Applied fault detection and diagnosis for industrial gas turbine systems

Abstract

Abstract:The paper presents readily implementable approaches for fault detection and diagnosis (FDD) based on measurements from multiple sensor groups, for industrial systems. Specifically, the use of hierarchical clustering (HC) and self-organizing map neural networks (SOMNNs) are shown to provide robust and user-friendly tools for application to industrial gas turbine (IGT) systems. HC fingerprints are found for normal operation, and FDD is achieved by monitoring cluster changes occurring in the resulting dendrograms. Similarly, fingerprints of operational behaviour are also obtained using SOMNN based classification maps (CMs) that are initially determined during normal operation, and FDD is performed by detecting changes in their CMs. The proposed methods are shown to be capable of FDD from a large group of sensors that measure a variety of physical quantities. A key feature of the paper is the development of techniques to accommodate transient system operation, which can often lead to false-alarms being triggered when using traditional techniques if the monitoring algorithms are not first desensitized. Case studies showing the efficacy of the techniques for detecting sensor faults, bearing tilt pad wear and early stage pre-chamber burnout, are included. The presented techniques are now being applied operationally and monitoring IGTs in various regions of the world.Keywords: Fault detection and diagnosis, hierarchical clustering, self-organizing map neural network. Introduction 1The purpose of fault detection is to automatically generate an 'alarm' or 'flag' to inform operators of impending or developing failure, whilst fault diagnosis aims to identify the location and predict the consequences of the failure [1] . The adoption of 'early warning' systems to identify and localize emerging faults has therefore attracted considerable attention due to the widely-recognized benefits of facilitating reduced down-time and assurance of safety, through the use of fault detection and diagnosis (FDD) [2,3] algorithms.Of the methods previously explored to date, FDD techniques can be broadly divided into three categories viz. knowledge-, model-and signal processing-based approaches [3,4,5] . Knowledge-based approaches often rely on monitoring residuals Manuscript received date; revised date * Corresponding author:Tel: +44 1522 837912; Email:cbingham@lincoln.ac.uk between multiple sensor measurements [6] , however, due to the high number of sensors used on modern industrial gas turbines (IGTs) and other complex industrial systems, the adoption of additional redundant sensors is prohibitively expensive. When using model-based approaches, a virtual sensor (a 'model' by some description) is employed to provide an estimate of expected measurements, from which residuals are then used as an indicator of potential failure modes being present [3] . However, for large IGT systems, which are often custom-designed to meet individual orders, the use of application specific materials and components (for example, to satisfy off-s...

A Review on Gas Turbine Gas-Path Diagnostics: State-of-the-Art Methods, Challenges and Opportunities

“…There is inconsistency in the literature on the terminology and definition of fault diagnostics. Some of the commonly used terminologies are fault diagnostics [60,61], fault detection and isolation (FDI) [62,63], fault detection and diagnostics (FDD) [64,65], fault detection, isolation, and identification (FDII) [66], fault detection, isolation and accommodation (FDIA) [67,68], fault detection, isolation and recovery (FDIR) [69] and identification and fault diagnostics [70]. This makes it difficult to understand the goals of the contributions and to compare the different techniques.…”

“…In this analysis, eight different fault cases were taken into account, which may cause the failure of three components: oil gauge, needle valve, and delivery valve. The performance of a hierarchical clustering (HC) and SOM based fault detection and diagnosis method has been evaluated by Zhang et al [64] using measurement deviations obtained from a group of 19 and 16 sensors of a single shaft industrial gas turbine. As demonstration case, fault scenarios, sensor faults, bearing tilt pad wear, and early-stage pre-chamber burnout were considered.…”

A Knowledge Transfer Platform for Fault Diagnosis of Industrial Gas Turbines

“…These approaches do not rely on prior expert knowledge, but rely on the precision and richness of operational and historical data collected on the mechanical system. Latent features are then extracted from the sensor measurements to characterise the health and fault conditions of the system [3,4].…”

Fault Detection for Gas Turbine Hot Components Based on a Convolutional Neural Network

“…Obviously, by simply monitoring the EGT discrepancies, the methods above cannot detect the faults as early as possible.Therefore, many researchers have focused on the fault detection method of the hot components based on the EGT model [11][12][13][14][15][16][17][18][19][20][21]. A normality EGT model is built [11][12][13][14][15][16][17][18][19][20][21]. If the estimated EGT based on the model is different from the actual EGT, the abnormalities of the hot components are identified.…”

“…Yan [15,16] took advantage of stacked denoising autoencoder to extract feature from EGT and used extreme learning machines (ELM) to detect abnormalities in combustors. Hierarchical clustering and self-organizing map neural networks were used for gas turbine pre-chamber burnout anomaly detection [17].Some EGT physical models were also presented. Basseville et al [18,19] presented an EGT physical model.…”