Aerospace and Electronic Systems Magazine July 2018 - 4

Feature Article:

DOI. No. 10.1109/MAES.2018.170045

EGT Prediction of a Micro Gas Turbine Using Statistical
and Artificial Intelligence Approach
Iman Koleini, University of Tehran, Iran
Alireza Roudbari, Shahid Sattari Aeronautical University of Science and Technology,
Tehran, Iran
Vahid Marefat, Urmia University of Technology, Urmia, West Azerbaijan, Iran

INTRODUCTION
Small size gas turbine application has increased with the advancement of technology. These engines have simple structures, fewer
number of operating components, and less design and construction
costs in comparison with large gas turbines. Small size gas turbines
can be used in different fields such as remotely piloted vehicles,
unmanned aerial vehicles, helicopters, and special purpose aircraft
such as vertical takeoff and landing aircrafts [1]. The advantages
of these engines compared to the large gas turbine engines leads
to high usage of small-sized engines and growing demand for research activities and experimental test results on the operating parameters of these engines, such as exhaust gas temperature (EGT).
In general, research on gas turbine engines, related with the present
work, can be divided into two parts. The first part includes studies
about past experiments and modeling of gas turbine engines, and
the second part focuses on data mining methods applied on gas
turbine engines. These two areas are discussed in the following
sections.

USING MODELING FOR STUDYING GAS TURBINE ENGINES
Modeling gas turbine engines especially small-sized ones was the
subject of studies [2]-[10]. Some of these studies were on numerical modeling of the AMT Olympus HP E-Start gas turbine engine
[2]-[4]. This is the engine used in the present work and it can be
used in applications such as radio controlled flying objects, processes for improving experimental flying objects, and audio research [11]. These engines are also widely used in educational and
Authors' current addresses: I. Koleini, Master Student, Faculty
of New Sciences and Technologies, University of Tehran, Tehran, Iran; A. Roudbari, Department of Aeronautical Engineering, Shahid Sattari Aeronautical University of Science and
Technology, Tehran, Iran, E-mail: (alirezaroudbari@ae.sharif.
ir); V. Marefat, Department of Mechanical Engineering Urmia
University of Technology, Urmia, West Azerbaijan, Iran.
Manuscript received February 3, 2017, revised June 8, 2017,
and ready for publication July 4, 2017.
Review handled by M. Jah.
0885/8985/18/$26.00 © 2018 IEEE
4

research projects at many universities and research centers around
the world [11].
A set of experiments was carried out by Leylek et al. [2] using a
small gas turbine engine. In their study, the performance modeling
and thermodynamic characteristics of the engine were evaluated.
They also did some analysis on strategies for increasing thrust.
Rahman and Whidborne [3] did a numerical investigation on the
engine bleed and its effect on performance of a single-spool turbojet engine. They did a real-time simulation, modeling to investigate
the effect of variable engine bleed on steady-state and transient
performance of a single-spool turbojet engine. The effect of thrust
vector on performance of the engine also was examined. They
showed that, for better consistency of modeling and experimental
results, the efficiencies of the components should be obtained with
more accuracy. They also found that increase in bleed extraction
error can be attributed to two factors. First, bleed extraction location affects the amount of work required by the compressor and
this may vary nonlinearly for the radial compressor of the AMT
Olympus engine. Second, the bleed extraction itself may adversely
influence the performance of the compressor, and the compressor
characteristics may change. Diamantis and Anastassios [4] performed some activities on data analysis and performance model
calibration of a small turbojet engine. They developed a simulation
model to evaluate the performance of a small turbojet engine. They
considered both random and systematic errors that occur in experimental examinations, and they also involved a correction factor
into the experimental data to validate the experimental results with
the simulation ones. They showed that a calibrated model using
total temperature and pressure measurements at the compressor
and turbine exits, is inadequate. Two possible reasons were examined for this inadequacy: Heat transfer through the turbine casing,
which was not taken into account, and the possibility of systematic
errors in the measurements used.
Liu [5] accomplished activities on the creation of a dynamic
model for small gas turbine engine. He developed a dynamical
model and provided a simulation platform for predicting micro
engine performance and control design. He also produced an estimation of the micro engine behavior under current design. He
analyzed and modeled the dynamical components of the micro
engine and then derived a nonlinear model, a linearized model,

IEEE A&E SYSTEMS MAGAZINE

JULY 2018

Table of Contents for the Digital Edition of Aerospace and Electronic Systems Magazine July 2018