Simulating the Refuelling Process for a Liquid Hydrogen-Powered Commercial Aircraft

Abstract

This thesis focuses on simulating the refuelling process for a liquid hydrogen-powered commercial aircraft. The liquid hydrogen is intended to flow from the refuelling truck, through a transfer line, into the aircraft’s tank using pressure feed. The numerical simulation was built upon an existing tank model, however it did not consider the effects of flashing within the transfer line, nor did it account for the thermal mass or imperfect insulation of the transfer line. This tank model was nevertheless not yet been validated against experimental data. In addition, only a small fraction could be filled on top of the receiving tank, as the complex underlying physics of liquid hydrogen droplet evaporation were not included in the model. First, the tank model was modified and validated using recently published experimental data. A sensitivity analysis was performed to identify the process parameters for which the model results aligned most closely with the experimental data. The parameters that exhibited the most significant impact on the results included the loss factor, droplet diameter, heat ingress due to radiation, transmission line delay constant, and vent-closing pressure. A loss factor of 11, droplet diameter of 6mm, transmission line delay constant of 5, and vent-closing pressure of 3bar were found to provide the best match with the experimental data. The considered heat ingress due to radiation also yielded satisfactory results, whereas other parameters did not significantly affect the simulation. The experiment was lacking the last 18 seconds of transient data. When comparing the model with the experimental data at the end of the refuelling process, discrepancies were observed. The model exhibited namely a significantly higher pressure, vapour temperature and mass compared to the experiment. These discrepancies were expected to be caused by the receiving tank’s venting valve in the experiment, which displayed odd behaviour during the last stage of fill. Despite these disparities at the end of the filling process, the transient behaviour throughout the process was found to be well aligned with the model results. As a result, the modified tank model was deemed suitable for simulating the LH2 refuelling process. Then, a case study was conducted to simulate the refuelling process for a commercial aircraft similarto the Airbus ZEROe Turboprop concept. The objective was to calculate the refuelling time and the associated losses resulting from venting the aircraft’s tank during the refuelling process. The refuelling time has been determined to be around 19 minutes, accompanied by venting losses from the aircraft tank totalling 36.7kg, which accounts for approximately 2.2% of the total fuel transferred. This refuelling time is more than three times longer than the refuelling time of the kerosene-powered equivalent, the Bombardier Q400. Nevertheless, the total time taken by the LH2 refuelling process was determined tobe well within the turn-around time if LH2 refuelling is possible while in parallel performing other tasks of the turnaround. However, future research is required to determine whether this is possible. Additionally, the aim was to find what process parameters impact the refuelling time and losses due to venting from the aircraft tank. Therefore, a second sensitivity analysis was conducted to assess the impact of various parameters on the refuelling process of the case study aircraft. This analysis revealed that the venting pressure of the aircraft tank, operational pressure in the trailer tank, transfer line length, and transfer line diameter emerged as crucial factors influencing the refuelling time. A general observation was done were an increase of 25% in transfer line length resulted in a increased refuelling time of 6%. In addition, it was noticed that a decrease in venting pressure by 0.1bar resulted in a decrease in refuelling time of about 1.2%. The percentage of spraying in the aircraft tank was identified as the most significant factor affecting venting from the aircraft tank, for which 100% spraying reduced the amount of venting by 30%. Next, a one-dimensional, homogeneous fluid model was developed to consider the effects of flashing, thermal mass, and imperfect insulation of the transfer line. The transfer line was discretised inn elements with equal length. A linear pressure distribution was considered between the supply and receiving tank, but an option was presented to account for the entrance length. Using the conservation laws for mass and energy, the enthalpy in each node along the axial location of the transfer line was computed. Using REFPROP, the corresponding fluid temperature was determined. The transfer line model was coded in Python using Scipy’s solve_ivp solver. The model was validated using both liquid nitrogen and liquid hydrogen. Discrepancies existed between the model and experimental data, which were expected to be caused by incorrect heat transfer correlations. Finally, the developed transfer line model was integrated into the tank model with the the goal of simulating the effects of flashing, thermal mass and imperfect insulation of an initially warm transfer line to more accurately model the refuelling process for the case study aircraft. However, it was discovered that the liquid temperature in the aircraft tank increased to infeasibly high values, which resulted in a lower liquid density, leading to a decreased mass flow rate in the transfer line, thereby increasing the refuelling time. The dominant factor contributing to these elevated liquid temperatures originated from inaccurately modelling flashing. The incapability of the integrated model to simulate two-phase flow within the liquid phases of the tanks was expected to be the cause behind the lacking capability to capture the flashing phenomenon accurately. As a result, it was concluded that the integration of the tank and transfer line models failed to accurately capture the phenomenon of flashing and so did not provide a more accurate representation of the refuelling process. As a result, the case study results obtained solely from the modified and validated tank model were considered to hold greater credibility.It was nevertheless recommended to perform research on restructuring the integrated model so that the liquid phases are described by temperature and pressure, instead of the saturated states. In addition to this, describing the liquid state by temperature and pressure would allow for simulating subcooling. Since subcooling aids to achieve a vapour-free flow to guarantee a reliable liquid mass flow, implementing the possibility of simulating subcooling is of interest.