A fuel cell produces no exhaust other than heat and water. With no moving parts, its design is simple in principle: a membrane is sandwiched between two electrodes. When the hydrogen fuel meets the anode, it is split into a proton and an electron. The proton passes through the membrane to the cathode, where it meets oxygen. The electron takes a longer route between electrodes, traveling through an electrical circuit. The flow of electrons creates the power for the motor. At the cathode, the protons, electrons, and oxygen combine to form water.
Using modeling and real-time simulation enables Nuvera’s engineers to iterate on their design quickly and allows for experimentation without putting a real engine at risk.
The science is simple, but perfecting the recipe for a high-performance power source is hard. Many factors govern the multiple reactions inside a fuel cell, and a software control system must account for them all to squeeze the most power and efficiency out of the device. The control system makes constant corrections based on feedback.
“One of the greatest design challenges is maintaining proper hydration to the cells,” says Pierre-François Quet, Nuvera’s chief engineer. “Not enough water and protons don’t pass through; too much and the cells flood.”
The control algorithms also account for many operating conditions. In simulation, Nuvera tests the system in low and high ambient temperatures, and in low- and high-humidity environments.
In order to experiment with their algorithms in a more realistic setting, Nuvera does hardware-in-the-loop testing: They load their engine model onto a custom computer made by Speedgoat that is tailored to have the same inputs and outputs as the physical engine, and can simulate its operation in real time. The same embedded computer that runs the fuel cell engine is connected to the Speedgoat box and is programmed from C code generated from Simulink.
This setup adds rigor while enabling Nuvera’s engineers to iterate on their design quickly. It also allows for experimentation without putting a real engine at risk.
Virtually all fuel cell vehicles are electrically hybridized, powered by both fuel cells and batteries. In some cases, fuel cells provide a trickle charge to keep the battery charged, while in other configurations both the fuel cell and battery power the motors through an electrical bus. Batteries are also employed to accept a vehicle’s regenerated power, such as when a forklift brakes or lowers a load.
Quet’s team first had to build a model of the lithium-ion battery in Simulink, based on data provided by the manufacturer and collected in-house. They also wrote algorithms that could estimate the battery’s state of charge based on things they could measure—its voltage and current. They then used Simulink to program the control algorithm. The system needs to maintain an ideal level of battery charge, so there’s always enough energy for peak load and enough capacity to reabsorb energy. The Nuvera team also designed the optimal size for various system components by testing the algorithms in a range of simulated forklift and load scenarios.