Hybrid Electric Propulsion Power Split via Soft Actor-Critic RL
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current hybrid electric propulsion (HEP) systems for aircraft face challenges in optimizing fuel consumption and battery life, particularly during varying flight durations where reliance on electric power alone is not feasible.
Innovation Solution
The method involves using a soft actor-critic agent with neural networks and a control barrier function (CBF) filter to generate a power splitting profile between electric motor and gas turbine power, optimizing fuel consumption while maintaining battery health.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If charge-depleting charge-sustaining approach is used to maximize electric power utilization, then fuel consumption is reduced, but battery life is adversely affected
Solution Approach 1:
The patent dynamically changes the power splitting ratio parameter between gas turbine and electric motor based on real-time battery state of charge, flight phase, and power demand. This allows the system to optimize fuel consumption while preventing battery over-discharge, thereby extending battery life. The controller adjusts the electric motor power contribution according to predefined operating regions that consider battery health constraints.
Solution Approach 2:
The patent implements a feedback control mechanism where the controller continuously monitors battery state of charge, power demand, and flight conditions. Based on this feedback, the system adjusts the power splitting strategy in real-time to maintain battery operation within safe charge ranges, preventing excessive discharge that would harm battery life while still maximizing fuel efficiency during appropriate flight phases.
2Use of energy by moving object
If electric power alone is used for short flights, then fuel consumption is minimized, but this approach is not feasible for varying flight durations
Solution Approach 1:
The patent employs a dynamic power splitting strategy that adapts to varying flight durations and power demands. The controller divides the flight profile into different phases (climb, cruise, descent) and adjusts the gas turbine-electric motor power ratio dynamically for each phase. This allows the system to maximize electric motor utilization during phases where it is most efficient while ensuring sufficient gas turbine contribution for longer flights or high-power requirements.
Solution Approach 2:
The patent creates a universal power management system that can handle various flight durations and conditions through a unified control architecture. The power splitting algorithm is designed to work across different flight scenarios by considering overall flight phase, instantaneous power demand, and battery state of charge simultaneously, making the system adaptable to both short and long flights without requiring separate control strategies.
3Use of energy by moving object
If complex power management strategies are implemented to optimize fuel consumption, then fuel efficiency improves, but system complexity increases
Solution Approach 1:
The patent segments the flight profile into distinct phases (climb, cruise, descent) and applies specific power splitting strategies for each phase. The flight phase is determined based on aircraft speed, altitude, and power demand characteristics. This segmentation allows the complex optimization problem to be broken down into manageable sub-problems, each with its own optimized control logic, reducing overall system complexity while maintaining fuel efficiency.
Solution Approach 2:
The patent introduces a power splitting ratio as an intermediary parameter that mediates between gas turbine output and electric motor input. This single controlling parameter simplifies the coordination of multiple power sources by reducing the control problem to optimizing one key variable (the power split ratio) rather than independently controlling multiple components, thereby reducing control system complexity while achieving fuel efficiency optimization.
Data Source
Figure 1
Figure 2A~2C
Figure 3
AI summary
A method (100) for a hybrid electric propulsion, HEP, system (10) includes utilizing a soft actor-critic agent (62), which includes at least one neural network, and a control barrier function, CBF, filter (64) to obtain a power splitting profile for an HEP system. The power splitting profile includes an electric motor power for an electric motor of the HEP system and a gas turbine power for a gas turbine of the HEP system. The electric motor power and gas turbine power collectively provide a combined HEP output power. The method also includes, during a flight of an aircraft that includes the HEP system, performing (108) an output action for the HEP system based on the power splitting profile. The utilizing is performed based on a predefined fuel consumption objective and a state of charge of at least one battery that powers the electric motor. A system (60) for a HEP system (10) is also disclosed.