Indycar Grand Prix

15 May 2017

INDIANAPOLIS, Indiana – Team Penske’s Will Power ended the weekend the same way he started it: in first place.

The driver of the No. 12 Verizon Chevrolet capped off a flawless weekend at the INDYCAR Grand Prix with a dominating win today on the Indianapolis Motor Speedway road course. After winning Friday’s Verizon P1 Award with a new track record in qualifying and leading all three practice sessions on the weekend, Power led 61 laps en route to his 30th Indy car win, beating Chip Ganassi Racing’s Scott Dixon (No. 9 NTT Data Honda) by 5.2830 seconds in a caution-free affair, the first such race since Long Beach in 2016.

Power’s win – which came in his 175th Indy car start – breaks a three-way tie with Rick Mears and Helio Castroneves for 11th on the all-time win list. It is Power’s second victory on the IMS 2.439-mile road course and the 190th Indy car win in the history of Team Penske. For Dixon, it was his 35th career runner-up finish, fourth all-time, coming on the day he passed Jimmy Vasser for second on the all-time list with his 212th consecutive start. Teammate Tony Kanaan (No. 10 NTT Data Honda) holds that title with 270 consecutive starts. Andretti Autosport’s Ryan Hunter-Reay (No. 28 DHL Honda) finished third for his first podium of 2017, followed by reigning series champion Simon Pagenaud (No. 1 Menards Team Penske Chevrolet) and Castroneves (No. 3 Verizon Team Penske Chevrolet).

Pagenaud and Dixon remain the only two drivers to score top-fives in every race of 2017. Graham Rahal (No. 15 Soldier Strong / Turns for Honda) and Max Chilton (No. 7 Gallagher Honda) came home with much needed sixth- and seventh-place finishes, respectively. Rahal advanced from a 20th-place starting spot and Chilton equaled his career-best finish.

There is little rest for Verizon IndyCar Series teams, as on-track activities for the 101st Indianapolis 500 presented by PennGrade Motor Oil begin in two days. Opening Day practice on the Indianapolis Motor Speedway’s hallowed 2.5-mile oval begins at noon ET Monday.

Technology - Design Simulation

05 May 2017

Technology – Design Simulation

Fluid situation

These days there’s much more to CFD than just colourful spaghetti, as Racecar discovered when it talked to those at the cutting edge of this exciting design technology.

To simulate the behaviour within an engine accurately, the CFD has to overcome more complex challenges than those found in general aerodynamics.

The Sahara Force India F1 Team filed a request to the FIA to modify the Sporting Regulations in favour of CFD last year. “What we have today is a wind tunnel-biased formula, and what we are looking for is a CFD-biased formula”, explained Bob Fenley, its deputy team principal. “We believe that in the foreseeable future – in three years or a bit more – there is a possibility that CFD could become the primary or only aero programme. “Although this was a strategic move to save millions on renting a wind tunnel, it does also demonstrate the importance of CFD. For a team who finished fourth last year to instigate F1’s route to virtual aero development highlights just how realistic modern CFD software is becoming across all areas of the car.

We are all familiar with the mesmerising images of multi-coloured streamlines surrounding a Racecar. What we don’t often realise is that CFD can also be used to optimise internal flows within an engine to increase its performance and efficiency. “Essentially, we use CFD to model the gas exchange and combustion processes within our direct injection engines”, explains Ian Whiteside, chief engineer of Advanced Projects at Ilmor Engineering. “For example, we investigate the amount of turbulence generated in the combustion chamber, the interaction and the distribution of the air and fuel mixture and the volumetric efficiency – all with the aim of optimising the combustion process and therefore generating more engine power”.

Ilmor is renowned for its high performance engines that have raced in an impressive array of championships including F1, Indycar, NASCAR and even the WRC. To maintain its competitive edge, Ilmor is utilising the advanced Converge CFD software to not only improve its technologies but also to increase the efficiency of its development process. This is exactly what has been achieved, as Ilmor reduced its development time by 50 per cent which effectively saved eight weeks on the design and manufacture of the 2016 Indycar engine whilst reducing prototype build costs by 75 per cent.

To simulate the behaviour within an engine accurately, the CFD has to overcome more complex challenges than those found in general aerodynamics. “Firstly, the software has to model the movement of components to complex geometries such as pistons and valves”, explains Stephen Ferguson STAR-CC+ marketing director, Siemens PLM Software. “The second complication is simulating the complex physics of the combustion process and the reacting flows which needs to be modelled explicitly. You also have to determine the influence of the extreme heat generated on components and how to manage this by ultimately rejecting the heat to airflow in and around the car. In my opinion, engine simulation is one of the most difficult engineering problems that the motorsport industry has to face”.

One recent development in CFD technology which is providing invaluable for both aero and engine simulations is the concept of Adaptive Mesh Refinement (AMR). This is essentially automatic meshing which occurs during the run time of the simulation.

“For Ilmor, one of the main benefits of the Converge software is the automatic meshing” Whiteside says. “Our CFD experts determine the general grid size and define the areas where they want to refine the resolution of the mesh to capture complex flow physics. The rest is automated, which not only saves time from manually meshing, but also ensures that the same answer is achieved every run, regardless of who is controlling the simulations”.

The Converge software uses an orthogonal mesh that is composed of cubic elements generated from the triangulated approximation of surfaces found in standard CAD output STL files. AMR is also able to refine the mesh itself by subdividing the cubic elements in complex areas where the gradients are high. Therefore, the accuracy of the mesh can be greatly improved but only in areas where it is needed and only at the required times during the simulation. The benefits this gives are dramatic reductions in run time, enabling engineers to be more creative and investigate more ideas within the same time-frame.

For Siemens, highly automated meshing is one of the main reasons why its software is so successful within Formula 1. “F1 is leading the way in terms of automation”, Ferguson says. “Due to the hundreds of simulations run every day, the workflow has to be highly automated to ensure efficiency, so that highly trained engineers can focus their attention on analysing results rather than meshing, which is what our tools aim to achieve. The whole process of sourcing the CAD geometry, generating the mesh and making the simulation is best done automatically to ensure repeatability when comparing results”.

As well as automating a single run, additional optimisation tools such as HEEDS from Siemens allow you to automate a sweep of runs. For example, engineers define particular parameters such as a wing angle and HEEDS drives the STAR-CCM+programme through hundreds or thousands of configurations and then post-processes the results into a report. Therefore, once up and running, the first time an engineer is involved in the processes is right at the end when analysing the results.


Design filter

CFD is still mainly used as virtual filter to ensure that only the most effective designs are prototyped and tested. This is particularly true for engine simulations where CFD is utilised in a more fundamental way as dyno testing occurs much later in the development process and can be extremely expensive. “There are some subtle differences between the results from the engine simulations and the dyno but overall we have a lot of confidence in the CFD” says Whiteside. “The question is, can you ever achieve 100 per cent correlation? I think it’s unlikely and although CFD is a fantastic tool for the overall process, Ilmor are still a firm believer that the dyno is the ultimate decider”.

However, with the continuing advances in computing power and simulation efficiency, a fully predictive CFD future is not that far away and in some perspectives, is already here. “In F1, CFD is competing against wind tunnels and engine dynos”, Ferguson says. “Therefore, in practical terms, its accuracy only has to be as good as those test beds, so you could say that CFD tools are already fit for purpose. In general, CFD engineers are more interested in the resulting trends rather than actual values. Once they are satisfied with the accuracy of their CFD they can then go on to simulating scenarios that cannot be tested in real life. For instance, a few years ago, Renault F1 were using STAR-CCM+ to test how hot exhaust plumes influence the rear end aerodynamics, which you would never be able to replicate in a wind tunnel. As the accuracy of the models and solvers continue to increase, engineers can now simulate on-track conditions, overtaking and cornering. Although these might not be 100 per cent accurate, neither are any other engineering tools”.


CFD for all

CFD is not just for F1 teams, though, and it is becoming more widely available, largely thanks to the Cloud. Traditionally, if you wanted to simulate the fluid flow around a Racecar using CFD, you needed two things: 1) A license to use a software programme and 2) Access to high performance computing (HPC). Both of which are expensive, yet both of which can also now be avoided by using the Cloud.

“The whole process of purchasing a license, along with an advanced computer, and then installing the software is very time consuming and expensive” explains Milad Mafi, academic program manager from SimScale. “With SimScale, you simply log in on your web browser, upload your CAD STL file that you want to simulate and everything else is stored and computed within the Cloud. This means you are no longer restricted, because you can work on simulations from wherever you are and on whatever device you want – you just need an internet connection”.


Access all aeros

This flexibility and reduced cost has opened the doors for smaller software suppliers, as well as customers and the keen individual, because now everyone can afford to run a simulation. It is this accessibility which is making Cloud based simulation solutions so successful. “Our goal with SimScale was to make simulation accessible to everyone and you can clearly see the effectiveness of this strategy by the number of Formula Student projects we have on our platform” Mafi says. “These teams don’t have the budget of an F1 team and we found that their biggest problem was, despite having access to the software, they had a lack of hardware and knowledge. Therefore, together with the organisers of the competition, we created a four day workshop, teaching the teams how to conduct a fluid flow simulation and then how to utilise this to improve the aerodynamic efficiency of their race cars”.

Over 40 Formula Student teams currently utilise the SimScale platform, which also enables a team to collaboratively develop their design, thus improving the team’s efficiency.


Free flow

SimScale’s ethos of accessible simulation is not only achieved through the convenience of the Cloud but also the cost. For you or I to run a simulation, regardless of the complexity of the model or fluid behaviour, will cost us absolutely nothing. That’s right – it’s free and your work will be automatically uploaded to the SimScale library. Cleverly, this has not only resulted in a vast database with hundreds of examples for users to read but also stimulates further ideas for simulations. This perpetuating cycle means that whatever question you have or scenario you are trying to simulate, there is likely to be someone who has already solved your problem. In this way, the platform provides the answers, increasing your knowledge, which is another philosophy behind SimScale’s platform.

However, when in competition, race teams are unlikely to want their designs available with a Google search. Therefore, to ensure privacy, a professional subscription has to be paid. Regardless of your subscription type, everyone has access to up to 32 cores. A core is a processing unit which receives instructions and performs the necessary calculations; in CFD this is using equations to solve large linear algorithms. “To run a CFD simulation, it will take a total amount of computational time and this depends on the complexity of your simulation”, Mafi says. “For example, the higher the number of degrees of freedom or the finer the mesh, the longer the computational time. Let’s say your simulation will take one hour when using one core. If you use two cores, it will take half an hour and so on. Usually, a full F1 car sim using 32 cores can take up to eight to ten hours, at the end of the day it depends on your required accuracy and how fast you need the results”.

To run a full car simulation, imagine that the CAD model is encased in a rectangular box which is the fluid flow domain. On one side there is the inlet to the system and on the other there is the outlet, both of which require specific boundary conditions as well as the floor. The relative velocities of the floor and wheels then need to be considered to ensure the car is actually driving through the air and the wheels are rotating. Many additional complexities can be added, such as the effect of the exhaust flow. It depends on how accurate you need to be.

“Most people prefer to run half car simulations and apply symmetry boundary conditions to save time” Mafi says. “We have three ways of meshing: tetrahedral, hexahedral and then both, so if necessary you have control of every cell within the mesh to refine to your requirements. At the end of the day, all simulation users need to question whether they can trust the results. There is no point in saving money on simulation if you cannot make valid conclusions. SimScale uses the code based on an open source software called OpenFOAM which is used by the Mercedes and Sauber F1 teams. We have added our own GUI and integrated our own improvements”.


Clearly cloudy

Cloud based simulation solutions seem to solve many of the current issues found with traditional software programs. In addition to being accessible, it reduces costs, it is convenient and even avoids the need for annoying updates, downloads or restarts, because every time you refresh your web browser you are using the latest version.

But there are always sceptics when it comes to new technology and in this case data security is the main concern. Yet data transferred between your computer and the platform is done via industry standard SSL encryption technology. External data centres and hard drivers on the company’s servers are also encrypted by the same AES technology used to protect government documents.

Clearly, the Cloud is proving to be the next step in the evolution of simulation.

Racecar Engineering June 2017 by Gemma Hatton

Phoenix Raceway

02 May 2017

AVONDALE, Arizona – In his eighth Indy car season and 105th race, Team Penske’s Simon Pagenaud found victory lane on an oval.

By taking the checkered flag in the Desert Diamond West Valley Phoenix Grand Prix at Phoenix Raceway, Pagenaud (No. 1 Menards Team Penske Chevrolet) scored his 10th career Indy car win and the 189th in the illustrious history of Team Penske.

Pagenaud’s teammate Will Power (No. 12 Verizon Chevrolet) followed the Frenchman across the line by 9.1028 seconds, with Ed Carpenter Racing’s JR Hildebrand (No. 21 Fuzzy’s Vodka Chevrolet) rounding out the podium. With the win, Pagenaud jumps into the Verizon IndyCar Series points lead after four of 17 races. He and Chip Ganassi Racing’s Scott Dixon (No. 9 NTT Data Honda), who finished fifth in Saturday’s race and is now second in the title hunt, are the only two drivers to finish in the top five in each of the first four races.

Pagenaud’s win notched a pair of additional milestones for Team Penske. It was the team’s 450th race win in any motorsports form and the team’s 100th Indy car win on an oval. For Hildebrand, it was his first top-three finish since he was second in the 2011 Indy 500 and his sixth career top-five.

The Verizon IndyCar Series now turns the page to the month of May in Indianapolis. The INDYCAR Grand Prix is next up, the fourth Indy car race on the Indianapolis Motor Speedway road course. Coverage begins at 3:30 p.m. May 13 on ABC and the Advance Auto Parts INDYCAR Radio Network. Then it’s on to the 101st Indianapolis 500 on May 28.


The hidden valley: Ilmor takes the motorsport powertrain lead

29 March 2017

In deepest Northamptonshire, one of motorsport’s greatest names is blending traditional engineering knowhow with state of the art simulation techniques, Chris Pickering discovers

The Chevolet Indycar V6 engine is developed by Ilmor. Driving round the outskirts of Brixworth you wouldn’t necessarily know it was a place of global significance. But this unassuming corner of Northamptonshire marks the northern tip of Motorsport Valley – the strip of central and southern England that’s home to some 75 per cent of the world’s top-level motorsport R&D companies.

In many ways, Ilmor Engineering typifies this industry. Based in a quiet industrial estate on the eastern edge of the village, it’s responsible for engines that have dominated the likes of Formula One and IndyCar yet you’d hardly know it was there. It’s also evolving; embracing new techniques and moving into neighbouring industries such as aerospace and defence.

Ilmor can trace its roots back to another great UK motorsport institution. Company founders Paul Morgan and Mario Illien met while they were working as engineers at Cosworth, just down the road in Northampton. They hatched a plan to produce their own engines and founded Ilmor in partnership with US motorsport mogul Roger Penske in 1984.

 Initially, the company focused on the IndyCar series – a quintessentially American form of racing where British companies have had a defining impact ever since Lotus showed up in the 1960s. This remains a hugely important market for Ilmor, which produces the Chevrolet engines used in around half the current grid. Elsewhere, the company retains a significant presence in Formula 1.

Back in the late nineties it engineered the McLaren-Mercedes V10s that powered Mikka Hakinnen to his back-to-back F1 World Championships. This was the continuation of a fruitful partnership with Mercedes-Benz, which had begun with a phenomenally successful IndyCar engine earlier that decade. It culminated in 2005 when the German giant completed a buyout of the company and created Mercedes-Benz High Performance Engines. 

Paul Morgan, a keen collector of historic aircraft, was killed in 2001 when his Hawker Sea Fury overturned on landing. Following the buyout, however, his fellow co-founders bought the non-F1 parts of the business back from Mercedes, along with the Ilmor name. The two companies now face each other – literally – over the road, but they’re completely independent. They even compete against each other on the track, albeit indirectly with Ilmor carrying out behind the scenes work for a well-known F1 team.

Throw in a liberal sprinkling of NASCAR, World Rally Championship and GT racing projects and you have one of the most versatile – and active – motorsport powertrain companies in the world. Even so, it pays to diversify. “Motorsport is good when it’s good, but it can be quite seasonal,” explains Ian Whiteside, chief engineer for Ilmor’s Advanced Projects group. “We need cutting edge facilities with plenty of capacity to support our racing programmes, but we also need to ensure there’s enough work to keep them busy during the off-season. We try and fill that with racing parts for external companies, along with prototype work for other industries like automotive OEM, aerospace, defence and marine.”

These facilities include a state of the art machine shop, a comprehensive metrology suite and just about every conceivable powertrain testing rig. The building is home to no less than seven different engine dynamometers, including a 20,000 rpm F1 dyno designed and built in-house. Elsewhere, there are more than half a dozen smaller rigs that cater to sub-assemblies and specific components, ranging from valvetrain parts to fuel injectors. Embracing the digital domain Ilmor is a company that places huge emphasis on empirical testing, but the past few years have also seen a dramatic increase in the amount of simulation work carried out. In particular, the company has invested significantly in its CFD capabilities, with the addition of a new 32-core computing cluster and a dedicated in-cylinder combustion modelling package. “Historically we’ve relied heavily on physical testing, usually starting off with a handful of port designs on the flow bench,” Whiteside explains. “In recent years we’ve used rapid prototyping to speed things up, but essentially we’d still pick a handful of RP parts that looked promising on the rig and then manufacture them in metal to test them on the dyno. There’s only a finite number of parts that you can try with this approach and it does get quite expensive.” Aside from the cost constraints, there’s also a fundamental limitation on how much information you can glean from physical testing, he says: “Charge motion is vital to understanding combustion. We do have the facility to measure turbulence on our flow rig and we have an injector rig where we can look at spray patterns with a high speed camera, but these static tests are never truly representative of the real engine. Likewise, while you can measure the results on the dyno that doesn’t necessarily help you to understand the physics that’s produced the effect. In simulation, however, you can look at the root cause.”

While external CFD is more or less universal in motorsport, in-cylinder combustion modelling is still a relatively new field. Modelling the ports on their own creates similar limitations to flow bench testing; even if the correlation is perfect between the CFD and the rig, it doesn’t necessarily reflect what’s happening in the real engine. On the other hand, it’s notoriously tricky to accurately model a full cylinder with moving geometry. Ilmor has turned to the Converge CFD code developed by US company Convergent Science. This uses a radically different approach to generating the mesh that defines the geometry of the simulation. Instead of relying on a manual mesh, Converge automates the process, based on user-defined parameters. Its makers claim this improves the repeatability of the mesh – removing the degree of manual artistry previously involved – and hence providing more consistent results. More importantly, though, Converge regenerates the mesh at each time step throughout the simulation. That may sound time-consuming, but by reducing the cell density in less critical areas of the mesh and increasing it in others – for instance, following the flame front as it propagates out across the cylinder – Convergent Science claims it has achieved a step change in the speed-to-accuracy trade-off. It also means the simulation can represent moving geometry, such as valve and piston motion, without incurring the deformation errors that arise from distorting a static mesh.

This proved to be the tipping point for Ilmor, Whiteside explains: “We still use the flow rig for correlation, but we now use CFD for most of the development work. The first major project we tackled with Converge was the revised cylinder head for the 2016 IndyCar engine. We think we saved around six to eight weeks in terms of development time and arguably got to a better solution. There’s undoubtedly a cost saving too. By screening the designs in the virtual world we probably only manufactured half the number of physical prototypes that we would have done previously. There is a degree of investment in the software, admittedly, but our licence costs less than a single rebuild on an IndyCar engine and it cuts down on running costs too.” In-cylinder CFD is just one part of the shift towards digital development. Ilmor also extensively uses 1D simulation codes like GT Power to model engines at a systems level, often coupling them to the more detailed 3D models.

Other programmes within the GT Suite are also used extensively to analyse things like the torsional vibration and tribology. Elsewhere, the manufacturing process has also been heavily digitised. Once a design is released it is assigned a part number and sent to a central server in read-only form. The manufacturing department picks up this file and uses it to set up the machining operations. Meanwhile, the planning and procurement area of the business establishes a record of the same part number, which goes through a scheduling program to determine when it can be manufactured. “It was quite a lengthy exercise to get this system up and running, but it’s a huge benefit now,” explains Whiteside. “We have much greater control of the work going through the machine shop. Each of the machines is linked to the scheduling system and each part has a job code assigned to it with a bar code, so they get scanned on and scanned off the machine to keep everything updated.” This software also informs the inspection department that there will be parts on the way, where the CMM machines can be programmed in advance using the central CAD model. Finally, the stores are notified of the incoming parts, so each individual component can be given a serial number, which is then used to track it throughout the rebuild life of the engine.

At Ilmor, every stage of the engineering process now features some degree of digitisation, from initial R&D concepts through to managing the service schedule of completed engines. Much of this is part of a gradual trend, admittedly, but there have been step changes, such as the adoption of in-cylinder CFD. Combined, they help to keep this sleepy little corner of Northamptonshire an unexpected focal point for global powertrain development.

By Chris Pickering 17th January 2017 borrowed from The Engineer

Page 1 of 20  1 2 3 >  Last ›