Robust and adaptive planning under deep uncertainties for space-based solar power
Robust and adaptive planning under deep uncertainties for space-based solar power
Category: Novel system concepts for SPS for applications on Earth, the Moon, or Mars
All authors
Abstract
Space-based Solar Power (SPS) has the potential to offer clean and abundant energy for the future needs of humanity. Planning and deploying such a huge and long-term project requires tremendous effort, from the development and demonstration of the necessary technologies, to its incremental deployment and its operational management. This must be done in a very complex environment under deep uncertainties such as the performance of alternative power sources, future power demand, technology development, climate change effects, etc. How can we then plan for the development and deployment of such a long term space project under these conditions? A lot of these uncertainties permeating many aspects of the problem are not well defined and cannot be described probabilistically or predicted adequately. Existing qualitative forecasting and planning methods involving expert deliberation are not able to deal entirely with such problems and might introduce various biases. Quantitative methods usually rely on probabilistic models and aim at prediction, an approach that can be counterproductive when studying highly complex problems. A novel approach called Decision Making Under Deep Uncertainty (DMDU) has been developed to deal with such problems. Within this suite, Robust Decision Making (RDM) is a method used to design robust and adaptive (rather than strictly optimal) strategies, by considering a multiplicity of plausible futures. In RDM computer models are used to facilitate human deliberation over exploration, alternatives, and trade-offs. RDM has been applied to similarly challenging planning problems such as climate change mitigation, infrastructure planning, etc. We will apply RDM to the problem of planning for SPS to, among others, a) identify vulnerabilities and key factors in the SPS planning problem, b) offer trade-offs between different strategies and finally c) give dynamic adaptive strategy pathways useful for the incremental deployment of the SPS programme.
ESA
konstantinos.konstantinidis@esa.int
- Co-funded research
Research question: How can we plan for the development, deployment, and operation of a large and long term space project such as space-based solar power (SPS) under the conditions of deep uncertainty that permeate many aspects of this problem? How can we do this in a way that is robust to these deeply uncertain parameters and to the many plausible futures?
Short research plan: Following the RDM process, the problem must be first broken down into a few categories: The actions or “policy levers” available to the decision maker, the environment in which the problem evolves and the “exogenous factors” affecting it, and the “measures of performance” of the system for the decision maker. These blocks must be connected by a series of relationships, usually given in the form of computer models. Moreover, uncertainties may affect each of the above parameters and the connections between them. In the problem of planning for SPS, the policy levers will be, among others, the gradual development, deployment and operation of a set of SPS mission architectures, collected by literature research and further developed where needed. The exogenous environment in which an SPS architecture operates will be the energy supply environment comprising of the various Earth-based energy sources. Two models will be combined to simulate their interaction: on one hand, parametric models will be used for a cost and schedule estimation for technology development and demonstration of each SPS architecture and its elements. On the other hand, for the energy environment we will use one of the various open source or internal models that exist. As an SPS architecture is then gradually deployed, it takes its place among the energy sources in this model, and its performance can be assessed. Using the above integrated model, we will then perform simulations to evaluate proposed strategies in each of many plausible paths into the future. Using the resulting data we will characterize vulnerabilities, and identify the key factors that best distinguish futures in which proposed policies meet or miss their goals. With this information we will evaluate multi-parametric trade-offs to best balance between the competing objectives. Finally, based on the above results we will design an adaptive strategy as a set of pathways that represents how the decision maker might move from one policy to another in response to evolving conditions.
Kostas Konstantinidis is an Internal Research Fellow in mission analysis at the ESA Advanced Concepts Team (ACT). He has extensive experience in systems engineering of ambitious space projects, including a broad study of architectures for in-situ astrobiology on the icy moons of the giant planets, the design of a concept for a lander mission to Saturn's moon Enceladus, the design of a post-mission disposal kit for space debris mitigation, and several other solar system exploration and Earth orbiting mission concepts. At the ACT he has investigated the application of optimization methods and computational approaches to trajectory design, space systems engineering, and space policy. His research at the ACT further includes the design of advanced GN&C and the study of complex systems in space.
Thomas Hamacher is a Full Professor for Renewable and Sustainable Energy Systems and the Director of the Munich School of Engineering at the Technical University of Munich (TUM). He has served as head of the Energy and System Studies Group of the Max Planck Institute for Plasma Physics and as acting head of the Chair of Energy Management and Application Technology of TUM. He is also a member of the Environmental Science Centre (WZU) of the University of Augsburg. His research focuses on innovative energy and systems analysis, and the methods and fundamentals of energy models and new optimization methods. The Chair of Renewable and Sustainable Energy Systems (ENS) which prof. Hamacher leads focuses on energy system modeling. It develops models for different scales of time and space to describe and understand present and future transition processes. Moreover, it uses advanced methods for modeling technical and economical interactions to find optimal solutions with regard to economic benefits, external costs and environmental impacts.
Anne Mergy is a Young Graduate Trainee in artificial intelligence at the ESA Advanced Concepts Team (ACT). She has a background in machine learning and in optimization. She has acquired a good knowledge of the energy sector thanks to previous experiences working for an energy consumption optimization start-up and for an electric utility company, in which she has worked on the acceleration of an optimization algorithm using machine learning algorithms. At the ACT, she is working on generative modelling for space data.
Advenit Makaya; Aidan Cowley
2nd Round idea
No Data to Display
STATISTICS
- Oct 30, 2020
- 1,536 Views
- 88 Visitors
- 0 Comments
- 2 Likes
- 1 Follower