|InterJournal Complex Systems, 1776
|Manuscript Number: |
Submission Date: 2006
|Eigen Analysis of Model Based Residual Spectra for Fault Diagnostics Techniques|
The energy distribution of Residual Spectra generated by Model Based Parity Space relationship drifts in the form of Eigen value. The Eigen vector in such a multi-dimensional Residual Space is used to maintain the degree and polarity of drift. This paper presents investigations into the issues related to such Eigen analysis. It was found that normalized residuals from multiple sources and parity space relations are neutralized in the form of unified representation of energy that can be used to form a generic framework for fault detection and isolation. The case studies into multi-sensor system were carried out and actual data from sensors is used to test this generic framework. It is being investigated, how the proper modeling of qualitative entities as energy, could lead to unified and neutral residual space while keeping the implementation cost reasonably low. Key words: Fault Diagnostics, Single Throw Mechanical Equipment (STME), Parameter Estimation, Eigen Values of Faults, Residual generation, Parity-Space, Low-cost limit sensors, Adaptive thresholds.
|Submit referee report/comment|