Adaptive Modeling and Simulation of Gas Turbine Based on Improved Genetic Algorithms

In order to estimate the gas turbine engine performance precisely, the component maps containing their own performance characteristics should be used. Because the components map is an engine manufacturer’s propriety obtained from many experimental tests with high cost, they are not provided to the customer generally. Some scaling methods for gas turbine component maps using experimental data or data partially given by engine manufacturers had been proposed in a previous study. Among them the map generation method using experimental data and genetic algorithms had showed the possibility of composing the component maps from some random test data. However not only does this method need more experimental data to obtain more realistic component maps but it also requires some more calculation time to treat the additional random test data by the component map generation program. Moreover some unnecessary test data may introduced to generate inaccuracy in component maps. The map generation method called the system identification method using partially given data from the engine manufacturer (Kong and Ki, 2003, ASME J. Eng. Gas Turbines Power, 125, 958–979) can improve the traditional scaling methods by multiplying the scaling factors at design point to off-design point data of the original performance maps, but some reference map data at off-design points should be needed. In this study a component map generation method, which may identify the component map conversely from some calculation results of a performance deck provided by the engine manufacturer using the genetic algorithms, was newly proposed to overcome the previous difficulties. As a demonstration example for this study, the PW206C turbo shaft engine for the tilt rotor type smart unmanned aerial vehicle which has been developed by Korea Aerospace Research Institute was used. In order to verify the proposed method, steady-state performance analysis results using the newly generated component maps were compared with them performed by the Estimated Engine Performance Program deck provided by the engine manufacturer. The performance results using the identified maps were also compared with them using the traditional scaling method. In this investigation, it was found that the newly proposed map generation method would be more effective than the traditional scaling method and the methods explained above.

Download Full-text

Component Map Tuning Procedure Using Adaptive Modeling

Volume 1: Aircraft Engine; Ceramics; Coal, Biomass and Alternative Fuels; Controls, Diagnostics and Instrumentation ◽

10.1115/gt2012-69688 ◽

2012 ◽

Cited By ~ 1

Author(s):

Michel L. Verbist ◽

Wilfried P. J. Visser ◽

Rene Pecnik ◽

Jos P. van Buijtenen

Keyword(s):

Path Analysis ◽

Gas Turbine ◽

Engine Performance ◽

Diagnostic Techniques ◽

Reference Points ◽

Performance Data ◽

Adaptive Modeling ◽

Real Component ◽

Modeling Errors ◽

Component Map

Performance models are effective tools for analysis of engine condition throughout the life cycle of a gas turbine engine. Component maps necessary for accurate performance modeling are typically not provided by the original equipment manufacturers. To compensate for the missing information, available maps of similar components are scaled to match component performance at one or more reference points. Although scaled maps can provide sufficiently accurate results close to the reference points, modeling errors tend to increase further away from these reference points. For applications such as gas path analysis, the resulting modeling errors can be of the same order of magnitude as the deterioration to be detected. This severely limits the application of such techniques. This article presents a component map tuning procedure that tunes maps with more detail than just scaling. The tuned maps are a closer match to real component performance. The tuning procedure combines the adaptive modeling capability of the Gas turbine Simulation Program (GSP) and on-wing measured engine performance data. On-wing measured engine performance data allows map tuning over a wider range of power settings compared to engine performance data measured in a test cell. Effects of measurement uncertainty and scatter, and effects of compressor bleed flows on the map tuning procedure are analyzed and discussed. The tuned component maps enabled more accurate component condition estimations, mainly characterized by less scatter. By improving the accuracy of gas path analysis with on-wing measured performance data, this work has enabled more effective use of performance diagnostic techniques in the aero-engine maintenance industry.

Download Full-text

Multidisciplinary modeling and simulation of a fuel cell/gas turbine hybrid power system

2009 IEEE Power & Energy Society General Meeting ◽

10.1109/pes.2009.5275756 ◽

2009 ◽

Cited By ~ 11

Author(s):

Atideh Abbasi ◽

Zhenhua Jiang

Keyword(s):

Fuel Cell ◽

Power System ◽

Modeling And Simulation ◽

Gas Turbine ◽

Hybrid Power System ◽

Hybrid Power ◽

Multidisciplinary Modeling

Download Full-text

Experience With GSP as a Gas Path Analysis Tool

Volume 2: Aircraft Engine; Ceramics; Coal, Biomass and Alternative Fuels; Controls, Diagnostics and Instrumentation; Environmental and Regulatory Affairs ◽

10.1115/gt2006-90904 ◽

2006 ◽

Cited By ~ 2

Author(s):

W. P. J. Visser ◽

H. Pieters ◽

M. Oostveen ◽

E. van Dorp

Keyword(s):

Path Analysis ◽

Gas Turbine ◽

Reference Model ◽

Measurement Data ◽

Engine Performance ◽

Analysis Data ◽

Calibration Method ◽

Analysis Tool ◽

Operational Environment ◽

Adaptive Modeling

SKF’s primary tool for gas turbine engine performance analysis is GSP (Gas turbine Simulation Program), a component based modeling environment that is developed at National Aerospace Laboratory NLR and Delft University of Technology, The Netherlands. One of the applications is gas path analysis (GPA) using GSP’s generic adaptive modeling capability. With GSP, gas path analysis has been applied to different aero engines at several maintenance facilities. Additional functionalities have been developed to analyze multiple engine operating points and combine results of different adaptive modeling configurations automatically, resulting in more accurate and reliable GPA results. A ‘multi-point calibration’ method for the reference model was developed providing a significant improvement of GPA accuracy and stability. Also, a method was developed using ‘multiple analysis cycles’ on different condition indicator subsets, which successfully generated values for all condition parameters in cases with fewer measurement parameters than condition indicators and where measurement data are unreliable. The method has been successfully demonstrated on the GEM42 turbo shaft engine. A number of case studies have shown GPA results corresponding to available maintenance notes and inspection data. The extension of the GSP GPA tool with a database system provides a useful tool for analyzing engine history and comparison of analyzed component conditions throughout the fleet. When a large amount of analysis data is stored in the database, statistic analyses, trending and data mining can be performed. Also maintenance work scope effect on engine performance can be predicted. In this paper, the newly developed GSP gas path analysis functionalities are described and experiences and results with the GEM42 engine operational environment are presented.

Download Full-text