Diagnosis of nonlinear systems using time series analysis
Diagnosis and analysis techniques for linear systems have been developed and refined to a high degree of precision. In contrast, techniques for the analysis of data from nonlinear systems are in the early stages of development. This paper describes a time series technique for the analysis of data from nonlinear systems. The input and response time series resulting from excitation of the nonlinear system are embedded in a state space. The form of the embedding is optimized using local canonical variate analysis and singular value decomposition techniques. From the state space model, future system responses are estimated. The expected degree of predictability of the system is investigated using the state transition matrix. The degree of nonlinearity present is quantified using the geometry of the transfer function poles in the z plane. Examples of application to a linear single-degree-of-freedom system, a single-degree-of-freedom Duffing Oscillator, and linear and nonlinear three degree of freedom oscillators are presented. 11 refs., 9 figs.
- Research Organization:
- Los Alamos National Lab., NM (United States)
- Sponsoring Organization:
- DOE; USDOE, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 5242934
- Report Number(s):
- LA-UR-91-2283; CONF-911241--2; ON: DE91016041
- Country of Publication:
- United States
- Language:
- English
Similar Records
Nonlinear system vibration---The appearance of chaos
State analysis of nonlinear systems using local canonical variate analysis
Related Subjects
71 CLASSICAL AND QUANTUM MECHANICS
GENERAL PHYSICS
99 GENERAL AND MISCELLANEOUS
990200 -- Mathematics & Computers
ALGORITHMS
DATA ANALYSIS
DEGREES OF FREEDOM
ELECTRONIC EQUIPMENT
EQUIPMENT
FREQUENCY ANALYSIS
ITERATIVE METHODS
MATHEMATICAL LOGIC
NONLINEAR PROBLEMS
OSCILLATORS