Solar Sail Momentum Management With Mass Translation and Reflectivity Devices Using Predictive Control

This paper proposes a novel Model Predictive Control framework for solar sails that effectively manages reaction wheel momentum and attitude by integrating active mass translation and reflectivity control devices through tailored discretization, PWM-inspired quantization, and an iterative backwards-in-time optimization approach to handle coupled nonlinear dynamics and discrete actuator constraints.

The Gist

Imagine a spaceship that doesn't use fuel to move. Instead, it has a giant, ultra-thin sail (like a piece of tissue paper the size of a football field) that catches the "wind" of sunlight. This is a Solar Sail. The pressure from sunlight pushes the sail, allowing the ship to travel through space forever without running out of gas.

However, there's a problem. Just like a real sailboat can get pushed off course by a sudden gust of wind, this solar sail gets pushed by uneven sunlight, structural wobbles, and other space debris. To stay on course, the ship uses spinning wheels (called Reaction Wheels) inside it. These wheels spin faster or slower to nudge the ship back into the right position.

The Big Problem:
Think of these spinning wheels like a person holding a heavy spinning top. If you keep trying to balance a wobbly table by spinning the top faster and faster, eventually, the top will spin so fast it flies out of your hands. In space terms, the wheels "saturate" (spin too fast) and stop working. If they stop working, the ship can't steer anymore.

Usually, spaceships use tiny rocket thrusters to slow these wheels down. But solar sails are supposed to be fuel-free! Using rockets defeats the purpose.

The Paper's Solution:
This paper proposes a new, clever way to slow down those spinning wheels without using fuel. It uses two special tools already on the ship:

1. The Active Mass Translator (AMT): A heavy weight inside the ship that can slide back and forth.
2. Reflectivity Control Devices (RCDs): Tiny switches on the sail that can turn a section of the sail from "shiny" (reflective) to "dark" (absorptive) instantly.

The authors created a super-smart computer brain called Model Predictive Control (MPC) to manage these tools. Here is how they explain it using simple analogies:

1. The "Sliding Weight" (AMT)

Imagine you are balancing a long plank on your finger. If the plank starts tipping to the left, you slide your hand to the left to counterbalance it.

• The AMT is like a heavy box on a track inside the ship. By sliding this box, the ship's center of gravity moves. This creates a gentle push from the sunlight that naturally counteracts the unwanted spinning of the wheels.
• The Innovation: The paper says, "Don't just jerk the box back and forth." Instead, use a First-Order Hold method. Imagine the box doesn't teleport; it glides smoothly from point A to point B. This smooth movement is more efficient and less stressful on the ship's structure.

2. The "Light Switch" (RCDs)

Now imagine the plank is also being pushed by a fan. To stop the plank from spinning, you can flip a switch on the fan to change how much air hits it.

• The RCDs are like light switches on the sail. They can only be ON (reflective) or OFF (dark). They cannot be "half-on."
• The Problem: If you just flip the switch on and off randomly, it's inefficient.
• The Innovation (PWM): The paper suggests using a technique called Pulse-Width Modulation. Imagine you need to let 50% of the air through, but you only have an On/Off switch. You turn the switch ON for half the time and OFF for the other half. The paper's computer brain calculates exactly how long to keep the switch on during each step to get the perfect average push.

3. The "Smart Brain" (MPC)

The real magic is the Model Predictive Control (MPC).

• Old Way (The PID Controller): Imagine a driver who only looks at the car directly in front of them. If they drift left, they turn right. If they drift right, they turn left. They react to what just happened. This often leads to over-correcting and jerky movements.
• The New Way (MPC): Imagine a driver who looks 20 seconds down the road. They see a curve coming up and a gust of wind predicted to hit in 10 seconds. They start turning before the curve and adjust before the wind hits.
• The "Backwards" Trick: The paper introduces a special "Backwards-in-Time" trick. Since the light switches (RCDs) are so rigid (On/Off), the computer solves the problem by planning the end of the sequence first, locking in the best "On/Off" decision, and then working backward to figure out the rest. This ensures the computer doesn't waste time trying to solve impossible math problems.

Why This Matters

The paper ran a simulation of a massive solar sail (like NASA's Solar Cruiser) and compared this new "Smart Brain" against the old "Reactive Driver" method.

• Result: The new method used significantly less movement from the sliding weight and fewer switches for the light devices.
• Benefit: This saves power, reduces wear and tear on the mechanical parts, and keeps the ship's steering wheels from spinning out of control for much longer.

In a Nutshell:
This paper teaches a solar sail how to balance itself not by jerking around, but by gliding its internal weight and flicking its light switches with perfect timing. It uses a computer that looks into the future to make smooth, efficient moves, ensuring the ship can travel deep into space without ever needing to refuel or fix broken parts.

Technical Summary

1. Problem Statement

Solar sails offer propellant-free propulsion for deep-space exploration by utilizing solar radiation pressure (SRP). However, they face significant challenges in attitude control:

• Disturbance Torques: Structural flexibility and imperfections in large-scale sails shift the center of pressure (CP), creating disturbance torques that must be rejected.
• Reaction Wheel Saturation: Standard attitude control uses Reaction Wheels (RWs) to counteract these torques. Over time, RWs accumulate angular momentum, eventually leading to saturation and a loss of control authority.
• Limitations of Current Solutions: Traditional momentum management (e.g., thrusters or magnetic torquers) is unsuitable for solar sails due to fuel requirements or the lack of external magnetic fields.
• Actuator Constraints: NASA's Solar Cruiser concept utilizes two unique actuators for momentum management:
  1. Active Mass Translator (AMT): Moves a portion of the spacecraft bus mass to shift the Center of Mass (CM), generating pitch/yaw torques. This is a continuous actuator with rate and position limits.
  2. Reflectivity Control Devices (RCDs): Thin-film membranes that modulate reflectivity to generate roll torques. These operate in a discrete on-off fashion.
• Control Challenge: Existing control strategies (decoupled PID with on-off thresholds) neglect the dynamic coupling between the AMT and RCDs and struggle with the non-convex, discrete nature of RCD inputs, often leading to suboptimal performance or instability.

2. Methodology

The authors propose a Model Predictive Control (MPC) framework tailored specifically for the unique dynamics and actuation constraints of solar sails.

A. Dynamic Modeling

• Nonlinear Dynamics: A rigorous nonlinear dynamic model was derived that accounts for the time-varying Center of Mass (CM) and Moment of Inertia (MOI) caused by the AMT's translation.
• Linearization: The nonlinear model is linearized around an operational point to create a Linear Time-Varying (LTV) state-space model suitable for real-time MPC implementation.
• Hybrid Discretization: To accurately capture the distinct actuator behaviors, a mixed discretization scheme is used:
  • First-Order Hold (FOH): Applied to the AMT to model its continuous, linear translation between time steps.
  • Zero-Order Hold (ZOH): Applied to the RCDs to model their discrete, constant torque pulses.

B. Control Strategies

Two MPC-based strategies are developed to handle the integer constraints of the RCDs without resorting to computationally expensive Mixed-Integer Programming (MIP):

1. Strategy 1: MPC with Post-Optimization Quantization

   • Solves a standard convex Quadratic Program (QP) assuming continuous RCD inputs.
   • PWM Quantization: The continuous optimal input is converted to a physical on-off pulse using Pulse-Width Modulation (PWM). The pulse duration is proportional to the continuous input magnitude.
   • Thresholding: A deadband threshold is applied to ignore small inputs, preventing unnecessary on-off cycling and extending actuator life.

2. Strategy 2: Iterative Backwards-in-Time MPC

   • Addresses the performance loss caused by post-optimization quantization.
   • Algorithm: The algorithm iteratively fixes the RCD input at the end of the prediction horizon as a quantized PWM value. It then re-optimizes the remaining horizon. This process repeats, moving backward in time, until the first time step is quantized.
   • Benefit: This incorporates the knowledge of quantization constraints directly into the prediction model, improving dynamic fidelity while maintaining the computational efficiency of solving a sequence of convex QPs.

C. Cost Function and Constraints

• Objective: Minimize RW angular momentum accumulation, AMT motion (to save energy), and RCD usage.
• Soft Constraints: Slack variables are used to penalize RW momentum approaching saturation limits, allowing the controller to react preemptively.
• Hard Constraints: Enforced limits on AMT position, velocity, and RCD torque magnitude.

3. Key Contributions

1. Novel Dynamic Model: Derivation of a solar sail dynamic model that explicitly accounts for the time-dependent changes in CM and MOI due to AMT translation.
2. Tailored MPC Formulation: Development of a hybrid discretization (FOH for AMT, ZOH for RCDs) and a convex QP formulation that avoids MIP, making real-time onboard implementation feasible.
3. Iterative Backwards-in-Time Algorithm: A computationally efficient method to integrate discrete actuator constraints (PWM quantization) into the MPC optimization loop, significantly improving performance over standard post-processing methods.
4. Performance Validation: Demonstration that the proposed MPC strategies significantly outperform the state-of-the-art decoupled PID approach used in current Solar Cruiser designs.

4. Simulation Results

The authors conducted numerical simulations using parameters from NASA's Solar Cruiser (sail area >1600 m², specific mass properties, and disturbance torques).

• Scenario: The spacecraft performed attitude maneuvers while subject to constant worst-case disturbance torques.
• Comparison: The proposed MPC strategies (with and without the iterative method) were compared against the standard PID-based momentum management strategy.
• Key Findings:
  • Actuator Efficiency: The MPC strategies significantly reduced AMT travel distance (by ~75-80% compared to PID) and RCD on-time.
  • RCD Cycling: While the total number of RCD on-off cycles varied, the total time the RCDs were active was lower in MPC strategies, reducing wear and power consumption.
  • Momentum Management: All MPC strategies successfully prevented RW saturation, whereas the PID-only attitude control (without momentum management) led to rapid saturation.
  • Iterative vs. Non-Iterative: The iterative backwards-in-time strategy (Strategy 2) showed superior performance during transient phases by better anticipating quantization effects, though it required slightly more AMT movement in steady state compared to the thresholded non-iterative approach.
  • Robustness: The controllers remained stable even with ±50% uncertainty in disturbance estimation, though overestimating disturbances yielded better performance.

5. Significance

This work represents a significant advancement in solar sail control systems. By moving from decoupled, heuristic PID control to a unified, predictive framework, the paper demonstrates:

• Feasibility of Complex Control: It proves that handling coupled, nonlinear dynamics with discrete actuator constraints is possible within the limited computational resources of a spacecraft.
• Mission Longevity: The drastic reduction in actuator usage (travel distance and on-time) directly translates to extended mission lifespans and reduced power consumption, critical for long-duration deep-space missions.
• Scalability: The methodology is applicable to other spacecraft utilizing mass-shifting actuators or discrete control surfaces, offering a robust solution for momentum management in fuel-free propulsion systems.