Aerodynamic modeling using flight data
AssigneeUnited States of America as represented by the Administrator of NASA
InventorsJay Brandon, Eugene A. Morelli
Life of this patent
Estimating the term…
Abstract
Systems and methods are described for development of a nonlinear global aerodynamic model using fuzzy logic modeling as well as multivariate orthogonal function modeling with splines. The systems and methods described herein may be utilized to more quickly develop a global aerodynamic model of an aircraft, allowing for development of a global aerodynamic model of an aircraft in real-time, during a test flight, and reducing time and cost associated with the model development.
Claims20
A method of in-flight production of an aerodynamic model of an aircraft, comprising:
- receiving an indication of control inputs at a full-scale aircraft during a test flight;
- responsive to the control inputs, obtaining sensor data from one or more sensors on the aircraft during the test flight;
- downsampling the sensor data received from the test flight based at least in part on a cutoff frequency to reduce a computation time of the aerodynamic model;
- selecting, in flight, one or more explanatory variables from the downsampled sensor data;
- calculating in-flight, one or more membership functions for each of the one or more explanatory variables;
- defining a plurality of fuzzy cell internal functions, wherein a fuzzy cell internal function from the plurality of fuzzy cell internal functions comprises a membership function from the one or more membership functions for each of the one or more explanatory variables from the downsampled sensor data;
- determining a model output in the time domain of the aerodynamic performance of the aircraft as a weighted output of the plurality of fuzzy cell internal functions, wherein the model output is non-linear and is determined in real-time while the aircraft is airborne and the model output corresponds to all six-rigid body degrees of freedom of the aircraft;
- predicting, in-flight, an aerodynamic performance characteristic of the aircraft using the model output;
- and
- using the model output to impact one or more aircraft maneuvers;
- wherein the maneuvers change the flight conditions of the aircraft.
A non-transitory computer-readable storage medium comprising computer-executable instructions that when executed by a processor are configured to cause a processor to perform, in-flight during a test flight of an aircraft, at least:
- receive, responsive to control inputs to the aircraft, sensor data from one or more sensors on the aircraft;
- select one or more explanatory variables from the sensor data;
- generate a plurality of multivariate functions using the one or more explanatory variables;
- generate a plurality of orthogonal modeling functions from the plurality of multivariate functions;
- generate an output model comprising one or more of the plurality of orthogonal modeling functions that minimize a predicted square error;
- wherein the output model is nonlinear and is determined in real-time while the aircraft is airborne;
- predict in-flight, a flight characteristic of the aircraft using the output model;
- use the output model to impact one or more aircraft maneuvers in a subsequent portion of the test flight, wherein the maneuvers change the flight conditions of the aircraft;
- and
- update, in flight, the plurality of orthogonal modeling functions based on sensor data from the subsequent portion of the test flight with a Bayesian approach.
A computer-implemented method for in-flight production of an aerodynamic model of an aircraft, comprising:
- applying control inputs to the aircraft during a test flight;
- receiving, responsive to the applied control inputs, sensor data from one or more sensors on the aircraft;
- selecting one or more explanatory variables from the sensor data;
- fitting a nonlinear model to the received sensor data as a function of the selected one or more explanatory variables;
- generating an aerodynamics output model comprising the fitted nonlinear model;
- wherein the output model is generated in real-time while the aircraft is airborne;
- displaying, in-flight, the fitted nonlinear model on a display device, such that the fitted nonlinear model associated with the output model is available to a user in real time;
- determining, in flight, at least one adjustment to the test flight to improve the fitted nonlinear model, wherein the at least one adjustment is a change to one or more aircraft maneuvers in a subsequent portion of the test flight;
- and
- displaying, in flight, a message indicating the at least one adjustment.
Dependencies are inferred from the claim text.