104 JANUARY2020|COMPUTERSHOPPER|ISSUE383
and integratethese parts to simulateacomplete
vehicle model,”Alakshendra adds.
That might sound good in principle,but how do
these specific benefits lead to apossible place on the
podium? It seems that virtually all aspects of the race
car can benefit from the simulation approach.
Alakshendra points out, “Wehaveseenthe teamsusing MATLAB
and Simulink forvarious applications such as control system design,
torque vectoring, lap-time simulation, optimisation, ECU
programming, battery modelling and suspension tuning, which
eventually improved their vehicle design.”
We looked at Formula Student because it’s abit of fun, albeit with
an important educational element, but there’s no shortage of
examples of simulations that have broader benefits. As with race
cars, road cars are also simulated extensively during the design
phase.The same is true of aircraft, and even the spacecraft that
NASA is designing to take man back to the moon.
It’s not just transport, either,with simulation being used
effectively to give us so much of the technology we rely on.
Coming closer to home,for example,it’s pertinent to point out
that the processor in your PC is amiracle of simulation. And in that
sense,it’s almost achickenand egg situation –microprocessors are
now indispensable in designing microprocessors.
STORMYWEATHER
It might seem obvious that if the mathematical model is correct,
then carrying out asimulation by solving that model will give
correct results. Surprisingly,thisisn’t guaranteed.
Meteorologists are fairly confident that the models of the
atmosphere used forweather forecasting, which are based on
well-understood science,are correct. While forecasts are undoubtedly
alot better than theywere in the early days of numerical forecasting,
however,it’s still not unusual foraforecast to miss the mark.
This is no fault of the model, nor of the
supercomputers used to generatethe forecast, but
it’s aconsequence of the weather being what
scientists call achaotic system. This is asystem
that displays, to usethe jargon, super-sensitivity
to initial conditions. In practical terms, this means
that tiny errors in the measured starting conditions of temperature,
pressure,wind speed and so forth will cause the forecast to
eventually differ wildly from reality.And with 100% accuracy in
the instruments used formeasuring these starting conditions
being an impossibility,there’s alimit to howfar intothe future a
forecast can be relied upon.
So do meteorologists just accept this as an inconvenient fact
of life, or are there ways the impact of chaos theory can be
alleviated? Dr David Walters, head of the Research to the
Operations team at the Met Office,suggested that the problem
isn’t as severe as sometimes suggested.
“In the description of the weather as achaotic system, people
often quoteEdLorenz –the father of chaos theory –who,among
other similar statements, said that abutterfly flapping its wings
in Brazil can produce atornado in Texas. The good news is that
while this is true,the influence of most of these small-scale
perturbations on short time-scales is damped down, leaving a
reasonable amount of predictability,atleast over relatively short
forecast periods,”hetells us.
“The evolution of individual weather systems is fairly predictable
over one to five days, and today, even asingle numerical forecast
will often capturetheir evolution accurately.”
However, this isn’t to saythatchaos doesn’t have an influence,
as Walters then explained.
“The problem is that from this single forecast, it is impossible
to determine when chaos will lead to asignificant change in
predictability and fordifferent potential outcomes to move down
ABOVE:Ensemble forecasting
allows the Met Office to judge how
susceptible the weather is to chaos
at aparticular time,and therebyassign
aconfidence level to aforecast
