Simulation Modelling Practice and Theory 19 (2011) 996–1006

Contents lists available at ScienceDirect

Simulation Modelling Practice and Theory

journal homepage: www.elsevier.com/locate/simpat

Real-time multi-rate HIL simulation platform for evaluation of a jet

engine fuel controller
Morteza Montazeri-Gh, Mostafa Nasiri ⇑, Soheil Jafari
Systems Simulation and Control Laboratory, Department of Mechanical Engineering, Iran University of Science and Technology (IUST), Narmak, Tehran, Iran

a r t i c l e i n f o a b s t r a c t

Article history: A new Hardware-In-the-Loop (HIL) platform is developed for testing of a turbojet engine
Received 2 September 2010 fuel control system using a multi-rate simulation platform. The HIL equipment consists
Received in revised form 12 December 2010 of an industrial PC and a commercial I/O board for jet engine simulation as the controlled
Accepted 28 December 2010
process and an Electronic Control Unit (ECU) as the fuel controller. The controlled process
Available online 4 January 2011
consisting of actuator, physical process and sensors is fully simulated in HIL simulation.
However, the high resolution signals of some components in the HIL simulation cause
Keywords:
the real-time simulation to become difficult due to the need of small time-steps. As a
Hardware-in-the-loop
Multi-rate
result, the disparity between the jet engine model sampling rate and these high resolution
Turbojet engine signals requires a multi-rate simulation. In this study, a multi step size simulation is devel-
Fuel controller oped using multiple processors. These processors are designed to synchronize the status of
Electronic control unit the engine model with the control system as well as to convert the raw data of the I/O
boards to actual input and output signals in real-time. These features make the HIL equip-
ment more effective and flexible. The HIL environment is proved to be an efficient tool to
develop various control functions and to validate the software and hardware of the engine
fuel control system.
Ó 2010 Elsevier B.V. All rights reserved.

1. Introduction

Generally, the construction of aircraft is costly and time consuming. Safety is also a primary issue that one is facing in
conducting actual flight tests. Consequently, Hardware-In-the-Loop (HIL) simulation can effectively evaluate the reliability
of the overall aircraft system. HIL simulation is characterized by the operation of real components in connection with real-
time simulated components. The simulated components are often the processes being controlled and/or sensors and actu-
ators. In particular, the framework can be used to examine the performance of aircraft subsystems and equipments such
as flight control system [1], camera hardware [2], fuel cell power [3], electro hydraulic actuators [4] and engine fuel control
system [5].
Several studies have been reported for HIL simulation of jet engine control system. Iserman et al. [6] provided an overview
of the various kinds of simulation considering of real-time simulators, architectures and the historical development of
hardware-in-the-loop simulation. Cheng [7] discussed HIL testing system for the mini-type turbojet engine to study the tran-
sient and steady state performance of a conventional digital control regulator for ground start, air start, stopping, automatic
acceleration and deceleration, and steady state regulation. The adjusting characteristics of air start and flying along the
ballistic trajectory are also evaluated. In [8], HIL simulation system using MATLAB/xPC-target was developed for an
electronic throttle idle speed control strategy based on ANFIS. Wang et al. [9] introduced and applied a method of non-fully

⇑ Corresponding author. Tel.: +98 913 3001343; fax: +98 21 77240488.

E-mail addresses: montazeri@iust.ac.ir (M. Montazeri-Gh), m_nasiri@iust.ac.ir (M. Nasiri), s_jafari@iust.ac.ir (S. Jafari).

1569-190X/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.simpat.2010.12.011
 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006 997

recovering LQG/LTR to design the aero-engine control system. In order to validate the performance of the designed control sys-
tem, the HIL simulation system was then designed. Watanabe et al. [10] designed a fuzzy logic controller and tested on a tur-
bojet engine in a simulated environment. The controller was tested in HIL simulation before being used with the real engine.
According to Fathy [11], a common problem in HIL simulation is virtual model stiffness, defined as a large disparity be-
tween the characteristic speeds of different components of a virtual model. When the disparity between the fast and slow
dynamics in a virtual model cannot be eliminated, it is common to simulate these dynamics separately at different sampling
rates. Such multi-rate simulation may take place on one processor via multithreading, but is more often achieved using mul-
tiple processors. The HIL literature describes the challenge to the simulation platform due to the high time resolution re-
quired for emulating high sampling rates signals of actuators and sensors [12,13]. The disparity between the jet engine
control system components have not been considered for HIL simulation development.
In this paper, a new HIL simulation platform is developed for the evaluation of jet engine ECU performance. In this appli-
cation, electrical signals including rotor speed encoder and servo valve drive signals are at high sampling rates rather than
the engine model sampling rate of simulation. So this system is characterized by a combination of subsystems working at
different time scales and with different needs of time resolution. In order to address this issue, the separation of the high
resolution signals for the real-time operation from the software requirements has been proposed. This separation can be
achieved by introducing a suitable hardware interface to catch and to generate these high resolution switching signals with
a high accuracy without the need of reducing the integration time steps of jet engine simulation.
The outline of this paper is organized as follows. In Section 2, a thermodynamic model and the integration method which
is proposed for the real-time simulation of single spool turbojet engine is presented. The structure of the min–max algorithm
employed in the ECU is illustrated in Section 3. In Section 4, the software and hardware framework for HIL simulation is pre-
sented. The interface between the simulation model and hardware consists of some convertors described in Section 5. Finally
the results of the simulation with the simulated ECU and actual ECU are compared in Section 6. Some concluding remarks are
presented in Section 7.

2. Turbojet engine real-time model

The process considered is a single spool turbojet engine with a convergent nozzle, without bypass and bleed flow. The
assumptions in the model presented here are that, there does not exist any heat transfer between the control volume and
in each control volume, the gas is perfectly mixed. These assumptions imply that only two gas states per control volume
are necessary to determine the condition there. Jet engine model has been composed of two parts. The first one refers to
a system of static nonlinear equations that describe the thermodynamic relations in compressor, turbine and nozzle as fol-
lows [14,15].

Compressor:
  sffiffiffiffiffi
d2 P3 N P2 T2
_ c ¼ pffiffiffiffiffi
m f1 ; pffiffiffiffiffi ; d2 ¼ ; h2 ¼ ð1Þ
h2 P 2 h2 P0 T0
 
P3 N
gc ¼ f2 ; pffiffiffiffiffi ð2Þ
P 2 h2
( " c1 #)
1
1P3 c
T 3s ¼ T 2 1þ ð3Þ
gc P 2
Turbine:
  sffiffiffiffiffi
d4 P4 N P4 T4
_ t ¼ pffiffiffiffiffi
m f3 ; pffiffiffiffiffi ; d4 ¼ ; h4 ¼ ð4Þ
h4 P 5 h4 P0 T0
 
P4 N
gt ¼ f4 ; pffiffiffiffiffi ð5Þ
P 5 h4
( "  c1 #)
 c 1
P4
T 5s ¼ T 4 1 þ gt ð6Þ
P5

Nozzle:
8 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
> qffiffiffiffiffiffi 1c  c1
>
< C d An pPffiffiffiffi 2c P e Pe c
5
RT c 1 P
1  P5
pn  pn
_n¼
m 5
qffiffiffiffiffiffi ð7Þ
>
> 2c
: 0:2588C d An pPffiffiffiffi 5
pn > pn
c1
RT
998 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006

Fig. 1. Turbojet engine.

The second part refers to dynamic nonlinear equations that describe transient process. It has been assumed that the pro-
cesses of accumulating the enthalpy and the mass of working medium take place only in combustion chamber and nozzle.
Similar progress in the inlet duct, the compressor and the turbine has been neglected. Disregard of the turbine duct volume
has resulted from the fact that it is a single stage turbine. In addition, the compressor duct has been modeled as a control
volume together with the combustion chamber. The transient equations due to combustion chamber volume are as follows
[15,16]:
dT b 1
¼ _ 3 ðcT 3  T b Þ þ m
½m _ 4 ðT b  cT 4 Þ þ m
_ f gb Hf  ð8Þ
dt mb
dPb RT P dT b
¼ ðm _ 4Þ b þ b
_ 3m ð9Þ
dt V b T b dt
And the following relations show transient equations due to nozzle volume.
dT n 1
¼ _ 5 ðcT 5  T n Þ þ m
½m _ e ðT n  cT e Þ ð10Þ
dt mn
dPn RT P dT n
¼ ðm _ eÞ n þ n
_ 5m ð11Þ
dt V n T n dt

Table 1
Symbol of parameters.

c Ratio of specific heats

d Normal pressure
h Normal temperature
g Efficiency
p Pressure ratio
m_ Mass flow rate
T Temperature
P Pressure
N Rotor speed
A Area
V Volume
R Gas constant
Cd Nozzle discharge coefficient
H Heating value
cph, cpc Specific heat of hot/cold air at constant pressure
ch, cc Ratio of specific heats for hot/cold air
f1, f2 Compressor maps
f3, f4 Turbine maps
Subscripts
c Compressor
t Turbine
n Nozzle
b Combustion chamber
f Fuel
s Isentropic process
 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006 999

Fig. 2. Real-time simulation flowchart.

Moreover, the Euler equation due to the rotational motion of the rotor is given by:
_ t cph
m _ c cpc
m
N_ ¼ ðT 4  T 5 Þ  ðT 3  T 2 Þ ð12Þ
JN JN
1000 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006

0.8

Fuel flow rate (Normalized)

0.7

0.6

0.5

0.4

0.3

0.2
0 50 100 150 200
time (sec)
Fig. 3. Fuel flow rate.
Rotor speed N (Normalized)

0.9

0.8

0.7

0.6 Experimental
Simulation
0.5
0 50 100 150 200
time (sec)
Fig. 4. Rotor speed.

0.8
CDP (Normalized)

0.6

Experimental
0.4
Simulation
0 50 100 150 200
time (sec)
Fig. 5. Compressor discharge pressure.

The parameters used in Eqs. (1)–(12) are defined in Fig. 1 and Table 1.
Physical modeling of turbojet engine results in systems of Differential and Algebraic Equations (DAEs). Eqs. (1)–(12) con-
stitute the behavior of the whole single spool turbojet engine. This model includes seven algebraic Eqs. (1)–(7) and five or-
dinary differential equations (ODEs) of (8)–(12). The ODEs are solved using Runge–Kutta method. The steady-state values of
variables at the starting time of simulation are given by solving a set of nonlinear algebraic equations using Newton–
Raphson method. The flowchart of this real-time simulation is illustrated in Fig. 2.
 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006 1001

In addition, the engine is equipped with some pressure transducer, thermocouples, flowmeter and a magnetic toothed
wheel to measure the operating parameters of the engine in the test. Moreover, a PC/104 embedded computer is employed
to receive the signals from the measurement devices at a desired rate. Furthermore, the fuel flow is changed by a PLA (power
lever angle) manually where the servo angle is varied according to the PLA command. Figs. 4 and 5 show comparison of error
between the engine thermodynamic model and test responses due to fuel flow variations shown in Fig. 3. These results show
a good agreement between the model and test responses that support the model used in this study.

3. Electronic control unit hardware

What the pilot usually wants to achieve is to get the engine to deliver a certain percentage of the thrust that is available at
the current flight conditions. Since thrust itself is not measurable in flight, the relative thrust command given by the PLA
(power lever angle) setting must be translated into a command change of a measured variable. The relative thrust corre-
sponds very well to the engine pressure ratio and this parameter can be used for thrust modulation in controller design [17].
In this study, a min–max fuel controller is designed based on this fact that the fuel flow is considered to have two parts
including steady and transient fuel flow. Steady fuel flow refers to equilibrium points. The steady-state controller is respon-
sible to meet the first requirement of the engine control system. In addition to the steady-state considerations, a transient
performance for a gas turbine plant must be considered. Transient flow is the variation of fuel flow with respect to its
steady-state value in transient mode. The steady-state controller is designed based on a gain scheduling manner from a
look-up-table between the corrected rotor speed and corrected fuel flow. The transient controls necessary to accelerate
and decelerate the engine while meeting stall, flameout, over speed and temperature limitations will then be added.
In this control algorithm, four control loops are defined for calculation of transient fuel flow, three of them are limiting
control loops and another one is thrust modulation for satisfying pilot demands [18]. The control loops are as follows.

1. Maximum rotor speed limitation loop: This control loop computes the fuel flow that would be necessary to drive the
engine exactly to the given spool speed limit.
2. Maximum acceleration limiting loop: To ensure constant acceleration times, the maximum acceleration of the spool
has to be limited.
3. Maximum deceleration limiting loop: To limit the deceleration of the spool and thus ensure enough compressor
surge margin.
4. PLA control loop: The Compressor Pressure Ratio (CPR) is selected as a reference to compare with the PLA. The cor-
responding control loop to be implemented has to comprise an integrator to drive the difference between com-
manded and actual CPR to zero.

The first three limiting loops organize a bounded safe zone for transient fuel selection. Whenever the transient fuel is se-
lected from this area, the safe physical operation of the engine will be attained.
As mentioned above, four proportional control loops are used for calculating the transient fuel flow. The proportional
gains of these control loops are named Knmax, Kdec, Kacc and Kpla respectively. Moreover, the fuel flow corresponding to each
control loops are named m _ nmax ; m
_ dec ; m
_ acc and m
_ pla .

Fig. 6. Strategy of min–max fuel controller.

1002 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006

The engine transient fuel flow is calculated using a min–max selection algorithm. This algorithm decides which of these
fuel flows is passed onto the fuel actuation system. The selection logic uses the following algorithm:

m _ dec ; m
_ transient ¼ minðminðmaxðm _ pla Þ; m
_ nmax Þ; m
_ acc Þ ð13Þ
The overall block diagram of this control system algorithm is depicted in Fig. 6.

4. HIL architecture

The Matlab/Simulink has been selected as the base software used for the Electronic Control Unit (ECU) development. The
toolbox of Real-Time Workshop (RTW) can be used to generate the C-code directly from Simulink model. The xPC-target is
then selected as a HIL simulation platform which can be employed to make a real-time system with the host PC and target
PC. It provides a high-performance host-target prototyping environment to connect with the physical systems, and then exe-
cute them in real-time on PC-compatible hardware. The xPC target also provides some Input/Output (I/O) interface blocks for
system engineer to add them into the Simulink model on the host PC, and then to automatically generate code with Real-
Time Workshop which can be downloaded to the target PC for running real-time applications. The host PC is operated under
Microsoft Windows operation system like a general PC. However, the target PC runs the simulation programs with the xPC
operation system for speeding up the computation, and hence to realize the real-time simulation. In addition, a TCP/IP net-
work connection is used as the connection between the host PC and target PC. The HIL simulator consists of a host computer,
target computer, Ethernet network card, Network cable and I/O devices. The block diagram of this HIL system is shown in
Fig. 7.
A personal computer with specifications shown in Table 2 is used as the target computer. A compact disc carries xPC tar-
get real-time kernel is used for starting the target computer. After starting the target computer, the application files of jet
engine model is downloaded from the host computer to the target computer via TCP/IP protocol. An ethernet local area

Host PC

TCP/IP

Target PC (Jet Engine Model) RPM (0-5 v )

Servo (8bit) PCL-812PG (AO)
PCL-812PG (DI)
CDP (0-5 v)
PCL-812PG (AO)

Voltage to Frequency
PWM to digital
Converter

Converter

Monitoring PC

RS-232

v
CDP (0-5 )

Servo (5.5-9.5% PWM) ECU

RPM (0-4.5KHz)
(PC/104) PLA (0-5 )
v
Receiver

RS-232

Pilot Stick

Fig. 7. Hardware/software interfaces of HIL simulation.

 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006 1003

Table 2
Specification of target computer.

CPU Intel core 2 due, 2.4 GHz

RAM 2 GB
COM ports 2
Slots 4 ISA, 6 PCI
LAN 1 Onboard
HDD SATA

Fig. 8. Simulation platform (1: host PC, 2: target IPC, 3: ECU, 4: interfaces, 5: monitoring PC).

network (LAN) is also used for connecting the host and target computers. The NE2000 network card is used for this purpose.
In order to connect the target computer and ECU hardware, the Advantech PCL-812PG I/O card is used as an interface be-
tween the software and ECU. Fig. 8 shows the real simulation platform in systems simulation and control laboratory.

5. Simulation to hardware interface

5.1. Transmitter

The aircraft is controlled using three Pulse Width Modulation (PWM) servo motors. The transmitter section uses a specific
protocol to send commands to the vehicle over an RS-232 connection. When manual control is enabled, the transmitter
sends a constant stream of ASCII packet to the vehicle flight control computer. These packets consist of five pieces of infor-
mation. These pieces of information include manual control status, engine fuel throttle, aileron angle, elevator angle and a
check byte. As shown in Fig. 7, the pilot can interact with the aircraft by stick-and-throttle interface.

5.2. ECU architecture

Flight control computer is responsible for flying the aircraft. The control computer is based on PC/104 which is a high-
performance rugged embedded computer with data acquisition and it is housed inside a protective box. The servo motor
of fuel control unit is controlled by means of managing PWM. The fuel control system is mainly designed to calculate the
required duty cycle to control the servo motor. The square waves of PWM signal is outputted 50 times per second. The width
of the square wave decides the horn of the servo oscillating angle, and the wave width is described according to the contin-
uous time. The width of square will change from 1.1 to 1.9 ms, and the horn of servo will rotate 0–170° angle. As described in
Section 3, the controller uses the engine rotor speed and Compressor Discharge Pressure (CDP) as feedback signals. The rotor
speed feedback signal to the ECU is a pulse train supplied by a magnetic speed sensor mounted in proximity to a toothed
wheel on the gas turbine compressor shaft.
1004 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006

5.3. Target PC architecture

An industrial PC with specifications shown in Table 2 is used as the target computer for jet engine simulation. This com-
puter is equipped by the ISA board of Advantech PCL-812PG. This card has two analog output channels. These channels are
used for exporting compressor pressure ratio and rotor speed as outputs of jet engine model. Moreover, the card has 16 dig-
ital input channels that 8 ports are used for reading servo variations.

5.4. Signal converters

The overall system interfaces and converters are shown in Fig. 7. For jet engine ECU testing, the output signal of rotor
speed sensor is at high frequency (0–4.5 KHz) and the input signal of servomotor PWM duty cycle is very low (5.5–9.5%).
As a result, emulation of the rotor speed sensor and servomotor PWM signals require too small time-steps. However, the
use of converters to attain acceptable precision or even stability and convergence removes the constraint of using very
small time-steps, which are impractical for real-time applications. In this study, the proposed converters are imple-
mented using AVR microcontrollers in order to provide an inexpensive means of accurately simulating the interfaces
signals.
In PWM to digital converter, an 8-bit digital output part of the AVR microcontroller is connected to the 8 bits of the digital
input port of Advantech PCL-812 PG data acquisition (DAQ) card. A PWM signal is also applied to one of the digital input
channel of the AVR controller in order to output the average duty value of the PWM signal proportional to the percentage
of the time the signal is on during the time step. Moreover, in voltage to frequency converter, an analog output channel of the
DAQ card is applied to analog input port of AVR controller and the pulse signal generated with microcontroller is connected
to the ECU as a feedback signal. In engine model, the rotor speed variable is transformed to a value between 0–5 and out-
ported using the card.

0.9

0.8
Normalized

0.7

0.6

0.5 PLA
CPR HIL
0.4 CPR simulation
0.3
0 20 40 60 80 100
time(sec)

Fig. 9. PLA input and CPR responses.

1
HIL
Servo displacement (Normalized)

0.9 Simulation
0.8

0.7

0.6

0.5

0.4

0 20 40 60 80 100
time(sec)

Fig. 10. Servomotor responses.

 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006 1005

0.9

RPM (Normalized)
0.8

0.7

0.6
HIL
0.5 Simulation

0.4
0 20 40 60 80 100
time(sec)

Fig. 11. Rotor speed.

6. Experimental results

The performance of the ECU for the fuel control of the turbojet engine is tested through the HIL system in order to eval-
uate and improve the controller performance by tuning the gain parameters. The engine is simulated at zero mach number
and only the fuel flow rate is varied during the simulation. In order to compare the response of the simulated and the actual
ECU responses, the parameters of the controller are the same as those used in the simulation and the same type of step in-
puts test is attempted. To test the controller at the variety of operating points, various step inputs of PLA are applied using
pilot stick/throttle interface. Fig. 9 shows the time history of the PLA input. As shown in Fig. 9, the whole PLA range from idle
PLA to nearly full is firstly passed in two PLA steps. The engine is decelerated again using one step input. The PLA command is
transformed into CPR demand and the controlled engine follows this demand using min–max controller. Fig. 10 shows the
performance of the controller in following the demand. The engine’s net thrust corresponds with the actual CPR value. The
agreement between the PLA and CPR validates that the min–max controller is designed successfully. Similar results are
achieved using the same PLA input applying to the simulation with simulated ECU. Comparison of the results for ECU
Model-In-the-Loop (MIL) and actual ECU in the loop has been shown in Figs. 9–11. After going through the results, it be-
comes clear that the HIL simulation results matches with those of MIL simulation. The difference between the two results
is mainly due to hardware realization and implementation. As shown in Fig. 9, the difference occurred in servo displacement
is because of the low resolution of the PWM to digital converter. Although a clear lag between the MIL and HIL simulation is
unavoidable because of the interfaces.

7. Conclusion

The purpose of this research is to develop a new HIL platform to simulate the dynamics of a turbojet engine for the pur-
pose of ECU rapid control prototyping. The simulator includes an engine model and interfaces for connecting to ECU hard-
ware. The host-target architecture is integrated in the system to provide the capability of rapid controller prototyping for
ECU system. To validate the HIL simulator, the real ECU testing dynamics are compared with the simulator dynamics and
the differences are characterized. This application includes PWM and encoder signals. This poses a major challenge to the
simulation platform due to the high time resolution required. So, a hardware interface is employed in the simulation plat-
form to separate the hardware requirements from the real-time simulation, since high time resolution is not necessary to
simulate with the jet engine real-time simulation in the target PC. An AVR-based signal conditioning hardware interface
has been adopted as the interface to the hardware under test. It greatly reduces the computational burden of fast signals.
The time resolution of the proposed interface hardware can also be improved by using faster and more accurate micropro-
cessors to improve the configuration of the interface.

References

[1] N.R. Gans, W.E. Dixon, R. Lind, A. Kurdila, A hardware in the loop simulation platform for vision-based control of unmanned air vehicles, Mechatronics
19 (7) (2009) 1043–1056.
[2] T.H. Bradley, B.A. Moffitt, D.N. Mavris, T.F. Fuller, D.E. Parekh, powerplant Hardware-in-the-loop testing of a fuel cell aircraft, Journal of Propulsion and
Power 25 (6) (2009) 1336–1344.
[3] M. Karpenko, N. Sepehri, Hardware-in-the-loop simulator for research on fault tolerant control of electrohydraulic actuators in a flight control
application, Mechatronics 19 (7) (2009) 1067–1077.
[4] S. Chumalee, J.F. Whidborne, Unmanned aerial vehicle aerodynamic model identification from a racetrack manoeuvre, Proceedings of the Institution of
Mechanical Engineers Part G: Journal of Aerospace Engineering 224 (7) (2010) 831–842.
[5] L. Jun, G. Ying-Qing, W. Hai-Quan, Rapid prototyping real-time simulation platform for digital electronic engine control, in: 2nd International
Symposium on Systems and Control in Aerospace and Astronautics, ISSCAA, 2008, Art. No. 4776230, 2008.
1006 M. Montazeri-Gh et al. / Simulation Modelling Practice and Theory 19 (2011) 996–1006

[6] R. Isermann, J. Schaffnit, S. Sinsel, Hardware-in-the-loop simulation for the design and testing of engine-control systems, Control Engineering Practice 7
(1999) 643–653.
[7] T. Cheng, Hardware in the loop simulation of mini type turbojet engine digital control regulator, Journal of Aerospace Power 19 (3) (2004) 383–386.
[8] Y. Cao, W. Teng, H. Zhang, Hardware-in-the-loop simulation for engine idle speed control based on ANFIS, in: Proceedings of the World Congress on
Intelligent Control and Automation (WCICA), Art. No. 4593421, 2008, pp. 3125–3137.
[9] H. Wang, Y. Guo, J. Lu, Design and validation of aeroengine control system with non-fully recovering LQG/LTR method, in: Second International
Conference on Innovative Computing, Information and Control, ICICIC0 2007, Art. No. 4428111, 2008.
[10] A. Watanabe, S.M. Ölçmen, R.P. Leland, K.W. Whitaker, L.C. Trevino, C. Nott, Soft computing applications on a SR-30 turbojet engine, Fuzzy Sets and
Systems 157 (22) (2006) 3007–3024.
[11] H.K. Fathy, Review of hardware-in-the-loop simulation and its prospect in the automotive area, Proceeding of SPIE, vol. 6228, Society of Photo-Optical
Instrumentation Engineers, 2006.
[12] B. Lu, X. Wu, H. Figueroa, A. Monti, A low-cost real-time hardware-in-the-loop testing approach of power electronics controls, IEEE Transactions on
Industrial Electronics 54 (2) (2007) 919–993.
[13] H.P. Figueroa, A. Monti, X. Wu, An interface for switching signals and a new real-time testing platform for accurate hardware-in-the-loop simulation,
in: IEEE International Symposium on Industrial Electronics 2, Art. No. 1571930, 2004, pp. 883–887.
[14] I.H. Ismail, F.S. Bhinder, Simulation of aircraft gas turbine engines, ASME Journal of Engineering for Gas Turbines and Power (1991).
[15] M. Henriksson, C. Breitholtz, in: Proceedings of the International Conference on Control Applications, IEEE, 2004.
[16] S.-T. Lin, J.-H. Lu, M.-C. Hong, K.-W. Hsu, Real-time dynamic simulation of single spool turbojet engines, Simulation 73 (6) (1999) 352–361.
[17] A. Kreiner, The use of onboard real-time models for jet engine control, MTU Aero Engines, Germany, 2002.
[18] M. Montazeri-Gh, A. Safari, S. Jafari, Optimization of turbojet engine fuel control system for safety consideration, in: The 7th Iranian Aerospace Society
Conference, Tehran, Iran, 2008.