Roll Stabilisation of a Flexible Spacecraft using Quantitative Feedback Theory

A.g. Sreenatha, Mohit GuptaJournal of Aerospace Sciences and Technologies 2023 · DOI

TL;DR

This paper demonstrates how Quantitative Feedback Theory (QFT) can effectively stabilize the roll axis of a flexible spacecraft, even when the vibration frequencies of its solar arrays are uncertain. The authors show that a relatively simple 2nd-order QFT controller can match the performance of more complex $H_\infty$ controllers while offering a more transparent and intuitive design process.

What Problem Does This Paper Address?

Modern spacecraft are growing larger and more power-hungry, leading to the use of massive, flexible solar panels. These structures introduce a significant challenge for attitude control: flexibility.

When a controller applies torque to rotate the satellite, it can inadvertently excite "flexural modes" (vibrations) in the solar arrays. Because these vibrations are lightly damped and their exact natural frequencies are difficult to model precisely on the ground, a standard controller might become unstable or fail to maintain pointing accuracy. The authors tackle the problem of designing a "robust" controller—one that maintains performance despite these modeling uncertainties and external torque disturbances.

Methods

The authors employ Quantitative Feedback Theory (QFT), a frequency-domain technique that explicitly relates the amount of feedback required to the level of plant uncertainty.

Key aspects of their approach include:

  • Two-Degree-of-Freedom (2-DOF) Structure: The design uses a feedback controller ($G$) to handle plant uncertainty and disturbances, and a prefilter ($F$) to shape the overall system response to commands.
  • Loop Shaping on Nichols Charts: Unlike "black-box" optimal control methods, QFT involves a visual design process. The designer shapes the "loop transmission" on a Nichols chart to stay outside of specific "bounds" that represent the worst-case performance requirements across all possible plant variations.
  • System Modeling: The satellite is modeled as a Single-Input Single-Output (SISO) system focusing on roll dynamics, incorporating the first flexural mode of the solar arrays.
  • Benchmarking: The resulting QFT controller is compared against $H_\infty$ (H-infinity) control results from previous literature (McFarlane and Glover, 1990).

Datasets

As a theoretical control paper, this study does not use a traditional "dataset." Instead, it utilizes a numerical simulation model of a flexible satellite.

  • Model Parameters: The state-space matrices are derived from established literature (Ref [4] in the paper), featuring a nominal flexure frequency ($\omega_n$) of 1.5339 rad/sec and a very low damping ratio ($\xi$) of 0.003.
  • Uncertainty Range: The design specifically accounts for a $\pm 10\%$ variation in the natural frequency of the flexible modes.

Key Insights and Ideas

  • Feedback is for Uncertainty: A core philosophy of QFT highlighted here is that feedback is only necessary because of uncertainty. If we knew the plant perfectly, we wouldn't need high-gain feedback.
  • Simplicity vs. Complexity: The QFT process yielded a 2nd-order controller, whereas the $H_\infty$ benchmarks required 4th or 5th-order controllers. Lower-order controllers are generally preferred in aerospace applications because they are easier to implement and verify.
  • Transparency: The authors argue that QFT is "physically more appealing" because the designer can see exactly which performance constraint (stability, disturbance rejection, or control effort) is driving the design at any given frequency.

Key Results

The QFT controller successfully met all design objectives:

  • Robust Stability: The system remained stable despite the $\pm 10\%$ variation in natural frequency.
  • Pointing Accuracy: For a 0.3 Nm step disturbance, the maximum deflection was limited to approximately $3.8 \times 10^{-4}$ radians, meeting the required accuracy.
  • Control Effort: The maximum control torque required was 0.36 Nm, well within the 0.5 Nm limit and slightly more efficient than the $H_\infty$ benchmark (0.41 Nm).
  • Settling Time: The step response settled in less than one minute, satisfying the performance criteria.

Limitations and Future Work

  • Decoupled Dynamics: The paper assumes the roll, pitch, and yaw axes are decoupled. In reality, flexible modes often cause cross-coupling between axes, which would require a Multi-Input Multi-Output (MIMO) QFT approach.
  • Single Flexural Mode: The model only accounts for the first flexural mode. Higher-order modes could potentially impact performance, especially in very large structures.
  • Linearity: The study relies on linear plant descriptions. Future work could explore how QFT handles non-linearities like actuator saturation or dead zones in the context of flexible spacecraft.

Summary updated

Abstract

Quantitative Feedback Theory (QFT) has achieved the status as a powerful design technique for the achievement of assigned performance over the specified range of plant uncertainties. This is due to the relative simple algebra and the transparency of the design methodology. A QFT solution to a Single Input Single Output (SISO) problem of roll stabilization of a flexible satellite is presented in this paper. A simple controller is designed which ensures robustness against parameter variations and disturbances. The results of the study, are compared with those obtained using an H* controller.
BibTeX
@article{sreenatha2023roll,
  author = {A.G. Sreenatha and Mohit Gupta},
  title = {{Roll Stabilisation of a Flexible Spacecraft using Quantitative Feedback Theory}},
  journal = {Journal of Aerospace Sciences and Technologies},
  year = {2023},
  doi = {10.61653/joast.v55i1.2003.853},
}
Figure