The revived Bloodhound project heads for SA's Hakskeen Pan next month for its first high-speed tests. Picture: SUPPLIED
The revived Bloodhound project heads for SA's Hakskeen Pan next month for its first high-speed tests. Picture: SUPPLIED

The Bloodhound supersonic car is back, under new management and preparing to renew its pursuit of the land speed record.

The project went into administration in 2018, unable to secure the financing needed to go racing — even though the vehicle was all but built. But with the purchase of the car by UK entrepreneur Ian Warhurst, Bloodhound has been put on a new footing.

The team expects to start high-speed trials at SA’s Hakskeen Pan in the Northern Cape as early as October, with reigning world land speed record holder Andy Green behind the wheel.

A British team is developing a car that will be capable of breaking the existing world land speed record of 1,228km/h. Powered by a rocket bolted to a jet engine, the vehicle aims to go progressively faster with the ultimate goal of reaching 1,000mph (1,610km/h).

Bloodhound driver and current land speed record holder Andy Green. Picture: SUPPLIED
Bloodhound driver and current land speed record holder Andy Green. Picture: SUPPLIED

Andy Green’s diary

Bloodhound LSR is a hive of activity as we prepare the car for its first high-speed test runs in SA.

The high-speed test runs will give us some key data to confirm the aerodynamics of the car. In the age of supercomputers and artificial intelligence, it may seem old fashioned to go out and do test runs — surely computers can predict everything nowadays? The simple answer is no, they can't.

An American statistics expert named George Box is credited with the saying: "All models are wrong, but some are useful."

That's all very well for statistics, but slightly worryingly, Bloodhound's chief engineer Mark Chapman has also started quoting it.

Mark and his team are building the car, based on computer models, that I'm going to drive at supersonic speeds. Should I be worried at this point?

No, I'm not worried, for the very reason that we are going to be doing high-speed test runs.

We'll explore how accurate our models are during our step-by-step testing, correcting the models as we go.

Cutting edge technologies from Formula 1 to spaceflight do a lot of modelling and a lot of testing. We're no different.

In advance of the high-speed test runs, there is still of lot of background work going on to understand what the computer models can tell us, and how we can check them against the test results once we start running.

This is the bit where the models will prove "useful", even if they are not perfectly accurate.

Another fascinating area for us is understanding how the high-speed solid metal wheels will behave on our desert track.

Hakskeen Pan in SA is a very firm, smooth, dry mud surface, which (we believe) is the ideal surface for supersonic record breaking.

Before we get to supersonic speeds, though, we want to know more about how the wheels will interact with the surface.

At slow speeds the wheels will simply make shallow ruts in the desert surface, perhaps 7-8mm deep, due to mass of the car (about 5.5 tons for our initial tests).

However, at high speeds the wheels will start to "plane" on the surface like high speed boat hulls, reducing the depth of the wheel ruts to around 3mm.

Of course, all these figures are based on models, so they are probably all wrong (see above!), but they are useful in understanding how the car will behave.

One of the side-effects of planing like a high-speed boat is that the wheels will throw up "spray" from the surface.

The first problem with the spray from the wheels is what Ron Ayers, our legendary aerodynamics and performance expert, has termed "spray drag" from the huge clouds of dust that it is throwing up.

Measuring the effect of this spray drag will give us a more accurate indication of how much power Bloodhound needs to get to 1,000mph.

The second problem with the dust spraying up from the wheels is that the rotating storm of supersonic air around the wheels will suck a lot of the dust up into the bodywork.

We saw exactly the same effect with Thrust SSC back in 1997, when the car on that occasion generated massive supersonic dust storms inside the bodywork.

This didn't matter very much for Thrust SSC, as the structure around the wheels was all metallic.

However, Bloodhound's front end is composite, raising the prospect of rapid erosion of the carbon fibre structure and (to quote our chief engineer again) the front end of the car "suddenly going all floppy"…

To counter any possible erosion, we are putting aluminium liners inside the composite wheel arches. This will also allow us to monitor the erosion effects (if any) of the desert dust, without any risk to the bits of the car that stop the nose from becoming unexpectedly floppy.

It's exciting to see Bloodhound sitting in its new workshop at Berkley UTC, gradually being modified from its "runway" test configuration (rubber tyres, partial bodywork, no brake parachutes) to its desert test spec.

This includes fitting the brake chutes, all the bodywork panels, and the desert wheels and fairings (complete with aluminium liners, of course), as we get the car ready to do some high-speed tests in SA. Won't be long now.