scholarly journals Aero Engine Component Fault Diagnosis Using Multi-Hidden-Layer Extreme Learning Machine with Optimized Structure

2016 ◽  
Vol 2016 ◽  
pp. 1-11 ◽  
Author(s):  
Shan Pang ◽  
Xinyi Yang ◽  
Xiaofeng Zhang
Keyword(s):  
Fault Diagnosis ◽  
Gas Turbine ◽  
Extreme Learning Machine ◽  
Network Structure ◽  
Gas Turbine Engine ◽  
Data Set ◽  
Turbine Engine ◽  
Engine Component ◽  
Learning Machine ◽  
Hidden Layer

A new aero gas turbine engine gas path component fault diagnosis method based on multi-hidden-layer extreme learning machine with optimized structure (OM-ELM) was proposed. OM-ELM employs quantum-behaved particle swarm optimization to automatically obtain the optimal network structure according to both the root mean square error on training data set and the norm of output weights. The proposed method is applied to handwritten recognition data set and a gas turbine engine diagnostic application and is compared with basic ELM, multi-hidden-layer ELM, and two state-of-the-art deep learning algorithms: deep belief network and the stacked denoising autoencoder. Results show that, with optimized network structure, OM-ELM obtains better test accuracy in both applications and is more robust to sensor noise. Meanwhile it controls the model complexity and needs far less hidden nodes than multi-hidden-layer ELM, thus saving computer memory and making it more efficient to implement. All these advantages make our method an effective and reliable tool for engine component fault diagnosis tool.

scholarly journals Aero Engine Fault Diagnosis Using an Optimized Extreme Learning Machine

2016 ◽  
Vol 2016 ◽  
pp. 1-10 ◽  
Author(s):  
Xinyi Yang ◽  
Shan Pang ◽  
Wei Shen ◽  
Xuesen Lin ◽  
Keyi Jiang ◽  
...  
Keyword(s):  
Gas Turbine ◽  
Extreme Learning Machine ◽  
Validation Data ◽  
Data Set ◽  
Swarm Optimization ◽  
Turbine Engine ◽  
Optimal Network ◽  
Aero Engine ◽  
Learning Machine ◽  
Hidden Layer

A new extreme learning machine optimized by quantum-behaved particle swarm optimization (QPSO) is developed in this paper. It uses QPSO to select optimal network parameters including the number of hidden layer neurons according to both the root mean square error on validation data set and the norm of output weights. The proposed Q-ELM was applied to real-world classification applications and a gas turbine fan engine diagnostic problem and was compared with two other optimized ELM methods and original ELM, SVM, and BP method. Results show that the proposed Q-ELM is a more reliable and suitable method than conventional neural network and other ELM methods for the defect diagnosis of the gas turbine engine.

An improved hybrid modeling method based on extreme learning machine for gas turbine engine

2020 ◽  
Vol 107 ◽  
pp. 106333
Author(s):  
Maojun Xu ◽  
Jian Wang ◽  
Jinxin Liu ◽  
Ming Li ◽  
Jia Geng ◽  
...  
Keyword(s):  
Gas Turbine ◽  
Extreme Learning Machine ◽  
Gas Turbine Engine ◽  
Modeling Method ◽  
Hybrid Modeling ◽  
Turbine Engine ◽  
Learning Machine

scholarly journals Comprehensive report on Materials for Gas Turbine Engine Components

2019 ◽  
Vol 9 (2S2) ◽  
pp. 155-158
Keyword(s):  
Gas Turbine ◽  
Turbine Blade ◽  
Thermal Stresses ◽  
Prolonged Exposure ◽  
Raw Materials ◽  
Turbine Blades ◽  
Gas Turbine Engine ◽  
Operating Conditions ◽  
Turbine Engine ◽  
Engine Component

In the past three decades, it is very challenging for the researchers to design and development a best gas turbine engine component. Engine component has to face different operating conditions at different working environments. Nickel based superalloys are the best material to design turbine components. Inconel 718, Inconel 617, Hastelloy, Monel and Udimet are the common material used for turbine components. Directional solidification is one of the conventional casting routes followed to develop turbine blades. It is also reported that the raw materials are heat treated / age hardened to enrich the desired properties of the material implementation. Accordingly they are highly susceptible to mechanical and thermal stresses while operating. The hot section of the turbine components will experience repeated thermal stress. The halides in the combination of sulfur, chlorides and vanadate are deposited as molten salt on the surface of the turbine blade. On prolonged exposure the surface of the turbine blade starts to peel as an oxide scale. Microscopic images are the supportive results to compare the surface morphology after complete oxidation / corrosion studies. The spectroscopic results are useful to identify the elemental analysis over oxides formed. The predominant oxides observed are NiO, Cr2O3, Fe2O3 and NiCr2O4. These oxides are vulnerable on prolonged exposure and according to PB ratio the passivation are very less. In recent research, the invention on nickel based superalloys turbine blades produced through other advanced manufacturing process is also compared. A summary was made through comparing the conventional material and advanced materials performance of turbine blade material for high temperature performance.

Design of Dynamic Tracking Filter for Model Based Diagnosis of Gas Turbine Engine

2016 ◽  
Author(s):  
Yongwen Liu ◽  
Yunsheng Liu
Keyword(s):  
Fault Diagnosis ◽  
Gas Turbine ◽  
Gas Turbine Engine ◽  
Influence Coefficient ◽  
Performance Parameters ◽  
Relative Gain ◽  
Turbine Engine ◽  
Dynamic Tracking ◽  
Tracking Filter ◽  
Turbine Model

There exist many approaches for gas turbine engine condition monitoring and fault diagnosis. Among them, gas path analysis depends on the relations between deviations of performance parameters and deviations of measurements, such as pressure, temperature, at some positions in the flow path. In the first author’s previous study, a dynamic tracking filter is combined with a nonlinear gas turbine model to form a fault diagnosis system. The dynamic tracking filter is composed with multiple negative feedback control loops in which the residuals between model outputs and measurements are driven to zero by adjusting the performance parameters. In the present study, an interaction analysis technique, named the Relative Gain Analysis (RGA), is introduced to design more convincing and formal tracking filter for a heavy-duty gas turbine diagnostic problem. The basic concept of the RGA method is introduced in this paper with a simple blending example. Then a gas turbine model built using the Simscape language and environment from the MathWorks Co. is presented. The effects of secondary air system on the performance of compressor and turbine are considered in this gas turbine model. The linear influence coefficient matrix for four performance parameters and four measurement parameters is obtained from the steady state simulation with proper disturbance of performance parameters. Then the relative gain matrix (RGM) is obtained from matrix operation. To evaluate the pairing rule proposed in the RGA method, four tracking loops are built up to form a tracking filter for the four performance parameters selected. Deviations of performance parameters are implanted into the gas turbine model by adjusting the scaling factors of performance maps; and then simulation results are taken as measurements needed for the tracking filter to run. Tracking results of performance parameters in different cases are given to show the tracking capability for isolated performance deviations and concurrent performance deviations.

OMP-ELM: Orthogonal Matching Pursuit-Based Extreme Learning Machine for Regression

2015 ◽  
Vol 24 (1) ◽  
pp. 135-143 ◽  
Author(s):  
Omer F. Alcin ◽  
Abdulkadir Sengur ◽  
Jiang Qian ◽  
Melih C. Ince
Keyword(s):  
Extreme Learning Machine ◽  
Matching Pursuit ◽  
Computational Cost ◽  
Machine Intelligence ◽  
Orthogonal Matching Pursuit ◽  
Training Data ◽  
Singularity Problem ◽  
Data Set ◽  
Learning Machine ◽  
Hidden Layer

AbstractExtreme learning machine (ELM) is a recent scheme for single hidden layer feed forward networks (SLFNs). It has attracted much interest in the machine intelligence and pattern recognition fields with numerous real-world applications. The ELM structure has several advantages, such as its adaptability to various problems with a rapid learning rate and low computational cost. However, it has shortcomings in the following aspects. First, it suffers from the irrelevant variables in the input data set. Second, choosing the optimal number of neurons in the hidden layer is not well defined. In case the hidden nodes are greater than the training data, the ELM may encounter the singularity problem, and its solution may become unstable. To overcome these limitations, several methods have been proposed within the regularization framework. In this article, we considered a greedy method for sparse approximation of the output weight vector of the ELM network. More specifically, the orthogonal matching pursuit (OMP) algorithm is embedded to the ELM. This new technique is named OMP-ELM. OMP-ELM has several advantages over regularized ELM methods, such as lower complexity and immunity to the singularity problem. Experimental works on nine commonly used regression problems indicate that the investigated OMP-ELM method confirms these advantages. Moreover, OMP-ELM is compared with the ELM method, the regularized ELM scheme, and artificial neural networks.

Research on Transformer Fault Diagnosis Based on Online Sequential Extreme Learning Machine

2019 ◽  
Vol 12 (5) ◽  
pp. 408-413
Author(s):  
Yuancheng Li ◽  
Xiaohan Wang ◽  
Yingying Zhang
Keyword(s):  
Fault Diagnosis ◽  
Extreme Learning Machine ◽  
Experimental Result ◽  
Dissolved Gas ◽  
Traditional Methods ◽  
Electric System ◽  
Diagnosis Method ◽  
Learning Machine ◽  
Hidden Layer ◽  
Transformer Fault

Background: Transformer is one of the most important pivot equipment in an electric system which undertakes major responsibility. Therefore, it is very important to identify the fault of the transformer accurately and transformer fault diagnosis technology becomes one topic with great research value. Methods: In this paper, after analyzing the shortcomings of traditional methods, we have proposed a transformer fault diagnosis method based on Online Sequential Extreme Learning Machine (OS-ELM) and dissolved gas-in-oil analysis. This method has better precision than some commonly used methods at present. Furthermore, OS-ELM is more efficient than ELM. In addition, we analyze the effect of different parameter selection on the performance of the model by contrast experiments. Results: The experimental result shows that OS-ELM has certain promotion in precision than some traditional methods and can obviously improve the speed of training than ELM. Besides, it is known that the number of neurons in the hidden layer and the size of dataset have a great effect on the model. Conclusion: The transformer fault diagnosis method based on OS-ELM can effectively identify the faults and more efficient than ELM.

Sign in / Sign up

Export Citation Format

Share Document