This paper focuses on the conceptualization and implementation of a Machine Learning (ML) algorithm for the temperature control of a cryogenic hydrogen cooling system.
Cryogenic hydrogen will be used as energy carrier for mobility applications such as maritime vessels or cargo carriers.
This paper takes a scientific approach to providing a concept to utilize the thermodynamical potential of cryogenic hydrogen when the fluid is pre-heated by the waste heat of the fuel cell or other heat dissipating components in the drivetrain. Thereby, the hydrogen thermodynamic parameters must be controlled in real time to assure an efficient and safe cooling with respect to pressure drops and material temperatures.
Following a detailed survey of the specific thermodynamic challenges using cryogenic hydrogen as a cooling agent, a Machine Learning algorithm has been developed to predict the temperature in the system in dependency on the applied power to the motor. Finally, the verification process of the control loop is elaborated and investigated.
The use of fossil fuel in mobility onshore and offshore must be substituted by alternative energy carriers to comply with climate friendly requirements. One possible approach is the use of hydrogen as an energy carrier to cross the borders between different sectors of energy demand, e.g. transport, civil engineering and electricity. Therefore, utilizing hydrogen to produce electric energy via a hydrogen fuel cell is the most favourable way for a clean energy transformation process. Figure 1 illustrates the energy transformation process in a hydrogen fuel cell powertrain, converting the chemical energy stored in hydrogen into electrical energy, and ultimately into mechanical energy for a propulsion system. In the present approach, cryogenic hydrogen serves as cooling agent for the thermal control system of an electrical drivetrain before entering the fuel cell. Due to restrictions in installation space and energy costs for the thermal control system, high efficiency is required for the thermal management system. In a previous research study, a cryogenic cooling system for a high efficient axial flux motor has been analysed, see (Baeten et al., 2023). This approach is based on heat transport via convection as the dominant cooling mechanism, implemented in the form of hollow conductor cooling. Using cryogenic hydrogen, no additional cooling medium is required, which results in reducing the overall complexity of the system. Until now, only one study has been identified considering hydrogen in a hollow conductor cooling (Vietze et al, 2022), but not for a complex thermal control system.