Adaptive model predictive control for hybrid electric propulsion

1. A hybrid electric propulsion system comprising:

a gas turbine engine comprising at least one compressor section and at least one turbine section operably coupled to a shaft;

an electric motor configured to augment rotational power of the shaft of the gas turbine engine; and

a controller operable to:

determine a plurality of hybrid electric propulsion system parameters based on a composite system model and sensor data from one or more sensors;

determine a prediction based on the hybrid electric propulsion system parameters and the composite system model;

determine a model predictive control optimization for a plurality of hybrid electric system control effectors based on the prediction using a plurality of reduced-order partitions of the composite system model, wherein the reduced-order partitions comprise partitions of a propulsion system model comprising a gas turbine engine model, a mechanical power transmission model, and an electrical power system model;

update a plurality of composite system model states of the composite system model based on detection of one or more faults, wherein the reduced-order partitions of the composite system model comprise a plurality of reduced-order partitions of the composite system model states; and

actuate the hybrid electric system control effectors based on the model predictive control optimization.

2. The hybrid electric propulsion system of claim 1 , wherein the controller is further configured to update one or more reduced-order values based on the reduced-order partitions of the composite system model states.

3. The hybrid electric propulsion system of claim 2 , wherein the one or more reduced-order values are reduced-order Jacobian values based on a plurality of Jacobian equations associated with the composite system model.

4. The hybrid electric propulsion system of claim 3 , wherein the reduced-order partitions preserve a plurality of dominant states for each partition.

5. The hybrid electric propulsion system of claim 4 , wherein the composite system model comprises the propulsion system model, an optimization objective function, and a plurality of constraints.

6. The hybrid electric propulsion system of claim 5 , wherein the Jacobian equations associated with the composite system model comprise a plurality of model sensitivity matrices that are updated based on the detection of one or more faults.

7. The hybrid electric propulsion system of claim 6 , wherein the model predictive control optimization uses the model sensitivity matrices to determine a set of changes to the hybrid electric system control effectors that optimizes the optimization objective function over a finite time horizon while maintaining the constraints.

8. The hybrid electric propulsion system of claim 1 , further comprising an electric generator configured to extract power from the shaft, wherein the composite system model comprises a plurality of electrical and mechanical physics-based models of at least the gas turbine engine, the electric motor, the electric generator, and one or more mechanical power transmissions.

9. A hybrid electric propulsion system comprising:

a gas turbine engine;

an electrical power system;

a mechanical power transmission operably coupled between the gas turbine engine and the electrical power system;

a plurality of hybrid electric system control effectors operable to control a plurality of states of one or more of the gas turbine engine and the electrical power system; and

a controller configured to control the hybrid electric system control effectors based on a model predictive control that is dynamically updated during operation of the hybrid electric propulsion system, the controller operable to:

determine a plurality of hybrid electric propulsion system parameters based on a composite system model and a plurality of sensor data from one or more sensors;

determine a prediction based on the hybrid electric propulsion system parameters and the composite system model;

determine a model predictive control optimization for the hybrid electric system control effectors based on the prediction using a plurality of reduced-order partitions of the composite system model, wherein the reduced-order partitions comprise partitions of a propulsion system model comprising a gas turbine engine model, a mechanical power transmission model, and an electrical power system model;

update a plurality of composite system model states of the composite system model based on detection of one or more faults, wherein the reduced-order partitions of the composite system model comprise a plurality of reduced-order partitions of the composite system model states; and

actuate the hybrid electric system control effectors based on the model predictive control optimization.

10. The hybrid electric propulsion system of claim 9 , wherein the controller is further configured to update one or more reduced-order values based on the reduced-order partitions of the composite system model states.

11. The hybrid electric propulsion system of claim 10 , wherein the one or more reduced-order values are reduced-order Jacobian values based on a plurality of Jacobian equations associated with the composite system model, and the reduced-order partitions preserve a plurality of dominant states for each partition.

12. The hybrid electric propulsion system of claim 11 , wherein the composite system model comprises the propulsion system model, an optimization objective function, and a plurality of constraints, and the Jacobian equations associated with the composite system model comprise a plurality of model sensitivity matrices that are updated based on the detection of one or more faults.

13. The hybrid electric propulsion system of claim 12 , wherein the model predictive control optimization uses the model sensitivity matrices to determine a set of changes to the hybrid electric system control effectors that optimizes the optimization objective function over a finite time horizon while maintaining the constraints.

14. The hybrid electric propulsion system of claim 9 , wherein the electrical system comprises at least two electric motors, at least two electric generators, and an energy storage system.

15. A method for controlling a hybrid electric propulsion system, the method comprising:

determining, by a controller, a plurality of hybrid electric propulsion system parameters based on a composite system model and a plurality of sensor data from one or more sensors;

determining, by the controller, a prediction based on the hybrid electric propulsion system parameters and the composite system model;

determining, by the controller, a model predictive control optimization for a plurality of hybrid electric system control effectors based on the prediction using a plurality of reduced-order partitions of the composite system model, wherein the reduced-order partitions comprise partitions of a propulsion system model comprising a gas turbine engine model, a mechanical power transmission model, and an electrical power system model;

updating a plurality of composite system model states of the composite system model based on detection of one or more faults, wherein the reduced-order partitions of the composite system model comprise a plurality of reduced-order partitions of the composite system model states; and

actuating, by the controller, the hybrid electric system control effectors based on the model predictive control optimization.

16. The method of claim 15 , further comprising:

updating one or more reduced-order values based on the reduced-order partitions of the composite system model states.

17. The method of claim 16 , wherein the one or more reduced-order values are reduced-order Jacobian values based on a plurality of Jacobian equations associated with the composite system model, and the reduced-order partitions preserve a plurality of dominant states for each partition.

18. The method of claim 17 , wherein the composite system model comprises the propulsion system model, an optimization objective function, and a plurality of constraints, and the Jacobian equations associated with the composite system model comprise a plurality of model sensitivity matrices that are updated based on the detection of one or more faults, wherein the model predictive control optimization uses the model sensitivity matrices to determine a set of changes to the hybrid electric system control effectors that optimizes the optimization objective function over a finite time horizon while maintaining the constraints.