Thermodynamic Modeling and Data Assimilation of Cryogenic Liquid Storage Systems

Cryogenic liquids are increasingly being used on a big scale: in many applications, a predetermined quantity of such fluids must be stored for a long time within an insulated tank. A typical cryogenic liquid storage system consists of a horizontal or vertical tank containing a multi-phase liquid-vapor mixture: these tanks undergo pressure and mass changes because of heat transfer in the liquid, vapor and tank walls, and of mass transfer due to evaporation/condensation processes. In any of these storage systems, the flow of gas into the ullage space (pressurization) or external heat-fluxes leak during long term storage (self-pressurization) lead to a rise in ullage pressure and the creation of a thermal gradient at the gas-liquid interface. Thermodynamic equilibrium may be disturbed by phenomena like external excitation, which can trigger sloshing: this event enhances heat and mass transfer rates and may result in undesirable pressure fluctuations. Beside the fundamental role cryogenic play in the aerospace field, in the framework of fuel for transportation sectors and in order to reach carbon neutrality, storage in liquid form is preferred in locomotives, drones, aircraft and ships. Concerning the naval sector, this work is also highly motivated by the ongoing Clean Hydrogen Propulsion for Ships (CHyPS) project involving Von Karman Institute in the development of a modelling suite to enable simulation of the power trains of the vessels of the future. In this context, this thesis aims to develop a numerical model that can accurately describe the thermodynamic evolution of a cryogenic liquid storage system. The implemented model is computationally inexpensive yet sufficiently accurate and versatile to predict the state of the system under the different operating conditions (i.e., filling, pressurization, fuel extraction, storage). The methodology for this project is composed of four main stages. First, the relevant heat and mass transfer processes are computed using correlation equations taken from the literature and implemented in a Python module with user defined functions and external libraries. Secondly, to assess the importance of different mechanisms in the thermodynamic evolution of the system, scaling and similarity analysis are performed. Then, an inverse method is used as a calibration tool for non linear system identification and retrieval of the closure coefficients (heat and mass transfer coefficients) needed to accurately reproduce the provided training data. Finally, the ability of the model to forecast the physics in the tank for different operating conditions is validated using literature cases. The results demonstrate how the model created can yield fresh insights of the cryogenic tank behavior, increasing the level of confidence in using cryogenic fuel as well as providing all the stakeholders with important scientific data in real time.