parts, and key among them is the braking system. When the 18in wheels were originally tested by Lotus the brake system Tyres are already mounted to 18in rims in LMP1. Low profile tyres are said to be more road relevant than current F1 rubber carried over from the current 13in rims, but there is great uncertainty over what the 2021 rules will include in terms of brake disc size. ‘I would welcome bigger brakes,’ Bell says. ‘I think for the new formula we will need them as we are getting close to the practical limits with the current brakes. I think it does represent an opportunity to redress some capacity in the braking system.’ Disc discussions But to date it seems that the group developing the new rules has yet to turn its attention to the braking system, and at the time of writing had yet to consult with the brake manufacturers about what the new rules could bring, and there have even been suggestions that the current brake regulations will carry over to 2021. ‘We don’t know what the regulations around the brake disc size will be yet,’ Andrea Pellegrini, Brembo’s F1 brake engineer, says. ‘We expect to talk with the FIA about the rules, I know there Mario Isola, Pirelli racing manager, alongside an 18in low profile tyre. These were fitted to a Lotus to test the idea in 2014 are some meetings coming up, as was the case

36 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 36 14/12/2018 11:14 ELECTRONIC THROTTLE BODIES MOTORSPORT ABS POWER BOX PBX90

CAS M ANTI”COLLISION RADAR MOTORSPORT MOBILE MOTORSPORT IGNITION COILS WHEEL ALIGNMENT

MOTORSPORT FUEL PUMPS MOTORSPORT SENSORS PROGRAMMABLE WIPER MOTORS

Motorsport Systems Ltd, are a UK based dealer of Bosch Motorsport products. With an innate understanding of the motorsport industry, we specialise in delivering a personal service that gets your parts to you, on budget and on time. For further information on the products detailed above, please do not hesitate to call or email [email protected].

01684 291122 www.motorsport-systems.co.uk

MOTORSPORT SYSTEMS FP RC DEC18.indd 1 24/10/2018 12:30 Advert Grid 230x300 AW.indd 1 12/09/2016 11:00 FORMULA 1 – 2021 REGULATIONS

with the thickness change for 2017, but we have not spoken to them yet. With the big rims we can increase the diameter of the discs, so it would be like GT or LMP1 where they have a similar rim size. I personally think that the diameter of the disc will increase, in fact I hope it does because from my point of view the big rim and a small disc would not look good.’ Science friction There are a number of variables which will need to be decided in terms of the brake system design before any serious work on the tyres can begin, as Pellegrini highlights. ‘In terms of the friction material, nothing will change. The discs we use on LMP1 cars are the exact same material we use in Formula 1, we change just the diameter, the thickness and the cooling. The rules relating to brake systems have yet to be announced yet the brakes play a key role in the performance of the tyre But these are options to consider; you could have a bigger diameter disc, but to keep the weight down maybe it would not be as thick as now, perhaps back to 28mm thickness or 30mm. The radius would increase with a bigger disc so you can reach more torque without any problem. The current discs are on the limit because we cannot increase the diameter, and 32mm is quite a thick disc. With the current regulations the braking torque is really high, and in terms of safety it would be better to have a bit of margin. If there was a significant performance increase it could be a problem.’ Brake time However, time could be an issue in terms of developing new brakes, as much work will be needed to be done by the brake manufacturers to be ready for 2021. ‘For us it is really important LMP brake discs have a larger diameter than those used in F1; it is possible the latter might use similar discs from 2021 to know the regulations in advance as the work required to get the discs ready in time is very long,’ Pellegrini says. ‘It takes around eight or nine months just for the manufacturing. There is also the cooling design, and the calipers too. The sooner we get the rules the better. ‘To give you an example, for the 2017 season the brake disc thickness increased from 28 to 32mm,’ Pellegrini adds. ‘This was done because the energy of the cars with the new tyres and increased downforce would have been too much for a pre-2017 brake disc, so the only way to increase the performance was to increase the thickness as we could not increase the diameter then as the discs are already as close as possible to the wheel rim. That change may seem from outside like it was not such a big deal, but in truth we had to design the caliper from scratch. That small change was a big job. Changing the diameter means you have to redesign the upright, the disc, the pads, the calipers, everything, so it is an incredibly big change. Just how much a current F1 car leans on its tyres during hard cornering is clear in this image of the McLaren at Silverstone Just take the disc cooling; the extra 4mm in Currently the sidewalls of F1 tyres play a major role in the suspension system and that role will change substantially

38 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 38 14/12/2018 11:14 Primary Designs are leaders in the design and manufacture of world-class high performance exhausts

Suppliers to championship winning Formula 1 fabrication of thin wall exotic alloys makes teams and world record breaking supercar us the ideal partner to design and build manufacturers. Our knowledge of exhaust design, high-performance racing exhausts for an alongside our expertise in the welding and extensive range of motorsport applications.

Tel: +44 (0)1844 216 057 Web: www.primarydesigns.co.uk Email: [email protected]

WE CONNECT TECHNOLOGY

VAD FOR SOURIAU 8STA CONNECTORS ULTRA MINIATURE HARSH ENVIRONMENT CONNECTORS MIXED POWER AND SIGNAL CONNECTORS

HERMETIC CONNECTORS

DEDICATED ASSEMBLY TOOLING

FUEL PROOF HEAT SHRINK BOOTS

ULTRA LIGHTWEIGHT ACCESSORIES

FUEL PROOF HEAT SHRINK TUBES

+44 (0) 1403 790 661 10TH-11TH JAN 2019 STAND NO. E760 lanemotorsport.com come and see us [email protected]

ONLINE SHOP CONNECTORS ACCESSORIES HEAT SHRINK BOOTS FOR AFTER HOURS ORDERING FORMULA 1 – 2021 REGULATIONS

thickness meant that we went from 1000 holes Talking of tyre temperature, this is another This change will require a completely new to 1450 holes, and that was a major project. That major consideration for 2021, with tyre warmers approach to the tyre compound development, would be a bigger challenge for a bigger disc.’ to be banned from that season on. This is according to Isola: ‘It is really difficult, as Brembo and the other brake manufacturers actually not something that is entirely unknown without blankets we will have to change all the also need to feed information on the disc design in motor racing, as Super GT, for example, has compounds,’ he says. ‘Right now the cars leave to the teams well ahead of time so they can a total ban on tyre warmers, while the WEC the garage with the tyres at 110degC but now design the cooling package around them, and has a ban on tyre blankets but uses tyre if they roll out with the tyres at 20degC it is a the size of the disc could have a fundamental warming cabinets. However, for Formula 1 completely different challenge. Then, additional impact on how that package might look. the challenge could be quite significant. to that, we have to design a number of ‘The cooling systems around the brake discs compounds to meet the tender. It will be more are very sophisticated these days. Originally Blanket ban than the four specified by the FIA, it will these were just used to control the temperature Isola certainly thinks it will be: ‘The pressure be six at least and we have to be able to deliver of the caliper and disc, but now there are a evolution is going to be important, with certain specified performance gaps at 20 or 21 lot of things being done to warm up the rims, the tyres starting cold, the temperature and different circuits. That is not easy.’ for example, a lot of things are done for tyre pressure difference will be a lot bigger,’ he temperature,’ Pellegrini says. says. ‘You can’t just start the tyres at 5psi cold Cold comfort because this is F1 and there has to be a certain While some have questioned the value of this performance level. It will be important to have a change, it has been met with cautious optimism ‘If it drives us towards minimum pressure which is enough for the car by others in F1. ‘If it drives us towards tyres that in the first few laps. We need to understand how have a much wider window to operate in, that tyres that have a much the tyres will work and how the temperatures could be a good thing,’ Lowe says. ‘I know at and pressures will evolve. It is not easy to make the same time they are talking about moving wider window to operate a tyre like that, but it is a technical challenge. qualifying perhaps to formats where there are We do it in F2, they start from cold, but the less laps, more criticality around doing single in, that could be good’ performance level of F2 is much lower than F1.’ laps and, again, if that’s around tyres that aren’t prepared with blankets that would drive us towards tyres with a wider window, then I think it would be a good thing for the sport.’ The tender (see box out) also specifies that there must be three different tyre compounds available at each race; a hard, medium and soft. One of the requirements of the tender document is that the performance gaps between these compounds will be significantly larger than they are at present and the degradation will be higher. The hard tyre will lose about two seconds a lap in terms of performance through degradation after 22 per cent of the race distance; the medium tyre will be around 1.2 seconds a lap faster than the hard but will lose two seconds of performance Tyre warmers will be banned for 2021 bringing challenges for both the chosen tyre manufacturer and the Formula 1 teams after 18 per cent of the race distance; while the soft will be the quickest tyre, around 2.2 seconds a lap faster than the hard but will lose two seconds a lap in performance after 10 per cent of the race distance. This tyre performance is expected at 75 per cent of the races, as it is clear that it will not be possible at every race circuit. This is all intended to force the F1 teams to move away from a one stop strategy and towards two and three stop races, something the rule makers believe will ‘improve the show’. On the edge Additionally, the degradation of the tyre will be non-linear, so the tyres will be designed to fall off the cliff part way through the degradation. It is suggested that an under-layer of a lower performance compound is designed below the main tread compound to achieve this. The tyres will also be required to quickly recover performance after a period of following another car closely or some aggressive driving over a lap. The tender document states: ‘It is anticipated The tyre supplier will need to provide three different compounds for each race, with large performance gaps between them that to achieve this, for any given compound,

40 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 40 14/12/2018 11:14 ‘The degradation of the tyre will be non-linear, so the tyres will be designed to fall off the cliff part way through the process’

the deviation of grip under a particular favours Pirelli as it already has a product at for the following year, and we track tested in condition of surface macro and micro roughness this size. Though the tyres will have to comply August. It’s not a problem for a tyre maker to and at a typical slip velocity will not drop by with the new performance criteria, in 2020 tyre react like that, it’s unfortunate but in Formula 1 more than five per cent in a temperature range warmers will still be used. everything is always quick.’ of plus/minus 15degC from that which achieves But Isola says of this: ’For 2020 we have to peak grip. Furthermore the cross linking of the design a new tyre anyway, so it is no advantage Race against time compound polymers must be robust enough for us to already have a product. We have a Quick or not, Formula 1 still faces a race to resist permanent damage due to high slip or new tyre every year, there has not been a single against time to get everything ready in time surface temperature conditions.’ season where we have not done a new tyre. for 2021, with many areas of the regulations As the contract also covers the 2020 season ‘The tight deadlines should not be an issue still undecided and a number of deadlines the winning tyre maker will also have to supply for us as it is the same as it was for us when fast approaching. By the time this is published high profile tyres to suit 13in rims for a single we came to Formula 1 in 2011,’ Isola adds. ‘We the initial selection of the tyre supplier will Formula 1 season, something some have said were only appointed at the end of June 2010 have been made subject to commercial arrangements, but the details the tyre makers will require are not likely to have been finalised. ‘We need additional details. We need to know what the brakes are, what the downforce levels are, anything that will help us start designing the tyre,’ Isola says. ‘Right now with no information it’s quite difficult. Ideally, it is already too late for 2021, we need at least a draft set of technical regulations which are not too far from the final rules really by the time we would be fully appointed, which will be September or October time, once the commercial deals are done. Then we will have to start immediately as it is not only the tyres we have to design. We will have to upgrade the machinery in the factory, we need to change our indoor test equipment. There are a lot of other minor factors which will result, too. When they changed to the wider In 2016 Pirelli used ‘mules’ to test the 2017 rubber, but it seems unlikely there will be a suitable test car for the 2021 tyres tyres in 2017, we had to increase the number of trucks we have and the number of flight Tyre performance targets cases, because you cannot fit the same number of tyres in the trucks if the tyres are bigger, we need to consider all of those things.’ • Tyre stiffness should vary above and below optimum at two per cent with respect to monotonically with working which grip is reduced by five normal loads encountered Testing headache range. In addition tyres with per cent on typical micro and during operation and must When Pirelli became the sole F1 tyre supplier in the highest compound stiffness macro roughness tracks and maintain these limits within the 2011 it tested its products extensively using an should also have the highest sliding velocities. normal variations in tyre wear. unbranded Toyota TF109, and ahead of 2017 it working temperature range and used a number of specially adapted mule cars the tyre temperature working • For 2021 the ratio of cornering • Peak cornering force in low from various teams to try out the new wider range should reduce as the tyre stiffness of the tyres should be speed corners should be rubber. For 2021 the problem is somewhat more compound stiffness reduces. compatible with a rearward achieved at six to seven degrees challenging with no car available at all which The variation in working range shift of the longitudinal centre of slip angle on the rear tyres will be fully representative of the new rules. between the compounds of gravity position of three and in high speed corners at ‘It’s like 2016, when we had mule cars, we should not be excessive. per cent from current values. around eight degrees. Peak need cars that are designed to test 18in tyres, Suggested values are shown Furthermore the change in cornering force on the front but I’m not sure we will get that,’ Isola says. ‘A below where ‘working range’ cornering stiffness ratio must tyres should be achieved at NASCAR style test programme would be the is defined as the temperatures remain a constant plus/minus slightly lower slip angles. ideal, we have it in GP2 and GP3 where the organiser has a test car with a small team and Variations in working range • Tyres should provide safe we can test when we need to, and that gives performance when leaving the no advantage to any team. But in F1 it seems Compound Optimum bulk temperature Working range pits cold. The glass transition unrealistic to have a test team just for the tyres.’ A (hardest) (x + 10 to 15)ºC temperature must be chosen The first data from the new tyre supplier B (x + 5 to 10 ºC Optimum bulk temperature so that the tyres are never in is expected to be shared with teams just two C x ºC a ‘glassy state’ when either weeks after the contract is finalised, with the +– 15 ºC D (x – 5 to 10) ºC the ambient or the track first rig testing tyres for the 2020 season E (softest) (x – 10 to 15) ºC temperature is above 10degC. delivered to teams on 1 January 2019.

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 41

RCE F1 digi mag.indb 41 14/12/2018 11:14 FORMULA 1 – COOLING

Hot and bothered Solving the problems of cooling a hybrid-powered car has been one of the major challenges of the current Formula 1 era, but thanks to genuine F1 team blueprints Racecar has been able to piece together the complex thinking, and plumbing, that goes into the design of these systems By SAMUEL COLLINS

f you walked around to the bin stores at any defunct Caterham F1 team, which detail the strict thermal control. The complete power F1 team’s factory (and a few WEC facilities development of the 2015 Caterham CT06. This is unit consists of the internal combustion engine too) in late 2013 or at any point in 2014 you the fi rst time such detailed information has ever (ICE) and the hybrid system. The hybrid system would have found the metal waste hoppers been published on a modern grand prix car. is made up of a number of core components; Ioverfl owing with scrapped radiator cores. It When you look at any competition car (or the motor generator unit-heat (MGU-H) and was a sign of the biggest shift in grand prix car indeed any car) you will quickly note that there the motor generator unit-kinetic (MGU-K), an cooling system design in decades, if not ever. are several sources of heat and those all need to energy storage (ES) as well as control units for Today, cooling system layout and be cooled to some extent. All of these sources each (CU-H, CU-K and CU-ES). optimisation remains a crucial part of the can be gathered into two main categories; those design of every single car on the grid and is a that only need to be cooled, and those that not ICE cool constant area of study. The bins at the back of only have to be cooled but also have to have The main source of heat the cooling system the F1 teams factories are still quite full, if not their temperature controlled and maintained has to manage is the internal combustion overfl owing as they were four years ago. This between the proper operational boundaries. engine. The cooling of the engine is against the article looks at the cause of those overly full bins One source of heat that does not need thermodynamic effi ciency; the cooling system – the process of the design, development and an accurate control of its temperature is the extracts heat from the combustion chamber optimisation of a current F1 car’s cooling system. racecar’s braking system. This system only that could be used to increase the pressure It draws on two main sources of information. needs to be cooled to keep its temperature within the cylinder to extract more work from Numerical Simulation of a 2018 F1 Car Cooling below the operational limit of the materials in the engine. However, the cooling system is System for Silverstone Circuit is a technical paper which the diff erent components of the system mandatory in an internal combustion engine by Victor Tizon Otero and Stephen Samuel of have been manufactured. due to the limits that materials composing the Oxford Brookes University, which was presented But falling entirely within the category of a internals of the engine have regarding higher at the WCX SAE World Congress Experience in source of heat which needs its temperature to temperature. Above certain temperatures, these early 2018. The second source of information be controlled is a Formula 1 power unit. Every components would fail, so their temperature is a set of internal design documents from the single component needs to be kept under has to be kept under safe conditions.

42 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 42 14/12/2018 11:14 While the cooling Figure 1: The fi rst step was to map out the expected cooling demands over a season, including the extreme temperatures demands for a combustion engine are well understood, those for a hybrid system are far more complex

In addition, oil loses its lubricating properties when the temperature is higher than around 175degC, thus increasing wear of engine components and maybe leading to premature failure of the engine. DLC (diamond-like carbon) Figure 2: The average conditions were also analysed. Both sets of data were then used to create the cooling specifi cations coatings and other treatments play a role here in reducing friction and, in turn, heat. Another area addition or rejection quantity. So one of the fi tted a reciprocating combustion engine to a where DLCs help is in the transmission – which fi rst challenges is to model the associated parts car in 1886. That said, a challenge not quite as also has a cooling demand, though signifi cantly and components in order to assess the thermal old, and very new to motor racing, is cooling lower than the power unit components. performance of the complete vehicle. a hybrid system alongside the combustion As mentioned, the main source of heat for engine. This was the reason behind the piles of Cool for CADs the cooling system is the 1.6-lire V6 engine. The experimental cooler cores found in those bins The design of a cooling system for a Formula 1 heat transfer takes place via two main fl uids. The mentioned at the start of this piece. car is a demanding task that has to be carried main coolant is the water, the fl uid in charge of With many more sub-systems requiring out following all the regulations that govern the removing excess heat from the cylinder and the cooling than before one of the fi rst challenges sport. In addition to the specifi c regulations for head. The second fl uid is the oil, used mainly to that the Formula 1 teams had to understand in the cooling package which place restrictions on lubricate engine components like the pistons, the run up to the introduction of the new power some materials, for example, its design has also the camshaft and the crankshaft, but also to units in 2014 was actually working out what the to take into account the regulations of other remove heat from pistons, the turbocharger and cooling demands really were. components of the car since they interact and other engine components. Additionally, the fuel For Caterham, the starting point for this was transfer thermal energy across the systems. used also has a cooling role. analysing the weather data from the race circuits For instance, the maximum pressure a pump But this challenge is nothing new and that are visited by Formula 1 over the course of a can deliver or the maximum energy that the combustion engine cooling systems have been season, with the weather extremes mapped out battery can hold can change the overall heat refi ned and optimised ever since Mr Benz fi rst (Figure 1) along with the average conditions

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 43

RCE F1 digi mag.indb 43 14/12/2018 11:14 FORMULA 1 – COOLING

This is the first time such detailed information has ever been published on a modern grand prix car

(Figure 2). This data was then used to create both a maximum cooling specification and a baseline cooling specification, as well as the Figure 3: Flow rates. T-air is ambient temperature, W is flow rate required by water coolers, O for oil coolers and E for ERS bodywork options for the racecar. A design report on the car states: ’For the maximum cooling specification one has to look at the extreme weather predictions as well as the average and consider the likeliness of the extreme weather to be occurring and whether there are other ways of reducing temperature (at the cost of performance obviously). Once established, these massflow numbers can be used to size the radiators needed (based on expected maximum possible massflows from an aero point of view). For the baseline cooling specification one needs to consider on how many circuits one does accept an aero penalty for average conditions.’

Packing heat Figure 4: In first ERS cooling circuit devised by Caterham the small central cooler would feed the control unit for the MGU-K Another factor which Formula 1 teams have to consider is the maximum operational temperature of various components. A typical 2015 Formula 1 power unit had the following maximum allowable temperatures for the various components in the power unit. The control units ranged from 60degC to 75degC, the MGUs around 85degC, and the energy store was between 73degC to 85degC. Additionally, the fluid temperatures in various systems are often defined by the power unit supplier. This is all factored in to allow for a cooling circuit (or indeed circuits) to be laid out and the various coolers sized and positioned in the car, as well as the various airflow requirements for coolers within the system. Figure 5: Team looked at integrating energy store cooling into the MGU-H circuit and making greater use of centreline cooler Those flow rates can be seen in Figure 3. Here the extreme conditions are shown in red and the average conditions are in blue. Based on the maximum flow rate for water The final Caterham CT06 at Singapore (1.01kg/s @ 32degC) and the cooler layout. Two charge Hungaroring (1.07kg/s @ 36degC) a maximum air coolers with ERS and oil flow rate for the water coolers was set at and water coolers are on the 1.02kg/s, similarly for oil (0.71kg/s at Singapore left side, water cooler on the and 0.75 kg/s at Hungaroring) a maximum right. The gearbox cooler is of 0.72kg/s was set. In terms of the ERS the above the turbocharger maximum was set at 0.47kg/s. While the cooling demands for a combustion engine are well understood, the demands of a hybrid system are far more complex and required a lot more investigation than would

44 www.racecar-engineering.com FORMULA 1 · WINTER 2018

RCE F1 digi mag.indb 44 14/12/2018 11:14 usually be conducted by a Formula 1 team. As a result, one of the first cooling circuits to be laid out was the ERS cooling, and this was to utilise a heat exchanger in the left hand side sidepod as well as a smaller centreline cooler. The sidepod duct circuit would cool the energy store, both MGUs and the control electronics for the MGU-H. The small central cooler would feed the control unit for the MGU-K (Figure 4). Chill out However, this layout was found to place a high demand on the left hand side radiator, so to overcome this the team looked at integrating the energy store cooling into the MGU-H cooling circuit and make greater use of the centreline cooler (Figure 5). But simulation suggested that while this layout would likely work for the average circuit temperature demands (Figure 3) it would struggle to meet the Figure 6: A second pump was added to serve the CU-K and this was to be fed via a cooler on the centreline of the racecar demands in the extreme conditions, it would also be a struggle to get adequate cooling from the centreline cooler. The issue was felt to be the low temperature requirement of the MGU-K control unit, which essentially caused a bottleneck in the system. To get around this a second pump was added to the system, primarily to serve the CU-K, and would be fed via a cooler on the centreline of the racecar – the additional pump would also have the benefit of reducing the airflow requirement of the centreline cooler (Figure 6). This layout would allow the CU-K to run at a cooler temperature while the other components would be able to run hotter, reducing the size of cooling Figure 7: The single pump would have been workable but the demands of the oil system and ERS cooling counted against it apertures across the racecar. Pump action However, there was still interest in a single pump layout and both the twin and single pump layouts were re-evaluated at a higher ambient temperature, but it was found that while the single pump solution would be workable the demands of the oil system and the ERS cooling could create a significant aerodynamic penalty (Figure 7). The double pump layout was found to still require less air flow into the centreline cooling duct, but made little difference in terms of the size of the sidepod duct and ultimately this layout, the same as Figure 6, was adopted. Once the layout had been confirmed the attention then shifted on to the coolers themselves. A set of design targets were established, not only based on the pure cooling demands of the power unit but also on the racecar’s overall aerodynamic goals. Additionally, the weight of each of the Airflow under the body and through sidepods was calculated in CFD to ensure adequate flow and a good level of efficiency components was also tightly monitored. The issue was the low temperature requirement of the MGU-K control unit, which essentially caused a bottleneck in the system

FORMULA 1 · WINTER 2018 www.racecar-engineering.com 45

RCE F1 digi mag.indb 45 14/12/2018 11:14 FORMULA 1 – COOLING

The final design called for the mounting of a Perhaps the most complex layout on double pass charge air cooler in the left sidepod display in 2018 is that of the Sauber C37 (see along with a double pass oil cooler and the ERS Racecar Engineering June 2018, V28N6), which cooler. In the right sidepod a second charge air is equipped with a multitude of different cooler would be installed (though of different circuits and coolers. This has been done for one dimensions to the one on the opposite side) overriding reason – aerodynamic performance. along with the engine water cooler. Indeed, the cooling system is a major challenge The transmission cooler would be mounted for the aerodynamic performance of the car above the turbocharger and fed from the and it is constantly adjusted and optimised. It centreline of the car. After a substantial amount is also one of the few areas of an F1 car which of optimisation of the coolers the weight of can be developed outside of the perimeter of the entire cooling system for the car was restricted aerodynamic testing, so the amount brought down to 16kg. The Caterham CT05 had of development is substantial. a cooling system weight of over 30kg, so the 2015 Caterham would have seen a substantial Playing it cool saving, had it been completed. The use of thermal barrier coatings on the inner surfaces of the car bodywork has seen the cars Out in the cold almost shrink-wrapped around the power unit But that’s the sad part. Ultimately, how effective at the rear, and climate specific aerodynamic this system would have been was never packages developed. But the teams still struggle determined as the CT06 was never constructed to get the heat out, scorch marks still appear because the team collapsed during the 2014 on bodywork and extra ducts and outlets have A set of cooler design targets was established and this was not season. It is fair to say, however, that the to be added at hotter circuits. All of which only based on the pure cooling demands of the power unit but complexity of cooling systems has increased create an aerodynamic penalty for the overall also on the racecar’s overall aerodynamic goals. Meanwhile substantially since 2015, with a significant car. So the never ending cycle continues, with the weight of each component was also monitored (above) shift toward centreline cooling, largely for testing optimisation and development, and all aerodynamic reasons, and with more coolers the while cooling systems offer a significant and elements added to the systems. potential performance gain. The use of thermal barrier coatings on the inner surfaces of the body

hasSST Advert seen half page the FIA.qxp_Layout cars almost 1 28/02/2018 shrink 14:17 Page wrapped 1 around the PU at the rear

at AutosportVisit us Stand E482

Design and manufacture of high performance exhausts, precision fabrications and RADIATORS AND HEAT EXHAUSTS AND PRECISION EXCHANGERS thermal management solutions FABRICATIONS

SST are the only UK manufacturer of the Additional Frontal SST Technology SST-DPS-ProTi Protection Halo AFP.003.18-T TITANIUM DRIVER (AFP-Halo) 8869-2018 PROTECTION SYSTEM

SUPPLIERS TO F1, INDYCAR, TOURING AND SPORTSCAR, RALLY AND PERFORMANCE AUTOMOTIVE SECTORS SST Technology Oxfordshire T: +44 (0) 1865 731018 E: [email protected] www.sstubetechnology.com

F1 Supp_Cooling_MBAC.indd 46 17/12/2018 10:27 Oil coolers and Intercooler cores

Reliability Flexibility Know-how Visit us at www.setrab.com/proline SUBSCRIBE TO MOTORSPORT’S LEADING TECHNOLOGY MAGAZINE

Get the latest product developments, technical insight and performance analysis from the leading motorsport technology publication every month.

INSIDE EACH ISSUE • Unrivalled analysis of major events including Formula One, the World Rally Championship and the World Endurance Championship • Detailed racecar design and innovation • Features from leading industry engineers print or and the latest product digital development access

TO ORDER www.chelseamagazines.com/CRCEF1D8 +44 (0) 1858 438 443 (quote code CRCEF1D8) AVAILABLE WORLDWIDE

DIGITAL EDITION ALSO AVAILABLE WORLDWIDE for tablet, smartphone and desktop. To order go to: www.chelseamagazines.com/racecar-CRC1F1D8

CRCEF1D8.indd 1 17/12/2018 14:44