Assoc. Prof. Dr. Thach Dinh 

Battery and fuel cell systems can be described using mathematical equations that capture their behavior over time. The states of these systems represent how they evolve and change. To gather information about the system states, physical sensors are used. However, the available information from these sensors is limited.

In my project, I want to reconstruct two important aspects of battery and fuel cell systems: the state of charge for batteries and the state of health. This is achieved by utilizing the partial information obtained from physical sensors. I will extensively study a specific class of algorithms called dynamic observers, also known as software sensors. These dynamic observers estimate the system's states by integrating a dynamic model. For a dynamic observer to work effectively, it requires the measurements from the physical sensors to contain enough information to determine all the system states. This property is known as observability. In this project, I focus on developing a particular type of observer called set-membership observer. Set-membership observers provide a set of possible states, including an interval in which the true states lie. The use of set-membership observers is especially advantageous when the system is subject to various disturbances, noise, and unknown inputs such as system faults or cyber attacks. This approach facilitates robust control and monitoring of battery and fuel cell systems as well as effective detection of faults.