Characteristics of identifying linear dynamic models from impulse response data using Prony analysis
The purpose of the study was to investigate the characteristics of fitting linear dynamic models to the impulse response of oscillatory dynamic systems using Prony analysis. Many dynamic systems exhibit oscillatory responses with multiple modes of oscillations. Although the underlying dynamics of such systems are often nonlinear, it is frequently possible and very useful to represent the system operating about some set point with a linear model. Derivation of such linear models can be done using two basic approaches: model the system using theoretical derivations and some linearization method such as a Taylor series expansion; or use a curve-fitting technique to optimally fit a linear model to specified system response data. Prony analysis belongs to the second class of system modeling because it is a method of fitting a linear model to the impulse response of a dynamic system. Its parallel formulation inherently makes it well suited for fitting models to oscillatory system data. Such oscillatory dynamic effects occur in large synchronous-generator-based power systems in the form of electromechanical oscillations. To study and characterize these oscillatory dynamics, BPA has developed computer codes to analyze system data using Prony analysis. The objective of this study was to develop a highly detailed understanding of the properties of using Prony analysis to fit models to systems with characteristics often encountered in power systems. This understanding was then extended to develop general ``rules-of-thumb`` for using Prony analysis. The general characteristics were investigated by performing fits to data from known linear models under controlled conditions. The conditions studied include various mathematical solution techniques; different parent system configurations; and a large variety of underlying noise characteristics.
- Research Organization:
- Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- AC06-76RL01830
- OSTI ID:
- 10114740
- Report Number(s):
- PNL-8450; ON: DE93005607
- Resource Relation:
- Other Information: PBD: Dec 1992
- Country of Publication:
- United States
- Language:
- English
Similar Records
Oscillation Detection Algorithm Development Summary Report and Test Plan
Advances in the identification of electrochemical transfer function models using Prony analysis
Related Subjects
99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE
POWER SYSTEMS
S CODES
I CODES
P CODES
T CODES
COMPUTER PROGRAM DOCUMENTATION
OSCILLATIONS
ELECTRICAL TRANSIENTS
DYNAMICS
MATHEMATICAL MODELS
MAXIMUM-LIKELIHOOD FIT
MATRICES
ALGORITHMS
240100
990200
MATHEMATICS AND COMPUTERS