Matthijs Damen, Forze Hydrogen Racing Team Delft
Building and testing a fuel cell system systematically, which is difficult because every part needs to communicate with other parts, and most parts need to be actively controlled to function properly.
Using LabVIEW and the LabVIEW PID and Fuzzy Logic Toolkit to easily control the full system, and using NI CompactDAQ modules to easily connect and monitor all the parts from a single piece of hardware.
Forze Hydrogen Racing Team Delft (Forze) is a student project associated with the University of Technology in Delft. Our team consists of about 70 students all working on the same car. Each year, the team builds a new hydrogen-powered race car. We combine the spectacle of racing with the development and promotion of sustainable technology.
The team was founded in 2007 with the vision of promoting hydrogen technology. In the first three years, we built hydrogen karts to participate in the Formula Zero competition. The competition organizers provided the fuel cells used in the karts. After three years, the team wanted to expand and changed to a Formula student-type car. This, of course, required designing and building a fuel cell system. Last year, we built our second Formula student-type car, the Forze V.
The main challenge in creating a fuel cell system was the balance of plant (BOP) design, development, and test. The BOP is a complex system of pumps, valves, and sensors that ensures that the fuel cell stack operates under the best conditions. The sensors monitor the system and guarantee safe operation. The BOP controller, which is responsible for control and communication, also must manage the reactant flow to the stack.
To produce power, the stack needs hydrogen and oxygen. The oxygen comes from the ambient air and is pumped to the stack by a supercharger. The air pressure in the stack is between 0 and 0.8 barg. Hydrogen comes from the tank, which has a pressure of 350 bar. The stack cannot operate at this pressure, so a proportional valve regulates the pressure to 0.2 bar above the air pressure. This pressure difference ensures flow from anode to cathode, and not the other way around.
Both the air pump and the proportional valve use a proportional integral derivative (PID) control on the BOP controller to manage the flows. Depending on the amount of power needed, the airflow can be calculated. A mass flow sensor monitors this flow, and is the process variable for the PID controller. The PID controls a radio-controlled helicopter motor connected to the supercharger. Hydrogen pressure is monitored by a pressure sensor in the low pressure box. This pressure sensor is the process variable for the PID on the hydrogen side. This PID controls the proportional valve, which regulates the pressure.
Beside the reactant inlet, there is a recirculation pump on the hydrogen side. This pump recirculates the hydrogen that comes out of the stack back to the inlet. This creates stoichiometry, which is necessary to maximize stack lifetime and stable operation. Stoichiometry is the ratio of actual flow rate over flow required to support the reaction, and is dependent on the current produced, so the pump needs to be controlled by the BOP controller as well.
Previously, the BOP could be tested only when the electronics were finished. Testing was difficult because the electronics didn’t always function correctly, and it was hard to make quick changes to the parameters (for this, the software on the controllers needed to be changed). This year, we started to use LabVIEW with NI CompactDAQ modules so we could test and control individual parts of the BOP as soon as they were available. We also used the NI LabVIEW PID and Fuzzy Logic Toolkit to control the proportional valve and the air pump.
The basics of LabVIEW are easy to understand and quite intuitive. Because even nonprogrammers understood how to make changes to the code, it was easy to make quick changes and expand the program when more parts of the BOP were assembled. Our intention was to create the whole controller in LabVIEW. Then, when the code was working and the electronics were finished, we could shift every task to the BOP controller one by one. First, the BOP controller took over both PID tasks, while LabVIEW was still responsible for all other tasks. Eventually, only the current setpoint was given in LabVIEW and sent to the BOP controller through a controller area network (CAN) message. For this, we used an NI 9862 XNET high-speed 1-port CAN module. The BOP controller passed on all the sensor values, which were shown in the LabVIEW front panel.
We took advantage of the following benefits of using LabVIEW and NI CompactDAQ modules:
Overall, NI products gave us a great advantage.
Matthijs Damen
Forze Hydrogen Racing Team Delft
Delft
Netherlands
m.damen@forze-delft.nl
In LabVIEW 2014 and later, the PID and Fuzzy Logic Toolkit is included natively within LabVIEW Full and Professional Development Systems and does not require a separate license, installation, or activation.