Available online at www.sciencedirect.

com

Procedia Engineering 12 (2011) 4753

2011 SREE Conference on Engineering Modelling and Simulation (CEMS 2011)

A General Multitask Simulation Package of Aircraft Engine

Wang Lifeng*
North China University of Technology, Field Bus Technology & Automation Lab, 1000441

Abstract

The general multitask simulation of aircraft engine has been developed in order to meet the needs coming from
aircraft, engine and academic for research & develop platform. In the package, not only can military and commercial
aircraft engine be visually modeled, but also different simulating tasks can be completed, such as flight performance,
linear state space model, transient & fault simulating. The objective of this paper is to provide the comprehensive
overview of general multitask simulation package of aircraft engine.

2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.

Selection and/or peer-review under responsibility of Society for Resources, Environment and Engineering

Keywords: aircraft engine simulation, modeling package, linear state-space model, engine control

1. Introduction

There is great interest in the development of integrated aircraft/engine design, engine advanced
control algorithm and Engine Health Management[1]. So the simulation model of aircraft engine is needed
in order to complete different design tasks. As different calculating requirement which comes from
aircraft, engine and academic departments, it is essential to build the general multitask simulation
package to solve the problem.
In recent years, the commercial engine simulation package based on Matlab/Simulink is developed,
which is tools for advanced control and Health Management[1,2]. Other engine simulation planning on
concept of intelligent engine control is supported by NASA and GE [3,4]. In the meantime, a graphical
simulation program which is developed using java language is released for distributed engine design
calculating [5,6].The GSP is developed by using Delphi in Holland, which aims at modeling many types Jet

* Corresponding author. Tel.: +8610-88803751; fax: +8610-88803751.

E-mail address: lf_wang@ncut.edu.cn.

18777058 2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
doi:10.1016/j.proeng.2011.05.009
48 Wang Lifeng / Procedia Engineering 12 (2011) 4753

engine and its control system[7]. In 90s years, many modeling methods [8] and simulations package[9,10,11,12]
are developed in China, such as using orientedobject coding technique, the visual modeling and
simulation environment is built using C++ language.
The general multitask simulation of aircraft engine is developed in C++ language visual tools. The
objective of this paper is to provide the comprehensive overview of general multitask simulation package
of aircraft engine. First, the general package means the visual modeling and simulation environment of
aircraft engine and many types of military and commercial aircraft engine can be modeled for different
aircraft design tasks. Second, different simulating tasks can be completed and the data results can be
output visually in the form of data table or curve graph. The multitasks are included flight performance,
linear state space model, transient & fault simulating etc. Now the details are separately described as
follows.

2. Visual Modeling Environment

The engine consisted of many components which are connected by flow paths and rotors, so every
component of engine may be considered as an object entity. The oriented-object method is adopted to
design engine model which is divided into three type objects, such as component, path and shaft. The
component objects ancestor relativity is referring Fig 1.
The oriented-object modeling and coding technique is more suitable to design visual simulation tools.
If you could drag component icons to modeling zone in order of flow path station, the engine modeling
process is finished by simple click actions in simulation environment, in which the typical commercial
turbofan engine model is referring Fig 2. But in the engine model, the first constraint is that the TCell
object is existing as well as TPath or TShaft object although it is automatically created in simulation
environment. The second constraint is that the model is starting with the inlet object and ending with
TNozz object along with flow path. For example, the typical commercial turbofan engine model is
modeled in Fig 2, certainly the military turbofan, turbojet or other types engine model can be easily built
in simulation environment.
In the graphical user interface, most of initial data of engine model can be filled in form, and
components characteristics are supply by data file. Especially, the sensor measure output and control
input can be freely setup in order to invalid the control configurations. In the meantime, sensor noise level
and data I/O interface can also be setup.

TObject

TCell

TInlet TFan TBst TCom

TShaft

TPath

TTurb TMburn TDuct TMixer TAburn TNozz

Fig 1Cell class inherited relationship in engine model

 Wang Lifeng / Procedia Engineering 12 (2011) 4753 49

Fig 2: Commercial turbofan engine model skeleton

3. Engine Performance in Flight Envelop

The engine model setup in modeling environments can be described in form of nonlinear state
space model, which is consisted of two parts, such as differential equations and algebraic equation,
observation equation and performance equations. As follows:
x f (x, , u, h)
(1)
0 g (x, , u, h)
y y (x, , u, h)
(2)
p p(x, , u, h)
The former is represented for dynamic characteristics and mass flow balance of engine, the latter is
represented for the measured sensor output and the unmeasured performance output. In the equations, the
x vector is state variable, vector is the tuning variable, u vector is the control input variable, h is
health parameter, y vector is measurable output, p vector is unmeasurable performance output ,such as
engine force, special fuel consume-sfc etc.
Engine is serviced for aircraft, especially in the integrated aircraft/engine design, the much
calculating iteration and optimization between aircraft and engine designer may happen. So the basic
requirement for engine model is the ability of calculating the performance in flight envelop. When the
engine control principle is inserted into the equation (1), the augmented equation which described as
engine steady state model is as follows:
0 f ( x, , u , h )
0 g ( x, , u , h ) (3)
0 x ns  x n
50 Wang Lifeng / Procedia Engineering 12 (2011) 4753

Where xns and xn is expected engine state and actual engine state, for example x stands for fan
rotor speed, compressor rotor speed or pressure and temperature limits. After the control principle, the
flight altitude and flight Mach number is setup, the performance of engine force and specific fuel
consumption-sfc is completed and output in graphical form. The typical flight envelop referred to Fig 3,
and The performance of engine is as Fig 4. 



Fig 3: Aircraft/engine flight envelop


Fig 4: Engine performance of entire flight envelop

4. Linear State Space Model

The linear state space model is basic for research on modern advanced control algorithms and
engine health management, for example Performance Seeking Control-PSC, engine fault-tolerant control,
Integrated Flight/Propulsion Control-IFPC, health parameter estimate and fault prognosis etc. The
nonlinear equations (1) is solved in Newton Raphson method, as the state variable derivative x  is
assumed to be given constant, then the [x, E ] is the result. So the so-called Offset derivative method
T

is adopt to numerically calculating liner state space model.

 Wang Lifeng / Procedia Engineering 12 (2011) 4753 51

 to the [x, E ] , the nonlinear equations[1] is linear into as

According to the process from the x
T

follows:
G x A1 B1 L1
GE | G x  Gu  Gh
A2 B2 L2
G y | C G x  D G u  M Gh (4)
G p | E G x  F G u  N G h
The Jacobi matrix A1> A2 @
T
is a narrow matrix and assumed that inverse matrix A11 of matrix
A1 is existed, and the part equation on GE is omitted because it is no relative to Gx equation. So the
above equations are changed as follows:
G x | A1 1G x  A1 1 B 1G u  A1 1 L1G h
G y | C A  G x  (  C A11 B 1  D )G u  (  C A11 L1  M )G h (5)
G p | E A  G x  (  E A11 B 1  F )G u  (  E A11 L1  N )G h
Now the above equation (5) can be changed to linear state space model which is treated as state
variable Gx and health parameter Gh , as follows:
G x | A G x  B G u  L G h
Gy | C Gx  D Gu  M Gh (6)
Gp | E Gx  F Gu  N Gh
The above equation is the classic engine linear state space model which is much suitable to estimating
the performance and health parameter of engine.

5. Close-loop Transient and Fault Simulating

The engine control system modeling and simulating can be graphically realized in Matlab /Simulink.
So it is crucial how to connect the above engine simulation environments with the Matlab /Simulink in
order to implement the close-loop simulation. The connection way is referred to Fig 5.
The data transfer is realized by using the UDP manner and the simulating time is synchronous by
separately real-time clock control. In the way, the intelligent control and advanced optimization algorithm
can be invalided in Matlab/Simulink which is linked to simulation environment. In the meantime, the
control system model, for example certainly some sensor and actuator model may be considered as ideal,
should be inserted into the control system model if it is necessary.
On the basis of commercial turbofan engine model, the dynamic performance of transient process or
fault events can be simulated according to equations [1].The simulation state of affairs is as follows: the
engine state is migrating from idle to maximum within 3.5 seconds, then at 6 seconds the fault of high
pressure turbine happens, where the efficiency is reduced by 2% and the flow ability is increased by 1%.
The measured data output is showed in Fig 6.
52 Wang Lifeng / Procedia Engineering 12 (2011) 4753

Simulation environment
Engine Model
Udp Receiving Data Udp Sending Data
Real-time clock

Udp Sending Data Udp Receiving Data

Real-time clock

Matlab/Simulink
Control System Modeling

Fig 5 Data Transfer between Matlab/simulink & engine simulation environment

Fig 6: The typical results of engine transient and fault simulating,

The transient and fault simulating results of commercial turbofan engine are showed in Fig 6. The
eight measure outputs are varying vs time, including rotor speed, pressure, temperature, such as
N0/N0design, N1/N1design, P2/Pam, T2/Tam, T22/Tam, P3/Pam, T45/Tam, P5/Pam. The Pam, Tam in
Fig stands for standard atmosphere pressure, temperature and the measuring noise is added in the curve
graph. In the accelerating processit takes about 4s and most of output are increasing except the inlet
pressure P2/Pam decreasing because of the inlet pressure loss increasing vs flow mass. Then at 6s, the
fault of high turbine happens which results in the P3/Pam is reduced due to flow ability enhanced and
 Wang Lifeng / Procedia Engineering 12 (2011) 4753 53

N1/N1design is reduced due to turbine work decreased. In the means time, the control system in
Matlab/Simulink maintains the rotor speed N0/N0design resuming the expected constant after a while.

6. Conclusion

The general multitask simulation package of aircraft engine is introduced in paper. Some details of the
visual modeling and simulation environment, a commercial aircraft engine model and the different
simulating tasks are described. The model equations are proved through the simulating results which is
output in curve graph. The package is more suitable for research on Integrate aircraft/engine design and
advanced control algorithm. Not only the money and time are saved, but also preciseness and
effectiveness are gained.

Acknowledgements

Thanks for the project which is sponsored by Beijing Municipal Education Commission funds
(KM200410009006,KM200610009001) & PHR(IHLB).

References
[1] Jonathan A. DeCastro; Jonathan S. Litt; Dean K. Frederick;A Modular Aero-Propulsion System Simulationof a Large
Commercial Aircraft Engine, NASA/TM2008-215303, September 2008.
[2]. Frederick, D.K., DeCastro, J.A. and Litt, J.S., Users Guide for the Commercial Modular Aero-Propulsion System
Simulation (C-MAPSS), NASA/TM2007-215026, October 2007.
[3]Nathan GibsonIntelligent Engine SystemsAdaptive Control NASA/CR2008-215240June 2008.
[4]. Parker, K.I. and Guo, T.H., Development of a Turbofan Engine Simulation in a Graphical Simulation Environment,
NASA/TM2003-212543, August 2003.
[5]. Reed J.A,Afjeh A.A.Computational simulation of gas turbines:Part I-foundations of component-based
models[R],International Gas Turbine & Aeroengine Congress & Exhibition,1999.
[6].Reed J.A., Afjeh A.A. Computational simulation of gas turbines: part II-extensible domain framework[R]. International
gas Turbine & aeroengine Congress & Exhibition,1999.
[7].Visser W .P.J,Broomhead M.J. GSP: A generic object-oriented gas turbine simulation environment[R]. ASME Turbine
Expo 2000,Munich, 2000.
[8].Tang Hailong, Zhang Jing. A study of object-oriented approach for aeroengine performance simulation[J ]. Journal of
Aerospace Power, 1999, 14 (4) :421-424.
[9]. Zhou Wen-xiang, Huang Jinquan, Dou Jianping, etc; Objected-oriented simulation platform for turbofan engine and its
control system, Journal of Aerospace Power , 2007, 22(1): 119-125.
[10]Cao Zhisong , Yu Longjiang , Piao Ying An extensible propulsion simulation platform Part I 2 foundations of
engineering model Journal of Aerospace Power, Vol. 22 No. 2,Feb. 2007.
[11].Cao Zhisong, Yu Longjiang , Piao Ying; An extensible propulsion simulation platform: Part 2 extensible framework
Journal of Aerospace Power, Vol. 22 No. 2, Feb. 2007.
[12].Wang lifeng; Study on Visual Dynamic Modeling Environment of Turbojet Engine, 12th conference on aerodynamic &
thermodynamic of propulsion system of China, Sichuan, 2009/9.