The Near Earth Asteroid Rendezvous (NEAR) mission is the first launch in the Discovery Program, a NASA initiative for small planetary missions. Following the launch aboard a 7925 Delta II, and nearly three years of semi-dormant cruise mode, the NEAR spacecraft will rendezvous with and orbit the near-Earth asteroid 433 Eros. While orbiting 433 Eros, NEAR will conduct the first systematic scientific exploration of an asteroid by measuring its bulk, surface, and internal properties. In order to accomplish this task, the NEAR spacecraft contains a suite of sensors comprised of an Infrared (IR) Spectrometer, a Multispectral Imager/Spectrometer, an X-Ray/Gamma-Ray Spectrometer, a Magnetometer, and a Laser Rangefinder. The structure supporting these sensors is part of a 3-axis stabilized, solar-oriented spacecraft that employs fixed coplanar solar arrays to provide a constant voltage, regulated power bus. The NEAR propulsion system incorporates the unique thermal design of a single 450 Newton (N) bi-propellant thruster embedded in the spacecraft, accompanied by four external 22.5 N and seven external 4.5 N mono-propellant thrusters.
Due to its externally mounted sensors, tight power budget, and widely varying mission conditions, the NEAR spacecraft presented a challenging thermal design opportunity. Further compounding this challenge were the cost and schedule constraints imposed by the Discovery program guidelines. This paper describes the design approach, problems encountered in the design process, and final thermal design. The thermal modeling of the NEAR spacecraft and instruments is discussed, as are the analysis techniques and software used to evaluate the thermal design. Results of spacecraft-level thermal vacuum testing are reported in context with comparisons to expected thermal performance, and discrepancies are highlighted. The post-test thermal model correlation effort required to resolve these discrepancies and produce revised on-orbit predictions is detailed. Early mission temperature telemetry is also included and correspondence with predicted performance is discussed.