This paper shows the Hardware-In-the-Loop (HIL) technique developed for the complete emulation of Solid Oxide Fuel Cell (SOFC) based hybrid systems. This approach is based on the coupling of an emulator test rig with a real-time software for components which are not included in the plant. The experimental facility is composed of a T100 microturbine (100 kW electrical power size) modified for the connection to an SOFC emulator device. This component is composed of both anodic and cathodic vessels including also the anodic recirculation system which is carried out with a single stage ejector, driven by an air flow in the primary duct. However, no real stack material was installed in the plant. For this reason, a real-time dynamic software was developed in the Matlab-Simulink environment including all the SOFC system components (the fuel cell stack with the calculation of the electrochemical aspects considering also the real losses, the reformer, and a cathodic recirculation based on a blower, etc.). This tool was coupled with the real system utilizing a User Datagram Protocol (UDP) data exchange approach (the model receives flow data from the plant at the inlet duct of the cathodic vessel, while it is able to operate on the turbine changing its set-point of electrical load or turbine outlet temperature). So, the software is operated to control plant properties to generate the effect of a real SOFC in the rig. In stand-alone mode the turbine load is changed with the objective of matching the measured Turbine Outlet Temperature (TOT) value with the calculated one by the model. In grid-connected mode the software/hardware matching is obtained through a direct manipulation of the TOT set-point.
This approach was essential to analyze the matching issues between the SOFC and the micro gas turbine devoting several tests on critical operations, such as start-up, shutdown and load changes. Special attention was focused on tests carried out to solve the control system issues for the entire real hybrid plant emulated with this HIL approach. Hence, the innovative control strategies were developed and successfully tested considering both the Proportional Integral Derivative and advanced approaches. Thanks to the experimental tests carried out with this HIL system, a comparison between different control strategies was performed including a statistic analysis on the results The positive performance obtainable with a Model Predictive Control based technique was shown and discussed. So, the HIL system presented in this paper was essential to perform the experimental tests successfully (for real hybrid system development) without the risks of destroying the stack in case of failures. Mainly surge (especially during transient operations, such as load changes) and other critical conditions (e.g. carbon deposition, high pressure difference between the fuel cell sides, high thermal gradients in the stack, excessive thermal stress in the SOFC system components, etc.) have to be carefully avoided in complete plants.
Fuzzy Coordinated PI Controller: Application to the Real-Time Pressure Control Process