YIP Graph-Based Optimization and Reconfigurable Control for Naval Aircraft Propulsion, Power, and Thermal Management Systems
Abstract
As aircraft systems become more complex and requirements on their performance become more challenging to achieve, tighter integration of Propulsion, Power, and Thermal (PPT) management systems is needed. Conventional design and control approaches rely on federated methods that neglect coupling between PPT systems and/or assume worst-case steady-state behaviors. These practices induce substantial conservatism and are unable to achieve the performance required of future platforms. To address this challenge, the proposed research will create new fundamental methods to coordinate the complex dynamics of PPT systems in both design and real-time control, with the objective of enhancing the capability, resilience, and readiness of military aircraft.Underpinning the proposed research is agraph-based modeling framework that applies conservation equations to capture the network structure of an energy system. This simultaneously models multi-domain (e.g., electrical, thermal, mechanical) and multi-timescale coupling within a common framework. Furthermore, the graph-based approach is modular, allowing complex system architectures to be constructed by interconnecting component models. The result is a system representation that exposes the sparsity of energy conservation for efficient optimization and reveals key parameters for design and control. The structure of this network representation will be leveraged to optimize designs and construct hierarchical control frameworks that adapt to changing model parameters and faults. The technical approach has three key thrusts:Thrust 1 will first create a user-friendly software toolbox for constructing graph-based models of PPT systems. This will support analysis, design, and control of both new aircraft in development and existing platforms in sustainment. Model-based design optimization algorithms will then be formulated to intelligently manage tradeoffs in performance, efficiency, and life-cycle costs, reducing degradation under transient operation while growing capability.Thrust 2 will broaden the utility of the graph-based modeling framework through a physics-informed lifting procedure that generates accurate linearized models while preserving conservation of energy. These models will be embedded in predictive controllers that reconfigure and adapt during operation. This will allow real-time controllers to leverage new degrees of freedom that may be available to next-generation vehicles, such as adaptive cycle engines. It will also provide resilience to damage, faults, and failures. Hierarchical frameworks will coordinate the dynamics of complex energy systems by decomposing decision-making into a network of communicating controllers that govern behavior across multiple timescales, subsystems, and physical domains.Thrust 3 will create parameter identification and fault detection algorithms that provide feedback to guide the reconfiguration and adaptation of controllers from Thrust 2. A non-convex set representation developed by the PI called the hybrid zonotope will be used to bound uncertain parameters and identify faults as a reconfiguration of a graph-based model#s network. Furthermore, hybrid zonotopes will be used to verify the safety and stability of these controllers without introducing the significant conservatism that is typical of other robust control techniques.New fundamental methods will be validated on scaled laboratory testbeds using hardware-in-the-loop to couple physical power and thermal systems to simulated propulsion systems. Initial case studies will focus on design optimization, control, and resilience for hybrid-electric vehicles.Close engagement with industry partners will facilitate technology maturation and transfer, and also strengthen an established workforce pipeline of engineering students to careers supporting the Navy#s science and technology objectives.Approved for Public Release.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 13, 2025
- Source ID
- N000142512051
Entities
People
- Herschel C. Pangborn
Organizations
- Office of Naval Research
- Pennsylvania State University
- United States Navy