scholarly journals Long Short-Term Memory Network-Based Normal Pattern Group for Fault Detection of Three-Shaft Marine Gas Turbine

Energies ◽  
2020 ◽  
Vol 14 (1) ◽  
pp. 13
Author(s):  
Mingliang Bai ◽  
Jinfu Liu ◽  
Yujia Ma ◽  
Xinyu Zhao ◽  
Zhenhua Long ◽  
...  
Keyword(s):  
Fault Detection ◽  
Gas Turbine ◽  
Gas Turbines ◽  
Short Term Memory ◽  
Fault Detection And Diagnosis ◽  
Normal Pattern ◽  
Long Short Term Memory ◽  
Detection And Diagnosis ◽  
Lstm Network ◽  
Pattern Group

Fault detection and diagnosis can improve safety and reliability of gas turbines. Current studies on gas turbine fault detection and diagnosis mainly focus on the case of abundant fault samples. However, fault data are rare or even unavailable for gas turbines, especially newly-run gas turbines. Aiming to realize fault detection with only normal data, this paper proposes the concept of normal pattern group. A group of long-short term memory (LSTM) networks are first used for characterizing the mapping relationships among measurable parameters of healthy three-shaft gas turbines. Experiments show that the proposed method can detect all 13 common gas path faults of three-shaft gas turbines sensitively while remaining low false alarm rate. Comparison experiment with single normal pattern model verifies the necessaries and superiorities of using normal pattern group. Meanwhile, comparison between LSTM network and other methods including support vector regression, single-layer feedforward neural network, extreme learning machine and Elman recurrent neural network verifies the superiorities of LSTM network in fault detection. Furthermore, comparison experiment with four common one-class classifiers further verifies the superiorities of the proposed method. This also indicates the superiorities of data-driven methods and gas turbine principle fusion to some extent.

scholarly journals Fault Detection and Diagnosis Using Combined Autoencoder and Long Short-Term Memory Network

Sensors ◽  
2019 ◽  
Vol 19 (21) ◽  
pp. 4612 ◽  
Author(s):  
Pangun Park ◽  
Piergiuseppe Di Marco ◽  
Hyejeon Shin ◽  
Junseong Bang
Keyword(s):  
Time Series ◽  
Fault Detection ◽  
Short Term Memory ◽  
Fault Detection And Diagnosis ◽  
Short Term ◽  
Term Memory ◽  
Long Short Term Memory ◽  
Detection And Diagnosis ◽  
Lstm Network ◽  
Types Of Faults

Fault detection and diagnosis is one of the most critical components of preventing accidents and ensuring the system safety of industrial processes. In this paper, we propose an integrated learning approach for jointly achieving fault detection and fault diagnosis of rare events in multivariate time series data. The proposed approach combines an autoencoder to detect a rare fault event and a long short-term memory (LSTM) network to classify different types of faults. The autoencoder is trained with offline normal data, which is then used as the anomaly detection. The predicted faulty data, captured by autoencoder, are put into the LSTM network to identify the types of faults. It basically combines the strong low-dimensional nonlinear representations of the autoencoder for the rare event detection and the strong time series learning ability of LSTM for the fault diagnosis. The proposed approach is compared with a deep convolutional neural network approach for fault detection and identification on the Tennessee Eastman process. Experimental results show that the combined approach accurately detects deviations from normal behaviour and identifies the types of faults within the useful time.

Development of a Companion Set of Charts—Soft Sensor and Directional Moving Range—for Fault Monitoring, Detection and Diagnosis With Application to Gas Turbine Engines

2008 ◽  
Author(s):  
Dorin Scheianu ◽  
Phillip A. Farrington
Keyword(s):  
Gas Turbine ◽  
Gas Turbines ◽  
Research Literature ◽  
Soft Sensor ◽  
Fault Detection And Diagnosis ◽  
Critical Parameters ◽  
Multivariate Statistical ◽  
Process Variability ◽  
Monitoring Process ◽  
Detection And Diagnosis

Gas turbines monitoring for fault detection and diagnosis is long desired to be embedded within control systems. Yet the general approach is to have alarms and shut downs when critical parameters exceed certain limits, and fault diagnosis is initiated on the behalf of experienced professionals and testing apparatus during scheduled maintenance time. Statistical methods for monitoring univariate and multivariate processes have been developed and publicized in the research literature. A gas turbine can be treated as a complex multivariate process with parameters depending both on control variables imposed by operator and on independent ambient parameters. The authors propose a set of companion charts that can be implemented on line and allows continuous monitoring both for fault amplitude — represented by a newly introduced soft sensor — and for process variability in the direction of interest. The control limits are introduced using multivariate statistical theory. The set of charts was applied at Wood Group LIT in a test cell, for monitoring process variability and for diagnosis and characterization of engine faults during tests. A second application is used for early detection of faults at the current serviced fleet of turbines.

Multiple Fault Detection and Diagnosis in a Gas Turbine Using Nonlinear Principal Component Analysis and Structured Residuals

2013 ◽  
Author(s):  
Amilcar Rincon-Charris ◽  
Joseba Quevedo-Casin
Keyword(s):  
Principal Component Analysis ◽  
Fault Detection ◽  
Gas Turbine ◽  
Principal Component ◽  
Component Analysis ◽  
Fault Detection And Diagnosis ◽  
Multiple Fault ◽  
Nonlinear Principal Component Analysis ◽  
Multiple Faults ◽  
Detection And Diagnosis

Multiple fault detection and diagnosis is a challenging problem because the number of candidates grows exponentially in the number of faults. In addition, multiple faults in dynamic systems may be hard to detect, because they can mask or compensate each other’s effects. This paper presents the study of the detection and diagnosis of multiple faults in a SR-30 Gas Turbine using nonlinear principal component analysis as the detection method and structured residuals as the diagnosis method. The study includes developing a mathematical model, software simulation with Matlab Simulink and implementation of algorithms for detection and diagnosis of multiple faults in the system using nonlinear principal component analysis and structured residuals. A real SR-30 gas turbine was used for our studies. The equipment is at the moment installed in the Inter American University of Puerto Rico, Bayamon Campus, and Department of Mechanical Engineering.

Comparison of a Few Fault Diagnosis Methods on Sparse Variable Length Time Series Sequences

2007 ◽  
Author(s):  
Lokesh Kumar Sambasivan ◽  
Venkataramana Bantwal Kini ◽  
Srikanth Ryali ◽  
Joydeb Mukherjee ◽  
Dinkar Mylaraswamy
Keyword(s):  
Time Series ◽  
Fault Detection ◽  
Gas Turbine ◽  
Fault Detection And Diagnosis ◽  
Variable Length ◽  
Training Data ◽  
Series Data ◽  
Support Vector ◽  
Detection And Diagnosis ◽  
Time Sequences

Accurate gas turbine engine Fault Detection and Diagnosis (FDD) is essential to improving aircraft safety as well as in reducing airline costs associated with delays and cancellations. This paper compares broadly three methods of fault detection and diagnosis (FDD) dealing with variable length time sequences. Chosen methods are based on Dynamic Time Warping (DTW), k-Nearest Neighbor method, Hidden Markov Model (HMM) and a Support Vector Machine (SVM) which makes use of DTW ingeniously as its kernel. The time sequences are obtained from Turbo Propulsion Engines in their nominal conditions and two faulty conditions. Typically there is paucity of faulty exemplars and the challenge is to come up with algorithms which work reasonably well under such circumstances. Also, normalization of data plays a significant role in determining the performance of the classifiers used for FDD in terms of their detection rate and false positives. In particular spherical normalization has been explored considering the advantage of its superior normalization properties. Given sparse training data how well each of these algorithms performs is shown by means of tests performed on time series data collected at normal and faulty modes from a turbofan gas turbine propulsion engine and the results are presented.

Gas Turbine Fault Detection and Diagnosis Using Nonlinear Feature Extraction Methods

2005 ◽  
Author(s):  
Joydeb Mukherjee ◽  
Venkataramana B. Kini ◽  
Sunil Menon ◽  
Lalitha Eswara
Keyword(s):  
Feature Extraction ◽  
Fault Detection ◽  
Gas Turbine ◽  
Principal Component ◽  
Feature Space ◽  
Component Analysis ◽  
Extraction Methods ◽  
Fault Detection And Diagnosis ◽  
Sensor Data ◽  
Detection And Diagnosis

Accurate gas turbine fault detection and diagnosis (FDD) is essential to improving airline safety as well as in reducing airline costs associated with delays and cancellations. In this paper, we present FDD methods based on feature extraction methods using nonlinear principal component analysis (NLPCA) and curvilinear component analysis (CCA). The underlying principle of both methods is to find the most representative feature space corresponding to gas turbine normal and faulty operations. During operation, new sensor data is located in this feature space and then it is determined whether a particular fault is indicated. NLPCA is an extension of linear PCA methods to the nonlinear domain; therefore, it is intrinsically better suited to nonlinear domains such as the gas turbine engine. The CCA method is another approach to clustering having superior properties for determining cluster manifolds automatically compared to the popular selforganizing map (SOM) method of clustering. The developed methods are tested with snapshot data collected at takeoff, both normal and faulty, from a turbofan gas turbine propulsion engine and the results are presented.

Sign in / Sign up

Export Citation Format

Share Document