Power plant vibration monitoring using wavelet feature extraction and functional design of experiments
No Thumbnail Available
Authors
Meeting name
Sponsors
Date
Journal Title
Format
Thesis
Subject
Abstract
In a nuclear plant power generation, analysis of vibration signal constitutes an integral part of predictive maintenance for rotating equipment such as pumps, motors, turbine generators, etc. Vibration signals are continuously monitored via sensors and thresholds for alarms maybe set up to identify equipment malfunction. Improved methods for decomposition and analysis of power plant vibration signals using wavelet feature extraction and functional design of experiment (FDOE) have not been sufficiently investigated for applicability in analyzing these signals for better detection of equipment faults. Chapter 1 introduces the general concepts and methods to be applied in our research study. In Chapter 2, we present the application of discrete wavelet transform (DWT) to decompose a reactor coolant pump vibration signal into frequency sub-bands and the generation of a number of features from some statistics. A principal component analysis (PCA) is used to reduce the large set of variables into a few principal components which can be applied in future monitoring of normal vibration signals. From the insights gained using PCA, Chapter 3 studies the linear discriminant analysis (LDA) to simulated vibration signals, to distinguish between normal and abnormal signals. In Chapter 4, we apply the functional principal component analysis (FPCA) in characterizing the vibration signals generated under several different levels of an environmental factor, a flow rate, associated with a condensate pump. An FDOE is applied with the target vibration curve and used to obtain an optimal flow rate. The obtained flow rate was found comparable to the theoretical pump curve best efficiency point (BEP) and recommended for use for optimal pump performance and reliability. In Chapter 5, we perform an extensive review of literature on FDOE and its applications. A standard FDOE framework was shown and demonstrated the five basic steps to be applied when using JMP Pro 17. Chapter 6 provided overall conclusions and suggestions for future work.
Table of Contents
DOI
PubMed ID
Degree
Ph. D.