Reactor controller design for steam and autothermal reforming for fuel cell applications

Motivated by the need for small scale distributed hydrogen generation, this study investigates two reformation methods which are suitable for implementing in small-scale reformers. Reformers as used for stationary applications have traditionally been in large scale and in continuous operation with minimum downtime. Only recently, studies have been made on the design, optimization, and control for small-scale reformer systems. These three aspects of reformers are dependent on one another, and they should be addressed in a unified fashion. To this end, this study explores implementation of a control algorithm, based on the design constraints of the reformer, and identifies various characteristics and limitations to gain insight for the further development of small-scale reformer systems.;In this dissertation, steam and autothermal reformation are studied. Coupled-physical models are made based on the similarities and differences between these two reformation methods. Since these two processes potentially can use the same catalyst and reactor but different chemical reactions, the models include the characteristics and necessary parameters to describe these reactions. Then, the critical physics are extracted for a control-orientated model for each reformation process, and a control algorithm is derived. The generic direct temperature feedback control has been used in past studies with the reformers used but exhibited significant temperature oscillation due to the inherent high thermal resistance in the catalyst bed. By using heat as the control variable, the new physics-based control algorithm improves temperature control of the reformers, thus maximizing fuel conversion and minimizing catalyst degradation. In steam reformation, the reaction is limited by mass and heat transfer. Maintaining sufficient heat in the catalyst bed improves the overall efficiencies of the process. Similarly, in autothermal reformation, the same limitation can be overcome by the heat generated on the surface of the catalyst by the catalytic combustion. However, the localized heat generated by the catalytic combustion must be controlled to avoid sintering of the catalyst. Based on the physics of the processes in both cases, using heat as the control variable is appropriate. Thus the model-based approach makes these controllers an effective heat regulator, and they have been tested at various conditions and have shown improved performance as compared to the simple temperature control algorithm.;The control-orientated model also gives insight to deriving an observer, which offers the possibility of sensor replacement inside the reactor. The observer has been used to replace the centerline thermocouple in the steam reformers with comparable results to the sensor-based controller. The observer allows online estimation of the reformer temperature by the reference input. Although catalyst degradation can influence the accuracy of the observer, this can be compensated for by additional temperature references inside the reformer or estimation by another observer.;This dissertation helps to extend the current understanding of reformation processes, and can be adopted for future reformer and reformation system design, and controller implementation. Data is presented showing the strengths of this control approach.