Identification of Longitudinal Aircraft Dynamics by Process Model Using Matlab/Simulink

International Journal of Scientific Engineering and Technology Volume No.2, Issue No.
4, pp : 207-210
(ISSN : 2277-1581) 1 April 2013
Identification of Longitudinal Aircraft Dynamics by Process Model using MATLAB/SIMULINK

Jitu Sanwale1, Dhan Jeet Singh2 Design Electrical Department, Aircraft Upgrade R&D Centre, Hindustan Aeronautics Limited, Nasik (MH) 1jituspec@iitr.ernet.in, 2dhanjeet1981@rediffmail.com
AbstractSystem-identification methods compose a mathematical model, or series of models, from measurements of inputs and outputs of dynamic systems. The estimated models allow the characterization of the response of the overall aircraft or component subsystem behavior. This paper discusses the use of System Identification Toolbox of MATLAB/SIMULINK for the estimation of aircraft flight dynamics in longitudinal channel. The data extraction, model selection and estimation of aircraft dynamics by process modeling along with data validation, time and frequency response analysis is illustrated using the system-identification simulation environment. KeywordsSystem Identification, Process model, MATLAB/SIMULINK. I. INTRODUCTION Design and analysis of control system is one of the most significant and imperative field of engineering and technology. Assorted methods [i] have been inhabited from the various fields of control engineering for enhanced performance. To design a better control system, the comprehension of system dynamics is indispensable. The system dynamics can be articulated in form of differential equations developed from physical principles or from transfer function models, which expresses the input-output property of the system. Some of the external parameters which affect system behaviour can have unknown or uncertain values, for example vortex formation on the control surface of the aircraft. In practical scenario, enhancement in the control system performance is constrained due to insufficient knowledge of the system behaviour or dynamics. Therefore estimation of the precise behavioural aspect of system is mandatory. facilitated by the rapid availability of accurate end-to-end and subsystem dynamic models. In this paper, section II illustrates the system identification and steps involved in estimation of dynamic behavior of the system. Section III deals with input-output data generation by experiment simulation, model selection and estimation by process modelling using MATLAB/SIMULINK and data validation. Results are discussed in Section IV and conclusion is provided in Section V. II. SYSTEM IDENTIFICTION System identification is the art and science of building mathematical models from measured input-output data. It allows estimating mathematical models of a dynamic system based on measured data. Essentially by adjusting parameters within a given model until its output coincides with the measured output. The most widespread models are difference equations descriptions, such as Process model, ARX and ARMAX models, as well as all types of linear state-space models. [ii] discusses the procedure of system identification using ARX model. The dynamics of the system can be estimated by using either parametric or nonparametric identification methods. Parametric Identification Methods are techniques to estimate parameters in given model structures. The numerical values of the parameters are find in an iterative manner that give the best agreement between the estimated output and the measured output. Nonparametric Identification Methods are techniques to estimate system behaviour without any model structure. Typical nonparametric methods include Correlation analysis, which estimates a systems impulse and step response. The spectral analysis is useful in frequency response estimation. The procedure to determine a model of a dynamical system from observed input-output data involves three basic ingredients:
System identification is a full life-cycle technology that The input-output data supports aircraft flight-control system development from design specification through flight-test optimization. Significant A set of candidate models (the model structure) reductions in development time and costs are realized by A criterion to select a particular model in the set, based on tracking open and closed-loop dynamic response characteristics the information in the data (the identification method). through the development process. System identification is especially effective in providing a transparent and integrated III. INDENTIFICATION PROCESS understanding of handling-qualities characteristics and system A. Input-Output data collection stability. Considerable improvements in system performance are
IJSET@2013
Page 207
International Journal of Scientific Engineering and Technology Volume No.2, Issue No. 4, pp : 207-210
The measurement data can be conceived by recording of input and output parameters from a live system. In this paper, we consider a conventional transport aircraft [iii] transfer function for collection of input-output data. The measurement experiment is carried out using MATLAB SIMULINK [iv] and block diagram representation is shown in Fig 1 with longitudinal aircraft transfer function given as below
(ISSN : 2277-1581) 1 April 2013
B. Model Selection and Estima tion This paper describes the system identification technique using process modelling. Process model allows to generate simple, continuous-time, dynamic linear models - characterized by static gain, time constants, and time delays. System Identification (SI) toolbox GUI of MATLAB have real and imaginary poles, zeros, delay and integrators as standard blocks. The SI GUI is opens by command ident on the command prompt. The time domain input and output data is imported from the MATLAB workspace by popup menu import data in SI GUI window. There is a provision in SI GUI window for examining and pre-processing the imported data set. However, for the sake of originality we have discarded the pre-processing stage.
Input_Delta_e
To Workspace3
Input Data Output Data1
s+3.1 s2 +2.8s+3.24
1 s Integrator
Output_Pitch_Angle
Random Number Generator
Transfer Fcn
To Workspace2
Fig.1. Experiment Simulink block diagram
The input and output data generated from simulation is shown in Fig 2 & 3respectively.
STABILIZER DEFLECTION CURVE
25 20 15
Stabilizer deflection [degree]
Fig. 4 System Identification Toolbox GUI window
10 5 0 -5 -10 -15 -20 -25 0
In this paper integrator, zero and undamped poles are considered for estimation. The selected transfer function coefficients are then estimated by using popup menu Estimate ->Process models. The estimated process model and its coefficients can be observed in window shown in Fig 5.
5 10 15 20
Time [sec]
25
30
35
40
45
50
Fig.2. Aircraft stabilizer deflection input curve

PITCH ANGLE CURVE
2
Pitch Angle [degree]
-1
-2
-3
-4
10
15
20
25
30
35
40
45
50
Time [sec]
Fig. 5 Estimated parameters in System Identification Toolbox GUI window Fig.3. Aircraft pitch angle output curve
IJSET@2013
C. Model Validation After the model estimation the verification need to be carried out for model perfectness. The data used to validate the accuracy of the model is shown by an icon, labelled Validation Data under Model Views. In this paper we have verified the estimated model by feeding the same aircraft stabilizer deflection input and obtained aircraft pitch angle output and discussed in section IV.
(ISSN : 2277-1581) 1 April 2013
Fig 8 & 9 reveal that the estimated models step and frequency response is identical with the actual model response.
STEP RESPONSE
6
Pitch Angle [deg]
IV. RESULTSANDDISCUSSION From Fig 6, it is clear that the waveform of measured data and estimated model output data are basically the same and the matching degree is about 95.54 percent. The pole-zero analysis of estimated model is also carried out in SI GUI and shown in Fig 7.
0 -2
-1
Time [sec]
.
MEASURED AND SIMULATED MODEL OUTPUT
2 Measured Simulated
10
2
Fig. 8 Step response of estimated system model
FREQUENCY RESPONSE
Amplitude
10
10
-2
Pitch Angle [deg]
10
-4
10
-6
10
-1
10
10
10
-1
-80 -100
Phase (deg)
-2
-120 -140 -160 -180 -200 -1 10
-3
10
10
10
-4
10
15
20
25
30
35
40
45
50
Frequency (rad/s)
Time [sec]
Fig. 9 Frequency response of estimated system model
Fig. 6 Comparison of measured and estimated system output

POLE-ZERO PLOT OF ESTIMATED MODEL
1.5
V. CONCLUSIONS Without prior knowledge of system behaviour, it is difficult to design good control system. For a flying platform, this can be problematic because poor control can cause the system to lose stability and crash, potentially damaging the aircraft. Even when stable gains have been implemented, it is still a time-consuming task to optimize the gains because it involves testing and retesting new gains on the aircraft. By characterizing the dynamics of the aircraft, the iterative guess-work in selecting control gains can be completely removed. In this paper, we have identified the aircraft dynamics with 95.54 percent accuracy using process modelling. The results indicate that process modelling is best suitable candidate identification of longitudinal aircraft dynamics involving integrator, zero and undamped poles.
0.5
Imaginary
-0.5
-1
-1.5 -5
-4
-3
-2
-1
Real
Fig. 7 Pole-zero plot of estimated system model
Apart from the above, time and frequency domain behaviour of the estimated model is also carried out. The results shown in
IJSET@2013
ACKNOWLEDGMENT The authors thank the management of Aircraft Upgrade Research & Design Centre, Hindustan Aeronautics Limited for providing the necessary facilities to undertake the above work. REFERENCES L Ljung, System Identification: Theory for the User, PTR Prentice Hall, Englewood Cliffs, New Jersy 07632, 1987. M. Li, C Chen, W Liu. Identification Based on MATLAB, Proceedings of the 2009 International Workshop on Information Security and Applicatio n, pp 523-525, November 21-22, 2009. J. H. Blakelock, Automatic Control of Aircraft and Missiles, Second Edition, A Wiley-interscience Publication, John Wiley & Sons Inc. 1991. L.Ljung, System Identification Toolbox Users Guide August, 1995. Biography
(ISSN : 2277-1581) 1 April 2013
i.
ii.
Jitu Sanwale received the B.E. degree in Electronics & Instrumentation Engineering from SGSITS, Indore in 2006. He also received the M.Tech degree in Communication Systems from IIT Roorkee in 2008. He is an associate member of Aeronautical Society of India (AeSI) and Institution of Electronics and Telecommunication Engineers (IETE). He joined the Aircraft Research & Design Centre, Hindustan Aeronautics Limited in 2008 as a design engineer. Presently, he is working as Deputy Manager (Design). His areas of interest include Inertial and Global Positioning System, Radio Navigation, Flight Control System and Adaptive Signal Processing. Dhan Jeet Singh received the B.Tech degree in Electronics & Communication Engineering from BIET, Jhansi in 2005. He is a life time member of Aeronautical Society of India.He joined the Aircraft Upgrade Research & Design Centre of Hindustan Aeronautics Limited in 2005 as a design engineer. Presently, he is working as Manager (Design) in Flight Controls & Navigation group of Design Electrical department. His areas of interest are radio navigation, flight control system and system identification.
iii.
iv.
IJSET@2013

Identification of Longitudinal Aircraft Dynamics by Process Model Using Matlab/Simulink

Uploaded by

Copyright:

Available Formats

Identification of Longitudinal Aircraft Dynamics by Process Model Using Matlab/Simulink

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Identification of Longitudinal Aircraft Dynamics by Process Model Using Matlab/Simulink

Uploaded by

Copyright:

Available Formats

International Journal of Scientific Engineering and Technology Volume No.2, Issue No.

(ISSN : 2277-1581) 1 April 2013