Thursday, July 29, 2010

Aeroelasticity in F1

Formula One's latest technical controversy surrounds the aeroelastic front wings currently used by Red Bull and Ferrari. The ends of the front wings on these cars appear to be deforming at high aerodynamic loadings, thereby generating ground effect downforce. Courtesy of Darren Heath's photographs, estimates suggest that the front-wing endplates are deflecting by up to 24mm.

McLaren, in particular, are currently working hard to understand how the effect is achieved. A cursory literature search, however, suggests that the effect probably depends upon the orientation of the carbon-fibre plies in the front wing, and Red Bull and Ferrari may even be using a method of coupling the bending of the front wing to the twisting of the endplates.

City University's Aeronautics department point to research revealing "the effect of ply orientation on the dynamic and aeroelastic behaviour of composite wings." Bristol University's Aerospace Engineering department provides further details, explaining that "Laminated composite materials designed adequately can present elastic coupling properties that can be used to induce an adaptive change. For instance, a composite presenting in-plane elastic coupling that is loaded under normal loads experiences a shear deformation. The proposed morphing design consists of a wing made of laminated composite materials presenting elastic couplings so as to induce twist when the wing bends. This concept, if proven, could provide a passive actuation for the control of the wing twist." Elsewhere, there are claims that "most new helicopters have composite elements in the hub/root of the main blades that are used to replace functions usually done by hinges on older designs."

Fascinatingly, a group of researchers in the Netherlands have also just published a paper entitled 'Aeroelastic tailoring using lamination parameters - Drag reduction of a Formula One rear wing', which also proposes a bending-torsion coupling, in this case to reduce rear-wing induced drag.

If these front wing effects are genuinely dependent upon the precise orientation of the carbon-fibre plies, then it will be interesting to see if McLaren can react on a time-scale consistent with winning this year's World Championship...

Sunday, July 25, 2010

Ferrari team orders

So what exactly was it that made Ferrari's decision to move Fernando Alonso past Felipe Massa in the closing stages of Sunday's German Grand Prix, so objectionable? Whilst team orders have nominally been banned since Ferrari's infamous prior offence with Schumacher and Barrichello in 2002, covert team orders have, of course, continued to be implemented. The teams have avoided issuing overt commands over the radio, but team orders have nevertheless been effected, either by issuing information which is interpreted as a coded instruction by the driver, or by timing pitstops in a manner which swaps the positions of the team-mates.

So, was it because victory in a Grand Prix was at stake? Well, it clearly wasn't this alone, because Massa himself had to sacrifice victory in the 2007 Brazilian Grand Prix in order that his then team-mate Kimi Raikkonen could win the World Championship. There was no media or public outcry on that occasion.

Was it because there are still many races to be run in the championship? Well, once again, clearly not, because in 2005 Juan-Pablo Montoya was forced to let McLaren team-mate Kimi Raikkonen past in the Hungarian and Belgian Grands Prix in order to assist Kimi's championship hopes, when the season had yet to reach its final stages. In the case of the Belgian Grand Prix, this cost Montoya a Grand Prix victory. McLaren even deliberately sabotaged Montoya's chances in the Canadian Grand Prix that year to permit Raikkonen first call at the pitstops induced by a safety-car.

Was it because the passing manoeuvre took place on-track, rather than during the pit-stops? Once again, it cannot be this alone, because at the 2008 German Grand Prix, Heikki Kovalainen let McLaren team-mate Lewis Hamilton past at exactly the same place, coming out of Turn 6, where Massa let Alonso past in this year's race. On that occasion, Kovalainen wasn't leading, but by letting Hamilton past he permitted Lewis to catch the leaders and win the race himself.

Whilst these factors exacerbate the offence, there is a crucial additional property which is common to both the 2002 Austrian Grand Prix scandal and Sunday's reprise: Massa, like Barrichello eight years ago, was labouring under the illusion that he would be permitted to beat his team-mate.

Ferrari could, if they so wished, have explained to Massa what they required of him before the race. Team Principal Stefano Domenicali could have told Massa beforehand that, 'if we're running 1-2 in the final stages of the race, with Fernando less than five seconds behind, then you must let him pass, for the sake of the championship.' The fact that Ferrari didn't do this, entails that they were hoping that the situation would never arise, that they could avoid de-motivating Massa by making their support for Alonso explicit. Thus, rather than acceding to a pre-agreed plan, we had a coded instruction, and a driver dis-illusioned whilst driving a car that was leading a Grand Prix.

Grand Prix racing will always be a confluence of sport, business and technology, but to maintain the revenue streams which depend upon the interest of millions of fans across the world, the teams need to understand that any coordination between team-mates must be seen to be done with the planned and grudging consent of both drivers.

Friday, July 16, 2010

Louvres, holes and algebraic topology

Autosport's technical triumvirate of Mark Hughes, Gary Anderson and Giorgio Piola, have spotted a beautiful touch on the engine cover McLaren were intending to introduce with their new exhaust-blown diffuser at last week's British Grand Prix. Whilst the teams have in recent years been forced, by regulation, to substitute single exit orifices in place of radiator exit louvres, McLaren have cleverly realised that if they cut the bodywork between each louvre in half, then the result is topologically identical to a single hole.

To recall, topology is the mathematical study of the connectedness and continuity of shapes and surfaces, irrespective of their geometry. Thus, the surface of a tea-cup is often said to be topologically identical to the surface of a doughnut, and the London Underground Map is said to preserve the topology of the capital city's subterranean transportation network, if not the actual length and shape of the tracks.

Now, the number of holes in a shape or surface M is typically characterised by an object from algebraic topology called the fundamental group π1(M). Algebraic topology is essentially the use of groups to characterise the topological characteristics of shapes and surfaces. Recall that a group is a set of elements which is equipped with a binary product operation, a unary operation called the inverse, and a special element called the identity element.

To understand what the fundamental group is, we need another concept, called homotopy equivalence. Basically, two curves or loops are said to be homotopically equivalent if one can be continuously deformed into another. If a pair of curves cannot be deformed into each other, then they belong to different homotopy equivalence classes.

Now, if we consider the set of all loops beginning and ending at the same point p in a shape or surface, then we can tag one loop onto the end of another to form a new loop. This concatenation operation gives us a product operation between different homotopy equivalence classes of loops at a point. Furthermore, by simply running around a curve in the opposite direction, we have an inverse operation, and the degenerate loop consisting of the point p, serves as the identity element e of a group. The homotopy equivalence classes of loops at a point can thus be treated as a group, and this group is called the fundamental group π1(M). (If the shape or surface is connected, it can be shown that the fundamental group at each point is isomorphic, hence the point chosen is arbitrary).

If a shape or surface has no holes in it, then all loops through an arbitrary point p can be continuously shrunk down to the point itself, hence the fundamental group consists of a single element π1(M) = {e}. However, if the shape or surface has a single hole, then there will be at least two homotopy classes of loops through a point: those which can be deformed down to the point, and those which cannot, because they circle the hole, and cannot be shrunk any smaller than the hole. In this case, the fundamental group consists of at least two elements. If there are two holes, then the loops around both holes cannot be shrunk to a loop around one hole, and the loops around one hole cannot be shrunk to a point, hence the fundamental group will contain at least three elements.

In the case of a radiator exit with n louvres, the fundamental group of the surface will contain at least n+1 elements. By cutting through the bodywork between the louvres, however, it becomes impossible to form a loop around anything but the entire louvre collection. Thus, topologically speaking, there is only a single exit orifice. Ingenious.

Sunday, July 11, 2010

Laser aerodynamics

One of the intentions of the new technical regulations introduced for the 2009 F1 season, was to exclude the existence of bargeboards between the trailing edge of the front tyres, and the leading edge of the sidepods. A crucial function of these devices was to guide the turbulent air from the front wing and front wheels, away from the vital airflow underneath the car, that ultimately feeds the diffuser.

Now, the teams responded to this with typical ingenuity, by shortening the sidepods, and installing mini-bargeboards in the newly created region of space. Nevertheless, the regulations were successful here in eliminating full-length bargeboards. When an item is banned in F1, however, the engineering tradition is to find some other means of achieving the same effect, so let us see if we can do just that. Note, however, that what follows is not intended to be a serious short-term practical recommendation, more an attempt to demonstrate the art of the possible.

Bargeboards were solid, 2-dimensional surfaces. When a solid is introduced into a viscous airflow, the surface of the solid provides a new boundary component, along which the airflow velocity must be zero. In the case of a bargeboard, this creates a stagnation point at the leading edge, and boundary layers down the inner and outer flanks of the boards. The stagnation point at the leading edge forces the streamlines of the airflow to go either side, hence the stagnation point functions as a branching point in the airflow. Thus, a solid, 2-dimensional surface is a hugely convenient device for enforcing the separation of airflow.

However, with bargeboards now banned, the question is whether there are other means of enforcing the separation of airflow in the region of space between the trailing edge of the front tyres and the leading edge of the sidepods. The rules prohibit solid substances in this region, so what else could we use?

Well, in principle we could use a plasma, but confining the plasma would be rather tricky, and would require the use of magnetic fields generated by superconducting magnets, which in turn would need to be cooled by liquid helium. That would be rather challenging to package.

So how about electromagnetic radiation? The regulations only prohibit the presence of solid substances in that sensitive region of space behind the front wheels. The space all around the car is already filled with natural and artificially generated radiation, so banning the presence of radiation would be very difficult. If electromagnetic radiation is permitted in the space between the front wheels and the sidepods, then it follows that coherent radiation, or laser light, is also permitted, and it may be that we can create virtual bargeboards out of laser light.

For aerodynamic purposes, the crucial property of radiation is that it is capable of applying and transferring pressure. Lasers provide the capability to inject pressure into an airflow at very precise locations, and could therefore be used, amongst other things, to create stagnation points in the airflow, and narrow high-pressure layers of air. By this means, lasers could, in principle, be used in F1 to replicate the function of bargeboards in achieving airflow separation.

Needless to say, this would not be the work of a moment. The radiation pressure would create high air temperatures as well as pressures, and the injection of heat energy into the airflow just in front of the sidepods is not necessarily ideal. There is also, of course, the question of laser energy consumption. One might wish to power F1 lasers with the energy stored by kinetic energy recovery systems (KERS), and such an aspiration will be assisted both by the gradual decrease in laser energy requirements over time, and the gradual increase in the energy harvested by KERS.

Nevertheless, given the heat energy injected into the airflow by lasers, it is perhaps at the rear of the car where they could most usefully be employed. Red Bull, of course, have re-introduced exhaust-blown diffusers to the sport this year, and the point about these is that they ultimately use heat energy generated within the engine for aerodynamic purposes. Craig Scarborough explains how Red Bull may be using retarded ignition in qualifying this year to maintain the flow of exhaust gases even when the driver is off the throttle. Perhaps a single KERS-powered laser inserted into the exhaust tract of the engine could achieve the same effect here...

Tuesday, July 06, 2010

Viscoelasticity and F1 tyres

Formula 1 aerodynamics is all about utilising a viscous fluid, to maximize the forces generated by a viscoelastic solid.

As Mark Hughes explained in last week's Autosport, an F1 tyre generates grip by two different mechanisms: physical grip and chemical adhesion. The latter is only triggered when the tyre reaches a certain critical temperature, and to reach that temperature, the physical grip must be used to repeatedly load and unload the tyre, the deformation cycle thereby heating up the carcass of the tyre.

However, if a rubber tyre is represented as an elastic solid, there is a puzzle here, for the theory of elasticity says that an elastic solid is non-dissipative; in other words, no heat is generated as a result of loading and unloading an elastic material. There is no net work done on a perfectly elastic substance during a load cycle. If rubber was genuinely elastic, it would deform, and then return to its initial configuration, at its initial temperature. Tyre friction between the tyre and the road surface would certainly generate heat, even if the tyre were an elastic solid, but if rubber were genuinely elastic, the carcass of the tyre would not heat up as a result of going through repeated load cycles.

To put the puzzle into context, and to understand the answer, we need to consider three types of material defined within continuum mechanics: elastic solids, viscous fluids, and viscoelastic solids.

An elastic solid undergoes elastic deformation up to a yield point, and thereafter undergoes a degree of irreversible plastic deformation, (until it finally fractures). Under plastic deformation, net work will be performed on the solid, and that net work will go into heating it up. In some respects, a solid undergoing plastic deformation behaves like an incompressible viscous fluid.

A viscous fluid, conversely, can be thought of as a solid with no elasticity and no yield point; it is something which flows under the action of an applied force, and which possesses internal resistance to shear forces.

A viscoelastic solid responds to an applied stress by undergoing simultaneous elastic deformation and viscous flow. There is a viscous flow for all stress levels, unlike plastic deformation, which only occurs in elastic solids after a yield point has been attained. The viscous flow produces heat. When the stress is removed, the material will return to its initial configuration via a different curve on a stress-strain graph. This phenomenon is referred to as hysteresis.

To reiterate, perfectly elastic materials do not dissipate energy as heat when a stress is applied and removed; there is zero net work done on an elastic material over a load cycle. The work done deforming the elastic solid will be stored as strain energy, and when the external stress is removed, the strain energy will be used to do work on the environment, resulting in zero net work being done on the elastic solid over a load cycle.

Now, rubber is slightly unusual in that it will heat up when deformed. In most elastic materials, the strain energy will be stored in the electrostatic bonds between molecules. In rubber, whilst some of the strain energy is stored in such bonds, a component of the strain energy is stored in the form of thermal energy, and when the external stress is removed, and the rubber returns to its original configuration, it will adiabatically cool.

If the unloading curve on a stress-strain graph followed the same path as the load curve, then the material would return to its initial temperature at the end of a load cycle. In contrast, a viscoelastic material dissipates a quantity of heat energy equal to the area enclosed by the curves on the stress-strain graph, and this is equal to the net work done on the material. Rubber is a viscoelastic solid. (This is also nicely explained in Pat Symonds's trilogy of articles on the science of F1 tyres, featured in the April/May/June issues of RaceTech Magazine).

Formula 1 aerodynamicists attempt to use the flow of a viscous substance over the car to maximize the downforce on the viscoelastic contact surfaces; this enables the drivers to corner at the speeds necessary to load and unload the viscoelastic contact surfaces to the point at which they dissipate sufficient heat that the chemical adhesion of the contact surfaces is then maximized.

It's all about resisting the flow.

Thursday, July 01, 2010

The F1 Fans' Forum

I thought I'd touch the hem of F1 today by attending the inaugural F1 Fans' Forum at the British Academy of Film and Television Arts in London's Piccadilly.

The idea of this was to provide a Question Time format in which fans could pose questions to a selection of key participants from the sport, comprising McLaren Team Principal Martin Whitmarsh, Lotus Racing Team Principal Tony Fernandes, Mercedes race engineer Jock Clear, Ferrari Press Officer Luca Colajanni, and Force India test and reserve driver Paul di Resta.

The overall concept here is a good one, for F1 fans are genuine financial stakeholders in the sport: the income streams upon which both the teams and Bernie Ecclestone's business are dependent, are ultimately predicated upon the existence of a huge global television audience. Alienate these people, then, at your peril.

Today's event was chaired in a polished fashion by ex-ITV F1 commentator, James Allen, and it was noticeable that every member of the panel was able to organise their thoughts and speak in an impressively coherent and intelligent manner. F1 folk clearly become very well trained in answering any and all questions popped at them.

There were no big surprises in any of the answers, although Martin Whitmarsh took the opportunity to endorse the plan for driver-adjustable rear wings next year, and there was wide general support for the return of F1 test days at Silverstone. One member of the audience made an interesting suggestion that each team could be given a finite fuel allocation for the season, to use as they see fit; teams with more fuel-efficient engines might therefore be able to gain testing miles as a reward for their environmental-friendliness.

PR-impresario Nav Sidhu was much in evidence, looking much like a finalist in The Apprentice, and I also spotted ex-F1 racing editor, Matt Bishop, looking rather well fed in his role as head of McLaren communications. Jonathan Legard and Ted Kravitz were also in attendance, Ted looking rather jolly, Jonathan looking quite animated. And, rather surprisingly, long-time F1 engineer Frank Dernie turned up, Frank suggesting from the back of the hall, in typically contrarian style, that there's no evidence whatsoever that increased mechanical grip makes for better racing.

All in all, it was a perfectly fine event. One suggestion for next time, to spice things up a little, would be to remove the pre-vetting of questions...