PhD Thesis

Nonlinear observer design for state and parameter estimation in PEM fuel cell systems

Work default illustration


  • Started: 01/01/2018
  • Thesis project read: 02/10/2019


Polymer electrolyte membrane (PEM) fuel cells are characterized by its high energy density, low volume, weight and reduced operating temperature, which makes it an attractive option for many portable applications. However, achieving an economically and technologically viable PEM fuel cell involves reducing its operational costs, increasing its lifetime and improving the system's reliability. From this point of view, a reliable PEM fuel cell requires a parameter estimation of a model that depicts the system dynamics, the design of a controller that drives the system and a monitoring subsystem that ensures the proper operation of the system. In such framework, there is internal information that is necessary to characterize the system's behaviour, said information, in general, cannot be measured by existing sensors. In consequence, one has to achieve an estimation of this internal information, which is usually referred as the observation problem. From the automatic control perspective, fuel cell systems are nonlinear multiple-input multiple-output systems with coupled internal variables, uncertainty in the parameters and sensor noise. Under this perspective, the fuel cell observation problem requires the development of new advanced nonlinear observation techniques, which is a rich engineering/mathematical field and the main focus of this doctoral thesis.

The work is under the scope of the following projects:

  • MdM: Unit of Excellence María de Maeztu (web)