A simulation study of the use of accelerometer data in the GRACE mission

The Gravity Recovery and Climate Experiment (GRACE) mission’s success in reaching its science goal of obtaining highly accurate gravity field estimates is dependent on how well it is possible to use the accelerometer data to model the effects of the non-gravitational forces acting on the satellites. The implementation of the accelerometer data into the conventional gravity estimation process was the focus of the research in this study. It provides an improved understanding of the issues associated with the use of an onboard accelerometer along with some usage guidelines. The improved knowledge was gained through simulation studies that were used to evaluate the simultaneous estimation of the accelerometer scale factors and biases along with the gravity field. The simulations were used to investigate various conditions that affect the accelerometer data, including the inherent accelerometer errors, the individual non-gravitational forces acting on the satellites, the orbit altitude, and the orbit orientation with respect to the sun. The non-conservative force effects on the accelerometer parameter recoveries were categorized individually as well as in combination. The orbit altitude ranged from an initial value of 500km down to a final useable altitude of 300km. The effect of the changing orientation of the satellite orbit plane with respect to the sun was evaluated over its full range. An error analysis of the attitude thruster model errors was performed. The alignment errors in the attitude thrusters cause linear accelerations on the satellite that appear in the accelerometer data and whose mismodeling consequently affect the gravity recovery. The correct modeling of the thruster effect on the accelerometer data was investigated. This required an evaluation of a range of integration time steps. The degradation in the gravity recovery due to the mismodeling of the thruster effects was significant and led to an analysis of the use of data filters in GRACE data analysis. To minimize the errors in gravity recovery when under the influence of the mismodeled attitude thrusting, the integrator was modified to ingest an integrated form of the accelerations directly as opposed to using the filtered accelerometer measurements to replace the force models.