Adaptive Monitoring, Fault Detection and Diagnostics, and Prognostics System for the IRIS Nuclear Plant
Abstract
Ideally, health monitoring of new, complex engineering systems should occur from initial operation to decommissioning. Health monitoring typically involves a suite of modules, including system monitoring, fault detection, fault diagnostics, and system prognostics. However, for systems which have not yet operated, this is challenging. Most available health monitoring modules are empirically based, meaning they are derived from available historic data. For new system designs, such data simply does not exist. This research proposes an adaptive modeling system which initially builds empirical models from high-fidelity simulated data. This data suffers from the common problems of data simulation caused by complicated physical models mechanisms and simplifying assumptions made in model development. As actual system data becomes available, the empirical models adapt in an automated and intelligent way to account for real-world, nominal data relationships. A key challenge in automatically adaptive empirical models lies in differentiating between faulted operation and nominal operation which is not well-described by the physics-based data. Nominal operation may extend beyond the simulated data for many reasons: the system may be operating in unanticipated environments; the assumptions made in model development may cause inaccuracies in the data; or the relationships modeled may simply be incorrect. Traditional fault detection methods such as those using the sequential probability ratio test are not able to distinguish between unexpected nominal operation and truly faulted operation. However, the main benefit of using adaptive models lies in their ability to accurately learn expanded nominal relationships while detecting and differentiating faulted conditions. For the purposes of accurately adapting a monitoring system, a principal component-based method is proposed to distinguish between these two cases.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 2010
- Accession Number
- ADA562536
Entities
People
- J. W. Hines
- Jamie Coble
- Matt Humberstone
Organizations
- University of Tennessee system