0
Research Papers: Gas Turbines: Aircraft Engine

A New Approach to the Gray-Box Identification of Wiener Models With the Application of Gas Turbine Engine Modeling

[+] Author and Article Information
Ehsan Mohammadi

Systems Simulation and Control Laboratory,
School of Mechanical Engineering,
Iran University of Science and Technology (IUST),
Tehran 16846-13114, Iran
e-mail: ehs_mohammadi@iust.ac.ir

Morteza Montazeri-Gh

Professor
Systems Simulation and Control Laboratory,
School of Mechanical Engineering,
Iran University of Science and Technology (IUST),
Tehran 16846-13114, Iran
e-mail: montazeri@iust.ac.ir

1Corresponding author.

Contributed by the Aircraft Engine Committee of ASME for publication in the JOURNAL OF ENGINEERING FOR GAS TURBINES AND POWER. Manuscript received August 31, 2014; final manuscript received October 31, 2014; published online December 23, 2014. Editor: David Wisler.

J. Eng. Gas Turbines Power 137(7), 071202 (Jul 01, 2015) (12 pages) Paper No: GTP-14-1521; doi: 10.1115/1.4029170 History: Received August 31, 2014; Revised October 31, 2014; Online December 23, 2014

In this paper, a new approach is presented for the gray-box identification of Wiener models (WM); and to evaluate the performance of the proposed method, it is used to estimate the dynamic behavior of a two-shaft industrial gas turbine (GT). The Wiener models, which have attracted a considerable attention due to their low computational demand and high accuracy, represent modeling techniques based on system identification. These models are composed of a linear dynamic part interconnected with a nonlinear static element, and the unknown parameters of these two parts are generally determined by black-box identification approaches. However, another identification method known as “gray-box identification” can also be employed, which uses the existing information about the static or dynamic behavior of a system to achieve the unknown parameters of the Wiener model. In this study, an innovative approach for improving the Wiener model’s capability of predicting the dynamic behavior of nonlinear systems is presented with the assumption that the static behavior of the examined system is known. In the proposed model called the enhanced Wiener model (EWM), the parameters of the linear dynamic part are allowed to vary with the operating conditions; and thus, this model provides a higher flexibility in estimating the dynamic behavior of the examined system compared to the conventional Wiener models. The EWM consists of a static nonlinear block and a linear dynamic block with varying parameters. Since gas turbine engines are essentially nonlinear in both the steady and transient conditions, the modeling of a gas turbine can be a suitable case for evaluating the effectiveness of the proposed model. In this regard, in order to estimate the parameters of a two-shaft industrial gas turbine, five multi-input single-output (MISO) EWMs with a special structure are employed in which the parameters of the dynamic part of each EWM is determined by an adaptive network-based fuzzy inference system (ANFIS). In order to evaluate the performance of the proposed model, the EWM results are compared with the result obtained by common system identification approaches like Wiener, Hammerstein, Wiener–Hammerstein, nonlinear autoregressive exogenous (NARX), and ANFIS models. The simulation results reveal that the proposed EWM not only is more flexible and effective in predicting the dynamic behavior of the examined gas turbine than the block-structured models, but it also outperforms the NARX and ANFIS models in estimating the static behavior of the gas turbine.

Copyright © 2015 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

A typical ANFIS architecture with multi-inputs and one output

Grahic Jump Location
Fig. 2

Block diagrams of the block-structured models: (a) Hammerstein model, (b) Wiener model, and (c) Wiener–Hammerstein model

Grahic Jump Location
Fig. 3

The schematic of the EWM training

Grahic Jump Location
Fig. 4

The overall configuration of the enhanced Wiener model for the two-shaft gas turbine modeling

Grahic Jump Location
Fig. 5

Schematic representation of the two-shaft gas turbine

Grahic Jump Location
Fig. 6

Schematic of the gas turbine control circuit

Grahic Jump Location
Fig. 7

Load variations in train mode (QAPRBS signal)

Grahic Jump Location
Fig. 8

NGG variations in train mode

Grahic Jump Location
Fig. 9

Fuel flow variations in train mode

Grahic Jump Location
Fig. 10

EGT variations in train mode

Grahic Jump Location
Fig. 11

Output power variations in train mode

Grahic Jump Location
Fig. 12

CPR variations in train mode

Grahic Jump Location
Fig. 13

CAMF variations in train mode

Grahic Jump Location
Fig. 14

NPT variations in train mode

Grahic Jump Location
Fig. 15

Variation of the “τ” in the EWM versus the WM for estimation of the NGG (training data)

Grahic Jump Location
Fig. 16

Variation of the “τ” in the EWM versus the WM for estimation of the EGT (training data)

Grahic Jump Location
Fig. 17

Variation of the “τ” in the EWM versus the WM for estimation of the CAMF (training data)

Grahic Jump Location
Fig. 18

Drop in the PC index when the examined models subjected to the untrained data

Grahic Jump Location
Fig. 19

Load variations in test mode (QAPRBS signal)

Grahic Jump Location
Fig. 20

Fuel flow variations in test mode

Grahic Jump Location
Fig. 21

NPT variations in test mode

Grahic Jump Location
Fig. 22

NGG variations in test mode (the selected region)

Grahic Jump Location
Fig. 23

EGT variations in test mode (the selected region)

Grahic Jump Location
Fig. 24

The percentage of error in estimation of the static behavior of the (a) CAMF, (b) NGG, (c) CPR, (d) EGT, and (e) OP

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In