The key is hooking up the two components of the turbo to their own individual electric motors, with an energy storage device in between. It's different than the electric supercharger systems you have seen on certain Audi products, for example. Those systems recover energy like a hybrid, store it, and then use it to drive an intake compressor. It supplements conventional turbochargers that harvest energy from the exhaust. In systems like Audi's, the electric supercharger is supplementing the sequential conventional turbochargers when they're not operating efficiently, at very low RPM in particular. It works well, but it's complicated, and it is a workaround for the limitations of a conventional turbocharger. See below for an animation of the Audi system.
Usually, optimizing a turbo is a compromise between figuring out what RPM is ideal for each side to spin at to generate power. A smaller compressor generates boost more quickly, but loses efficiency at higher RPM. But there's way more energy in high-RPM exhaust gasses. By hooking up the turbine to an electric motor instead, you can harvest energy from the exhaust throughout the rev range, and particularly when the engine is pushing lots of gasses through. And you can store that energy in a battery if it's not needed at that moment.
The intake-side compressor also has a reversible electric motor attached. It is not physically connected to the turbine, so it can operate at any time the computers decide it's beneficial. As engine RPM increases, the compressor doesn't have to increase its speed beyond its optimal range, so there's less energy wasted. And at low RPM situations, when a conventional turbocharger wouldn't have enough exhaust gas passing through its turbine side to generate useful boost in the compressor side, the electric motor can spin up Ferrari's divorced compressor to provide some boost. That practically eliminates turbo lag.
Lastly, in high-RPM operation, the turbine can provide more than enough power to run the compressor at full speed. So it can also send some additional energy to a third motor/generator mechanically attached to the driveline. That means the turbine is providing boost through the compressor and additional mechanical energy by way of the crank- or transmission-mounted motor/generator, increasing efficiency. And of course, making total engine output even greater.
There are other benefits. Packaging, for one. Since the compressor and turbine aren't physically connected to each other, they can be placed wherever is best in a tight engine bay. With a conventional turbo, even the best laid-out systems still end up having a lot of piping, routing both exhaust gasses and compressed intake charge to and from the same location. It puts a very hot turbine in close proximity to the intake tract, reducing efficiency by heating up the intake charge. But if the compressor is separate, you can put it pretty much anywhere, and have a shorter, more direct path to the intake manifold. It could even eliminate the need for a charge-air cooler (an intercooler) in certain applications, saving weight and complexity.
Sound quality is another. In a typical turbocharged car, using the exhaust gas pressure to spin a turbine necessarily quiets the car down. Ferrari's system can actually, in certain situations, use electricity to spin the turbine to generate some additional exhaust pressure to increase the sound quality — at the expense of some proportion of theoretical mechanical power. It can be spun faster for higher-pitched sounds, or slowed down with the electric motor to produce lower ones. It's all governed by a computer that seeks to balance power output with an overall sound quality profile. This system can eliminate or supplement exhaust systems with active valves.
It's certainly a fascinating design that could be beneficial to both supercars and commuter cars. It also seems pretty feasible, since we've already seen car companies use electric compressors on their own for performance applications. We're excited to see what future Ferraris might get this technology, and what other companies might try something similar.
By Joel Stocksdale and Alex Kierstein