Neural Networks Modelling of Aero-derivative Gas Turbine Engine: A Comparison Study

Ibrahem, Ibrahem M. A.; Akhrif, Ouassima; Moustapha, Hany; Staniszewski, Martin

doi:10.5220/0007928907380745

Cited by 7 publications

(5 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A final comparison [24] investigated two multiple-input multiple-output (MIMO) neural networks for modelling a gas turbine engine. The study found that the NARX model was better at generalising than the standard NN, however it did take longer to train.…”

Section: Previous Comparisons In Literaturementioning

confidence: 99%

Effect of Time History on Normal Behaviour Modelling Using SCADA Data to Predict Wind Turbine Failures

et al. 2020

View full text Add to dashboard Cite

Operations and Maintenance (O&M) can make up a significant proportion of lifetime costs associated with any wind farm, with up to 30% reported for some offshore developments. It is increasingly important for wind farm owners and operators to optimise their assets in order to reduce the levelised cost of energy (LCoE). Reducing downtime through condition-based maintenance is a promising strategy of realising these goals. This is made possible through increased monitoring and gathering of operational data. SCADA data are useful in terms of wind turbine condition monitoring. This paper aims to perform a comprehensive comparison between two types of normal behaviour modelling: full signal reconstruction (FSRC) and autoregressive models with exogenous inputs (ARX). At the same time, the effects of the training time period on model performance are explored by considering models trained with both 12 and 6 months of data. Finally, the effects of time resolution are analysed for each algorithm by considering models trained and tested with both 10 and 60 min averaged data. Two different cases of wind turbine faults are examined. In both cases, the NARX model trained with 12 months of 10 min average Supervisory Control And Data Acquisition (SCADA) data had the best training performance.

show abstract

Section: Previous Comparisons In Literaturementioning

confidence: 99%

Effect of Time History on Normal Behaviour Modelling Using SCADA Data to Predict Wind Turbine Failures

et al. 2020

View full text Add to dashboard Cite

show abstract

“…In this paper, to get the optimal NARX model structure which can represent the ADGTE dynamics, we perform an extensive comparative performance study using different combinations of NARX neural network architectures, training algorithms and activation functions while using different numbers of neurons. As a result, a comprehensive computer program was developed in the MATLAB environment (Ibrahem et al 2019). This program generates 240 NARX models with different structures by performing the following:…”

Section: Engine Real Time Modelmentioning

confidence: 99%

“…ADGTE are widely used as a mechanical drive in oil and gas application and power generation. These widespread and increasing applications have sparked a great interest among manu-facturers to improve the performance and increase the reliability of the engine, which in turn requires an accurate and real time model to simulate the gas turbine engine dynamics (Ibrahem et al 2019). GTE modelling methods can be categorized into two main groups including physics based modelling methods (white-box models) and data driven based modelling methods (black-box models).…”

Section: Introductionmentioning

confidence: 99%

“…Based on the literature survey, in the area of MIMO ANN modelling of gas turbine engines, the research activities used mostly one of the following two methods to generate a nonlinear model for the MIMO engine: Either, by building a neural network model for each output parameter (MISO) with the same structure for each one of them and trained with the same training algorithm (Bahlawan et al 2017, Asgari et al 2016, or by building one block neural model to represent the MIMO system (Asgari et al 2014, Tarik et al 2017, Salehi & Montazeri-Gh 2018, Ibrahem et al 2019. However, it is more powerful, as will be shown in this study, to use a different neural network's structure for each output.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Real time modeling and Hardware in the loop simulation of an aero-derivative gas turbine engine

Ibrahem¹,

Akhrif²,

Moustapha³

et al. 2020

Proceedings of Global Power &Amp; Propulsion Society

Self Cite

View full text Add to dashboard Cite

Modelling and simulation of gas turbines plays a key role in manufacturing and improving performance of gas turbine engines. In recent years, the importance of ADGTEs in the energy industry has sparked a great interest among manufacturers to improve the performance and increase the reliability of the engine, which in turn requires an accurate and real time model to simulate the engine dynamics during the full operating range. For this purpose, a real-time modelling and hardware in the loop (HIL) simulation of an aero-derivative gas turbine engine (ADGTE) are developed in this paper. A non-linear autoregressive network with exogenous inputs (NARX) is used to develop this model in MATLAB environment using operational closed-loop data collected from Siemens (SGT-A65) three-spool dry low emission ADGTE. A multiple-input single-output (MISO) NARX models with different configurations are used to represent each of the ADGTE output parameters with the same input parameters. First, data preprocessing and estimation of the order of these MISO models are performed. Next, a comprehensive computer program code is developed to perform a comparative study and to select the best NARX model configuration, which can represent the system dynamics. Finally, to demonstrate the performance of the MISO NARX model developed in this study, a HIL simulation is performed. In this simulation, the physical ADGTE is replaced with the MISO NARX model, and the physical engine controller is used to control that model. In addition, the neural model is validated with experimental data. The simulation results show that the proposed MISO NARX models followed the targets precisely and can predict the response with high accuracy and reliability. In addition, the proposed model presents high computation speed, which allows for real time applications.

show abstract

“…ML-based techniques like ANN have shown the capability to predict dynamic behavior of GTs without having access to information about the system physics. Different ANN-based methodologies have already been investigated and developed in order to disclose complex nonlinear behavior of aero gas turbines (Agrawal and Yunis, 1982;Chiras et al, 2001Ruano et al, 2003;Torella et al, 2003;Sarkar et al, 2012Sarkar et al, , 2013Salehi and Montazeri, 2018;Ibrahem et al 2019). These efforts have covered a variety of approaches such as…”

Section: Introductionmentioning

confidence: 99%

Recurrent Neural Network Based Simulation of a Single Shaft

Asgari¹,

Ory²,

Lappalainen³

2021

Linköping Electronic Conference Proceedings

View full text Add to dashboard Cite

In this study, a model of a single shaft gas turbine (GT) is developed by using artificial intelligence (AI). A recurrent neural network (RNN) is employed to train the datasets of the GT variables in Python programming environment by using Pyrenn Toolbox. The resulting model is validated against the Test datasets. Thirteen significant variables of the gas turbine are considered for the modelling process. The results show that the RNN model developed in this study is capable of performance prediction of the system with a high reliability and accuracy. This methodology provides a simple and effective approach in dynamic simulation of gas turbines, especially when real datasets are only available over a limited operational range and using simulated datasets for modelling and simulation purposes is unavoidable.

show abstract

Neural Networks Modelling of Aero-derivative Gas Turbine Engine: A Comparison Study

Cited by 7 publications

References 4 publications

Effect of Time History on Normal Behaviour Modelling Using SCADA Data to Predict Wind Turbine Failures

Effect of Time History on Normal Behaviour Modelling Using SCADA Data to Predict Wind Turbine Failures

Real time modeling and Hardware in the loop simulation of an aero-derivative gas turbine engine

Recurrent Neural Network Based Simulation of a Single Shaft

Contact Info

Product

Resources

About