Stochastic design optimization of modular, reconfigurable, persistent support platforms in Earth orbit

Perez, Jeff

Permalink

https://hdl.handle.net/2142/104870 Copy

Description

Title

Stochastic design optimization of modular, reconfigurable, persistent support platforms in Earth orbit

Author(s)

Perez, Jeff

Issue Date

2019-04-23

Director of Research (if dissertation) or Advisor (if thesis)

Ho, Koki

Department of Study

Aerospace Engineering

Discipline

Aerospace Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• Stochastic design optimization
• hosted payload platform
• persistent support platform
• modular satellite
• spacecraft design optimization
• multistage stochastic programming
• reconfigurable
• on-orbit satellite servicing
• servicing spacecraft
• robotic servicer
• reconfigurable satellite

Abstract

This thesis focuses on the design and optimization of modular, reconfigurable, hosted payload platforms operating in Earth orbit. Recent advancements in on-orbit servicing technologies and robotics are creating a market for hosted-payload platforms which can support multiple payloads with varying requirements. Such platforms can employ on-orbit servicing and robotic manipulation to repair or replace modules, enhance the platform’s capabilities over time, and reconfigure modules to optimize performance. Traditional spacecraft design is often driven largely by payload requirements. For the case of persistent platforms, however, not all payloads will be known in the initial design phase. This presents a unique challenge to designers, who must account for the uncertainty of future payloads by trading off between the costs of adding more capability to the platform initially, which assumes the risk of wasted costs due to over-designing the platform, and the costs of utilizing an on-orbit servicer to add capability as needed. The hosted payload platforms considered in this thesis consist of platform modules and payload modules and uses a standardized interface for intermodular and customer payload connection. Each platform module contains a critical satellite subsystem that is necessary for on-orbit functionality. As payloads are added to the platform over time, their demands may exceed the current capability of the platform, at which point additional platform modules can be added to increase the platforms capabilities. This thesis proposes an approach using a multi-stage stochastic programming method to create an initial platform design that is robust and flexible enough to support a wide range of payloads and minimizes the expected costs of future platform additions. Probability distributions for future payload selections are created based on a survey of active satellites. These distributions are then used to create samples of payload selection scenarios. Using a simple cost model, the expected costs associated with the addition of new payloads and the required platform modules are computed for each scenario in the sample. A genetic algorithm is used to find an optimal initial platform size that minimizes the combined total of the initial cost of the platform and the expected on-orbit servicing costs associated with adding future payloads and platform modules for each scenario. Platform designs are compared for a range of on-orbit servicing costs to determine the cost at which the optimizer begins to utilize servicing over adding more capability initially. Finally, a sensitivity analysis is performed to assess the variations in platform design due to the randomly selected payload scenario samples. The results of this work are a first step towards a solving a unique challenge presented by an emerging and increasingly relevant mission concept.

Graduation Semester

2019-05

Type of Resource

text

Permalink

http://hdl.handle.net/2142/104870

Copyright and License Information

Copyright 2019 Jeff Perez

Owning Collections