Fast data

Current E Mahindra Racing Blueprints Jan 2016 Formula E sensor network

Our modern world is built on data and computing power – and Formula E is no different. A comprehensive array of sensors and electronic devices nestle under the sleek skin of the all-electric Formula E speed machine (see graphic above; key below); together, this “neural network” gathers some 1.2GB of performance-critical data every race day. Here’s how Mahindra Racing goes about plugging into this information.

“An electric powertrain is only good as its inputs,” says Vinit Patel, chief engineer at Mahindra Racing. “Good controls software and great controls strategies are vital to achieving the best performance from the hardware. Software development is becoming the ‘holy grail’ of Formula E.”

To visualise how the network functions, think of the nervous system in your body. Using a vast array of sensors, the brain can capture all sorts of information about touch, taste, smell, sight and sound. It can then process this data and react. Sometimes these reactions are made consciously (such as activing the muscles in your fingers or moving your legs to walk, for example); sometimes they are involuntary, such as the signals that keep your heart beating, activate sweat glands to cool the body or constrict blood vessels to retain heat.

In a similar way, the Formula E car captures a wealth of data through its sensor network, which is delivered to the McLaren-sourced ECU, the “brain” of the car. Some of this data can be accessed in real time by race teams to help optimise performance and head off potential problems, but live telemetry is far more limited in Formula E than other motorsports. “During all sessions, teams are restricted to a few channels to check the ‘health’ of the car, such as water pressures, system voltages and battery cell temperatures,” explains Patel. “Crucially, energy consumption and regeneration numbers are not transmitted by telemetry.”

In fact, the only way that teams can get information on energy consumption during a track session is when the driver radios in the values displayed on the steering wheel display. (The FIA has access to live energy data to ensure drivers are not exceeding energy or power limits.)

“This sometimes results in engineers tearing their hair out,” Patel says. “In Beijing, as Nick Heidfeld crossed the line for third place, one of our engineers suspected that Nick may have been over the energy limit. But Nick simply replied over the radio, ‘No worries guys, it was easy – the dashboard said I had 0.07kWh left!’ Restricting team telemetry puts the show much more within the driver’s control but it means that we have a lot to do when the car gets back to garage.”

This is in marked contrast with other top level motorsports, where teams have at their disposal a huge amount of live data from the car, which can be analysed by small armies of engineers while the cars are rolling. Some live data is even beamed directly to factories in other countries, where even more technical experts can provide additional support or troubleshooting to tweak car settings during track sessions.

The first job for the Mahindra Racing engineering team when the driver returns to the pits is to plug into the ECU and download the data. (“Wireless systems could be faster but could cause security issues,” says Patel.) There can be a lot to get through: some 600MB of facts and figures from the free practice sessions, 150MB from qualifying and 450MB from the race itself.

“At the track, we have a team of engineers looking at different elements to ensure the best reliability and performance on that day for that set of climatic and track conditions,” Patel explains. “Because practice, qualifying and the race all take place in a single day, the key to matching or bettering performance improvements elsewhere in the pitlane is to be fast but accurate with analysis and solutions. Away from track, we have more time to examine the data, and we can feed our findings into our suite of simulation tools to ensure we’re making continual progress throughout the season.”

While the principal goal of improving speed and reliability is shared with other motorsports, the engineering challenges in Formula E are unique in other respects. “We’re working with a very different set of parameters to cars with internal combustion engines, although the end goal of optimising performance under given regulations is similar,” Patel explains. “With both types of powertrain, you want to optimise power and torque. In other series, these are not usually capped but, rather, are controlled by the powertrain architecture (V6, inline four, turbo size and so on). The electric motor system can generally deliver peak torque all of the time until the motor becomes RPM-limited; this is something ICE engines cannot do and so torque delivery must be optimised or compromised against the peak power band. Programming and mapping an electric motor powertrain is therefore simpler in some ways, as we can map it to deliver the regulation power all of the time.”

Control software is much more open to teams in season two to modify, which allows new strategies to be adopted and gives trackside engineering teams an arsenal of tools with which to hone performance throughout race day. “We can make high level changes, such as smoothing and shaping torque requests to ensure we are using the all of the power we’re allowed to, all of the time,” says Patel. “This involves using a table of values, with RPM and throttle positions as column and row headers. It’s manual and labour intensive right now – hard work rather than magic! We are developing a low-level artificial intelligence system that will allow the car some autonomy over certain parameters. We’re at very early stages so I can’t reveal any more than that.”

The team can also make adjustments to other parts of the drivetrain which may yield more marginal results, such as analysing gear shifts to make them as smooth and as quick as possible. The ultimate goal is to balance efficiency and outright pace, while keeping a close eye on thermal performance.

“We upload new maps or software at least five minutes before the beginning of the next session,” Patel says. “That gives us time to power up the car to make sure everything is working and that we haven’t created any bugs!”

Once the changes have been made and uploaded, the drivers head out onto track to see whether the updates have worked and to collect more data to be pored over by engineers in their never-ending question for better performance.

Current E Mahindra Racing Blueprints Jan 2016 Formula E sensor network plus key

4 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *