Parameter Estimation of Spacecraft Fuel Slosh

Abstract

The nutation (wobble) of a spinning spacecraft in the presence of energy dissipation is a well - known problem in dynamics and is of particular concern for space missions. The nutation of a spacecraft spinning about its minor axis typically grows exponentially, with the rate of growth being characterized by a Nutation Time Constant (NTC). For launch vehicles using spin -stabilized upper stages, fuel slosh in the spacecraft propellant tanks is usually the primary source of energy dissipation. For analytical prediction of the NTC this fuel slosh is often modeled using simple mechanical analogs such as pendulums or rigid rotors coupled to the spacecraft. Identifying model parameter values which adequately represen t the sloshing dynamics is the most important step in obtaining an accurate NTC estimate. Analytic determination of the slosh model parameters has met with mixed success and is made even more difficult by the introduction of propellant management devices and elastomeric diaphragms. By subjecting full -sized fuel tanks with actual flight fuel loads to motion similar to that experienced in flight and measuring the forces experienced by the tanks these parameters can be determined experimentally. Currently, the identification of the model parameters is a laborious trial -and -error process in which the equations of motion for the mechanical analog are hand -derived, evaluated, and their results are compared with the experimental results. The current research is an effort to automate the process of slosh model parameter identification using a MATLAB/SimMechanics -based computer simulation. Two different parameter estimation and optimization approaches are being evaluated and compared in order to arrive at a relia ble and effective parameter identification process. The first approach is conducted using Newton's method for nonlinear least squares. The second estimation method is a black box approach using MATLAB's Parameter Estimation Toolbox. To evaluate each pa rameter identification approach, a simple one -degree -of -freedom pendulum experiment is constructed and motion is induced using an electric motor. By applying an estimation approach to a simple system with known characteristics, its effectiveness and accur acy can be evaluated. The same experimental setup can then be used with fluid -filled tanks to further evaluate the effectiveness of the process. Ultimately, the proven process can be applied to the full -sized spinning experimental setup to quickly and a ccurately determine the slosh model parameters for a particular spacecraft mission. Automating the parameter identification process will save time and thus allow earlier identification of potential vehicle performance problems. This, in turn, can reduce the cost and schedule penalty associated with needed design changes.