Solid Oxide Fuel Cells (SOFCs) are considered suitable for alternative energy solutions due to advantages such as high efficiency, fuel flexibility, tolerance to impurities, and potential for combined cycle operations. One of the main operating constraints of SOFCs is fuel starvation, which can occur under fluctuating power demands. It leads to voltage loss and detrimental effects on cell integrity and longevity. In addition, reformer based SOFCs require sufficient steam for fuel reforming to avoid carbon deposition and catalyst degradation. Steam to carbon ratio (STCR) is an index indicating availability of the steam in the reformer. This work takes a holistic approach to address the aforementioned concerns in SOFCs, in an attempt to enhance applicability and adaptability of such systems. To this end, we revisit prior investigation on the invariant properties of SOFC systems, that led to prediction of fuel utilization U and STCR in the absence of intrusive and expensive sensing. This work provides further insight into the reasons behind certain SOFC variables being invariant with respect to operating conditions. The work extends the idea of invariant properties to different fuel and reformer types. In SOFCs, transient control is essential for U, especially if the fuel cell is to be operated in a dynamic load-following mode at high fuel utilization. In this research, we formulate a generalized abstraction of this transient control problem. We show that a multi-variable systems approach can be adopted to address this issue in both time and frequency domains, which leads to input shaping. Simulations show the effectiveness of the approach through good disturbance rejection. The work further integrates the aforementioned transient control research with system level control design for SOFC systems hybridized with storage elements. As opposed to earlier works where centralized robust controllers were of interest, here, separate controllers for the fuel cell and storage have been the primary emphasis. Thus, the proposed approach acts as a bridge between existing centralized controls for single fuel cells to decentralized control for power networks consisting of multiple elements. As a first attempt, decentralized control is demonstrated in a SOFC ultra-capacitor hybrid system. The challenge of this approach lies in the absence of direct and explicit communication between individual controllers. The controllers are designed based on a simple, yet effective principle of conservation of energy. Simulations as well as experimental results are presented to demonstrate the validity of these designs.

Experience Mapping based Prediction Control (EMPC) is a new type of controller presented in literature, which is based on the concept of Human Motor Control (HMC). During the developmental phase, called the initial learning phase, the controller records the experience in a knowledge base, through online interactions with the system to be controlled. This knowledge base created using the experience maps is termed as Experience Mapped Knowledge Base (EMK). The controller envisages the development of EMK only through interaction with the system, without the need for knowledge of the detailed plant model. The EMPC controls the system through prediction of actions based on the mapped experiences of EMK. Depending on the nature of control required for the system chosen, various strategies can be used to achieve control using the EMK. The above controller has previously been utilized for motion control applications. In the present work an effort has been made to study the suitability of the EMPC for the voltage regulation of switching converters. The plant chosen for the control study is a discontinuous conduction mode (DCM) buck converter. The parameter to be monitored for the purpose of control is the load voltage. The control input from the EMPC to the converter is a duty ratio value based pulse-width modulated (PWM) signal. Two strategies of control have been proposed: steady state control and transient control. Steady state control action maintains the steady state output voltage at the required value for a given load. The transient control action is used to improve the transient performance of the system. Iterative predictive action and iterative transient actions are used to facilitate convergence of the output voltage to within the required range in presence of non-linearities and uncertainties in the system. Impulse action is introduced to further improve the transient performance of the system. The EMPC is compared a proportional-integral (PI) controller for the given DCM buck system.