Flight Vehicle System Identification in Time Domain
Synopsis:
The scope of application of system identification methods has increased dramatically during the last decade. The advances in modeling and parameter estimation techniques have paved the way to address highly complex, large scale and high fidelity modeling problems. The objective of this two-day course is to review the recent advances in the time-domain methods of system identification from flight data, both from the theoretical and practical viewpoints. Starting from the fundamentals, a systematic approach will be presented to arrive at the solution. Benefits derived from flight validated models applying system identification will be highlighted. The course will provide an overview of key methods of parameter estimation in time domain, cover many examples covering both fixed-wing and helicopter applications, and address model validation in both time and frequency domain. The course will be supplemented with an overview of software tools available.
Key Topics:
- Parameter estimation, Consistency checking and Aerodynamic database validation methodology
- Insight and familiarity with modern time domain techniques and intricacies
- Large scale systems and high fidelity modeling
- Real world problems and possible solutions
- Demonstration of MATLAB based software tool and hand-on experience with test cases
- Establish contacts with leading organization with vast practical experience
Who Should Attend:
The two-day short course is designed for beginners and those wishing to refresh and broaden their knowledge in flight vehicle system identification. Emphasis will be on practical aspects with some theoretical background. Those working in mathematical model identification from experimental data with emphasis on aerodynamic characteristics estimation and model validation will benefit from this course.
Course Information:
Type of Course: Instructor-Led Short Course
Course Level: Intermediate/Advanced
Course Length: 2 days
AIAA CEU's available: Yes
I. Background
A. What is system identification, parameter estimation, and simulation?
B. Systematic Quad-M approach
II. Design of flight maneuvers for system identification
A. Practical approach to design multi-step inputs with an overview of different techniques
III. Kinematic consistency checking of recorded flight test data
A. Stochastic and deterministic approach to determine systematic sensor errors
B. Hand-on experience, covering various steps, with test case using provided software and flight data
IV. Parameter estimation methods in time domain
A. Least-Squares and Output error method
B. Filter error methods accounting for atmospheric turbulence
C. Different optimization methods and practical issues
D. Nonlinear systems and large scale problems
V. Applications
A. Trim point identification
B. Global model (ATTAS, C-160, DO-328, X-31A)
1. Aircraft mass properties
2. Nonlinearities in control surface effectiveness
3. Functional dependencies on flow variables
C. Unstable aircraft
D. Identification of nonlinear phenomenon (ex. Stall hysteresis)
E. On-line parameter estimation
VI. Model and Aerodynamic database validation
VII. Broad classification: Statistical properties, residual analysis, and model predictive quality
VIII. Proof-of-Match procedures for simulator certification based on FAA tolerances
IX. Approach to validate and update pre-flight aerodynamic databases using flight data
X. Demonstration of and hands-on experience with software tools
XI. Matlab based software tool for parameter estimation in the time domain; hands-on experience with test cases on sample flight data
XII. Summary of a modular and integrated software tool ESTIMA (Fortran-77 based) with post processing options
XIII. Wrap-up
A. Summary
B. Open discussion and possibilities of addressing actual problems raised by attendee
Dr. Ravindra Jategaonkar is a senior scientist and a group leader at the German Aerospace Center, DLR-Institute of Flight Systems, Braunschweig, Germany. He has been extensively involved in the development and application of system identification methods to flight vehicle modeling, covering nonlinear models, generation of Level-D aerodynamic databases for flight simulators and estimation in the presence of atmospheric turbulence. He is primarily responsible for the commercially available integrated software tool ESTIMA for parameter estimation from large-scale nonlinear systems and post-processing visualization. He is the author and co-author of more than 40 scientific papers and 45 technical reports, co-author of the invited AIAA survey paper on Evolution of Flight Vehicle System Identification (Journal of Aircraft Vol. 33, Number 1, 1996); and author of a book “Flight Vehicle System Identification: A Time Domain Methodology”, published August 2006 under Progress Series by AIAA. He is recipient of DLR Title “Senior Scientist” for scientific excellence, and of the AIAA Sustained Service Award 2007.
AIAA Training Links
For information, group discounts,
and private course pricing, contact:
Lisa Le, Education Specialist (lisal@aiaa.org)