Under control

Electrical engineering prof develops algorithm used in all-electric vehicle

When Tesla Motors unveiled the prototype for its first electric vehicle in July 2006, the automotive world took notice. The Tesla Roadster—a sleek and stylish sports car—promised high performance and zero emissions in a single package. And when the company's first production vehicles hit the road earlier this year, they delivered on that promise. The all-electric sports car can accelerate from 0 to 60 mph in 3.9 seconds and travel 244 miles on a single charge. Featuring an EPA rating of 135 mpg equivalent, the Tesla Roadster has received high marks from the automotive press and sports car enthusiasts alike.

What many people don't know, however, is that the 2008 Tesla Roadster has a Penn State connection.

Heath Hofmann, associate professor of electrical engineering, became involved with Tesla Motors in December 2006. As part of a team that included Tesla Motors employees Drew Baglino, Troy Nergaard, and Greg Solberg, Hofmann helped design a digital replacement for the prototype vehicle's unreliable analog controller. The new controller features a unique design that offers robust, efficient performance over a wide range of speed and torque. Hofmann was responsible for developing the algorithm used in this new device.

"There are lots of controllers in all vehicles in general," he says. "What we're talking about here is the control algorithm for the traction, or propulsion, drive for the vehicle."

In an electric vehicle, propulsion comes from an electric motor. Because the motor must operate over a wide range of speed and torque, special power electronic circuitry is used to apply AC voltages whose magnitude and frequency can be controlled.

"The algorithm controls this power electronic circuitry," Hofmann explains. "It reads in information from the machine, such as the electrical currents flowing into the windings and the rotational velocity of the rotor. From this information, it determines appropriate commands to the circuitry. These commands cause the circuit to apply voltages to the motor that achieve the desired torque that the motor applies to the shaft, and hence the desired amount of traction force that the wheels apply to the road. The controller also allows regenerative braking, where the kinetic energy of the vehicle is converted to electrical form and stored on the battery for future use."

Switching to the new controller proved advantageous on several fronts.

First, going digital made it easier to change the control. "With digital, the control algorithm is a computer program implemented on a microprocessor," Hofmann states. "You change the control by writing new code, whereas with the analog controller, you have to redesign the hardware—different chips and electrical components have to be changed to change the control."

In addition, Hofmann notes, the new controller is much more robust than the original. "It deals with some of the challenges that Tesla Motors had, which were that they were looking at a very high-performance vehicle with very high acceleration capability," he says. "What that means is that the motor has to provide a large amount of torque per unit weight. Because it's a sports car, it's not very big. They have a limited amount of space, so they have designed a very high power density motor. But to actually utilize all of that power, you need to have a control algorithm that can handle some of the issues that come up when you're trying to generate a lot of torque."

One of these issues, says Hofmann, is magnetic saturation. "When you're trying to get a high torque density out of one of these machines, you enter what's called a saturation region. The machine has magnetic materials in it, basically iron, that magnetically saturate as the magnetic fields in the machine get larger. The magnetic fields are what generate the torque, and as a result, the model of the machine under those conditions becomes very complicated, very nonlinear. Your controller needs to be very robust so that you can operate well in that saturated region."

Another issue is that the controller needs to work well over a wide speed range, from negative to very high. "The Roadster's motor spins up to 13,000 revolutions per minute," Hofmann explains. "Through accurate control of the operating point over the entire range of torque and speed, the motor can be run at its optimum efficiency."

Despite these challenges, Hofmann says development of the new controller went extremely well and it ended up going into production much earlier than expected.

And as an added bonus, he was lucky enough to get some personal drive time in the Roadster. "The acceleration is amazingly impressive; you feel like someone is pushing you into your seat," Hofmann remarks. "I recall getting motion sickness afterwards. It was as if I had just been on a roller-coaster ride."

—Jane Harris

Dr. Hofmann can be reached at hofmann@engr.psu.edu. More information about Tesla Motors and the Tesla Roadster is available online at www.teslamotors.com.