Since the first launch of the Kinéis constellation, the operational team has precisely refined its 3D modelization of the satellites, to better predict the effect of the atmospheric drag.
In this paper, we will demonstrate how this precision was obtained : we will discuss the various models used to compute drag, by comparing the operational results with basic models considering each part of the satellite (body, solar arrays, antenna, …) independently without accounting for shadowing versus more global models with realistic drag surface that accounts for shadowing.
This precision now allows us to bring out very small differences of area, due to an additional small (2 centimeters wide) antenna on some specific satellites.
Aside from power consumption considerations and solar arrays not being in the optimal position with respect to the sun, the satellite design allows us to freely tweak the yaw angle without impacting the payload. With the solar arrays being quite large with respect to the satellite body, we can more than double the drag surface by exposing the solar arrays perpendicular to the velocity.
This knowledge can be used to impart a differential drag relative to its neighbors, by commanding the satellites attitudes. This drag can then be used to control the relative position of each satellite on a given orbital plane, reducing the need for propulsion.
Other use-cases also include reducing the drag of satellites by minimizing the solar arrays exposition to the velocity when the power consumption allows it (typically during eclipses).
A specific algorithm dedicated to the in-plane rendezvous of satellites using only differential drag was developed and can be used operationally at any time.