International Journal of Electrical and Computer Engineering (IJECE) Vol.
No.
August 2017, pp.
ISSN: 2088-8708.
DOI: 10.
11591/ijece.
A New Induction Motor Adaptive Robust Vector Control based on Backstepping Mhamed Madark.
Ba-Razzouk.
Abdelmounim.
El Malah ASTI Laboratory.
FSTS.
University Hassan 1 st .
Settat.
Morocco.
Article Info
ABSTRACT
Article history:
Received Jun 9, 2016 Revised Nov 20, 2016 Accepted Dec 11, 2016 Keyword:
Adaptive control Backstepping control Indirect field oriented control Induction machine In this paper, a novel approach to nonlinear control of induction machine, recursive on-line estimation of rotor time constant and load torque are developed.
The proposed strategy combines Integrated Backstepping and Indirect Field Oriented Controls.
The proposed approach is used to design controllers for the rotor flux and speed, estimate the values of rotor time constant and load torque and track their changes on-line.
An open loop estimator is used to estimate the rotor flux.
Simulation results are presented which demonstrate the effectiveness of the control technique and on-line estimation.
Copyright A 2017 Institute of Advanced Engineering and Science.
All rights reserved.
Corresponding Author:
Mhamed Madark.
Department of Physics.
Science and Technical Faculty.
Hassan 1st University.
Settat.
Morocco.
Email: m.
madark1987@gmail.
INTRODUCTION
Induction machine constitutes the workhorse of the industry and is used in a lot of industrial Induction machine is robust, necessitates very little maintenance and is generally less expensive than synchronous or permanent magnet motors.
Induction machine is a class of highly coupled nonlinear multivariable system with two crucial control inputs .
tator voltage.
and two output variables .
echanical speed, rotor flu.
required to track given reference signals.
Parameters of induction machine are time varying with temperature, magnetic saturation and frequency.
Therefore, machine control performance deteriorates as the parameters vary and the controller loses its tuning.
It is therefore desirable to track on-line changes in the system parameters and to continually update the controller parameters to ensure optimal tuning all time.
The speed and rotor flux control is difficult because of the complexity of the model of the induction machine.
Many control strategies to solve the above problem have been proposed.
One of these techniques is the vector control.
This technique allows an induction machine to achieve torque control performance similar to that of a dc machine in many high performance drives.
Vector control can be realized in a direct or indirect fashion.
The principal drawback of indirect vector control is its sensitivity to parameters variations especially the rotor time constant .
The latter varies with time, thus using a robust controller with fixed parameters will not guarantee the performance required in all possible operating conditions.
Many algorithms have been proposed to estimate the rotor time constant .
Practically, load torque is measured using either contact or non contact type torque sensor where the latter is not economic.
Load torque estimation of induction machine is essential in trajectory tracking control Journal homepage: http://iaesjournal.
com/online/index.
php/IJECE A ISSN: 2088-8708 of rotor speed, so the observed load torque is fed forward to increase the robustness of the induction motor speed drive.
Methods for estimating load torque are interesting many researches and many solutions were proposed .
Vector control enables an asymptotic decoupling between torque and rotor flux of the machine as opposed to the nonlinear control that offers global decoupling .
The implementation of a new nonlinear control law with parametric adaptation by combining the Backstepping technique and indirect vector control is necessary to improve tracking of trajectories, guarantee stability, robustness to changes in parameters and rejection of disturbances.
The nonlinear controller and the adaptive law are designed for the mechanical speed and rotor flux using a fourth order model of induction machine.
This paper is organized as follows: Section 2 describes modeling of the system.
Section 3 introduces the theoretical aspects of vector control, control law and the adaptation laws.
Section 4 presents simulation results and section 5 introduces the perspectives of practical implementation.
MODELING OF THE SYSTEM
Modeling of Induction Machine The dynamic equations of induction machine in the synchronously rotating reference frame are expressed as:
Where , , , , are dq axis stator current, dq axis stator voltage, dq axis rotor flux, dq axis rotor current, rotor electric angular pulsation, stator electric angular pulsation, dq axis stator flux, stator resistance and rotor resistance respectively.
The stator and the rotor flux are defined as .
Where , and M are stator inductance, rotor inductance and mutual inductance respectively.
The load torque acts as a disturbance via the mechanical rotation .
Where , , are load torque, moment of inertia, viscous friction coefficient and number of pole pairs respectively.
The model of an induction machine can be expressed in a state space representation.
By considering the stator currents, the rotor flux and the rotor electric angular pulsation as state variables and two control Modeling of the Inverter The various switches are assumed perfect.
We can replace the arm of the inverter with one or two switches positions modeled by a logic function.
The latter is defined as follows:
and the switching function of the switch Let , with i * associated with the arm i of this inverter.
The relationships between these different functions are expressed by:
IJECE Vol.
No.
August 2017 : 1983 Ae 1993
IJECE
ISSN: 2088-8708
The potentials of the nodes A.
B and C of the inverter relative to the point M are given by:
We can also express phase voltages from line voltages as follows:
Using .
, the expression of phase voltages are formulated by .
We choose Let[ ] [ ].
( )[
] .
( )
( )
With ( ) ( )[ ] is d-q transformation.
The inverter model in the d-q reference frame is given by .
INVERTER
K11
K21
K31
K12
K22
K32
Figure 1.
Structure of a three phase inverter Modeling of the System The model of an induction machine can be formulated in a state space representation.
considering the stator currents, the rotor flux and the rotor electric angular pulsation as state variables and the switching functions ( ) yielding the following fifth Aeorder model.
( )
With:
UU is the state vector, ] is the control vector A New Induction Motor Adaptive Robust Vector Control Based on Backstepping (Mhamed Madar.
A ISSN: 2088-8708
( )
) .
) AE(
) .
AE .
AE .
From the state space representation, we can see that the dynamic model of induction machine and the inverter is strongly coupled nonlinear multivariable system.
The aim objectives is to choose such a way as to force the speed and the rotor flux to track given reference values denoted by respectively, to design an adaptation laws to on-line estimate the value of rotor time constant, estimate the value of load torque and track their variations.
Indirect Field Oriented Control This control strategy is based on the orientation of the flux vector along d-axis, which can be expressed by considering .
Using .
, we eliminate all the terms with quadrature component of rotor flux and reduce the five equation in .
to this expression of the synchronous angular speed.
If the above condition is satisfied, the decoupling of the rotor voltage equations is realized.
To what extent this decoupling is actually achieved will depend on the accuracy of motor parameters used.
On-line parameter adaptive technique is employed to tune the value of these parameters used in the controller which combines Backstepping and the indirect field oriented (IFO) control techniques.
To design the nonlinear control and the adaptation law based on the technique of Backstepping, we use a simplified model of induction machine in order to develop a simple algorithm that can be implemented The simplified model of induction machine is obtained by setting ( ) With:
And ( ) to zero in .
UU is the state vector, ( ) ( ) ( ) [ ( )] is the control vector Which is a fourth order model.
ADAPTIVE BACKSTEPPING CONTROL
Backstepping is a recursive Lyapunov-based scheme for the class of strict feedback systems.
The main idea of Backstepping is to design a recursive control by choosing some variables a virtual control and designing for them intermediate control laws.
This approach guarantees global or regional regulation and tracking proprieties.
An important advantage of the design of Backstepping design stabilizing controllers is IJECE Vol.
No.
August 2017 : 1983 Ae 1993
IJECE
ISSN: 2088-8708
following a step-by-step algorithm.
This technique has also the flexibility to avoid cancellations and achieve stabilization and tracking.
Many researches were proposed using the adaptive robust control .
The principal objective: Choose the control vector in such a way that:
The output to be controlled is:
( )
Control Law Step1: We define the variable representing the error between .
With is the desired value of rotor electric angular pulsationand Let E Consider a Lyapunov function Taking the time derivative of , we have .
In this step, our objective is to design a virtual control law which makes Define a virtual control and let ) with ( ) means the expression of the steady state.
We can now select an appropriate virtual control by choosing N N ) .
Where is a positive constant.
Let be the variable representing the error between the actual and virtual control Then the time derivative of .
Step2:We define C.
/C .
/C .
With is the unknown parameter.
Consider a Lyapunov function The time derivative of Lyapunov function candidate can be computed as following A New Induction Motor Adaptive Robust Vector Control Based on Backstepping (Mhamed Madar.
A ISSN: 2088-8708
EAE )
E (
)C ( )
Where C ( ) The control input
C ( )
is given by )C ( ) C ( ) C ( ) ) ( .
N becomes Using ( EAE
E (
Step3: We define the variable representing the error between .
With is the desired value of rotor flux and Consider a Lyapunov function candidate .
Its derivative with respect to time is .
In this step, our objective is to design a virtual control law which makes Define a virtual control ( ) with ( ) means variable state We can now select an appropriate virtual control by choosing N C ) .
C )C ( ) Where Let
) C
in the steady )) C is a positive constant.
representing the .
between the actual and virtual control Then the time derivative of )) E Step 4: We define .
Consider a Lyapunov function IJECE Vol.
No.
August 2017 : 1983 Ae 1993
IJECE
ISSN: 2088-8708
Its derivative with respect to time can be found us
C ( )
E E
Where The control input
C ( )) (
is formulated by Uing ( .
N becomes E E Adaptation Laws Step 5: Consider a Lyapunov function Its derivative with respect to time can be found us Using .
the adaptation laws are expressed by AE
RESULTS AND ANALYSIS
The algorithm proposed above, is applied to the nonlinear control of an induction machine, and has been implemented in the MATLAB/SIMILINK environment to validate its performance.
The parameters of the induction machine simulated are as shown in Table 1:
Table 1.
Parameters of Induction Machine Parameter V/U Value 220/380V 2,25 7
0,1232H
0,1122H
0,038Kgm2 0,0124Kgm2 A New Induction Motor Adaptive Robust Vector Control Based on Backstepping (Mhamed Madar.
A ISSN: 2088-8708 The controller gains have been chosen as follows The adaptation gains are chosen as The choice of controller gains and adaptation gains is not arbitrary.
Therfore, if this choice is correct, we obtain the high performances of the controller.
the error between the rotor flux and its reference is increasing.
So we lose the information which enables the adaptation law to track changes of rotor time constant.
To obtain good performance of the controller, we must respect this order of controller gains , and The initial value of the rotor resistance is chosen as the nominal value 0,7 and it increases rapidly to 0,84 .
The initial value of the load torque is chosen as 0 Nm.
In this study, the motor is running unloaded, at t >5s it set at 14Nm.
The rotor flux is required to track a constant reference value of 1Wb, the rotor speed is required to track constants references values of 50 rad/s for 0s< t<7,5s and 65 rad/s 7.
5s< t<9s.
The Simulink model of the drive is given in Figure 2.
The time histories of the actual and desired value of the flux, speed, rotor time constant and load torque are reported on Figures 3.
Figure 4.
Figure 5 and Figure 6.
The bloc induction machine AuIMAA is based on the fifth order model.
The bloc AAFLUX ESTIMATORAA is the estimator of the rotor flux and its frequency.
The block adaptive LawAA is the adaptation law of rotor constant time and load torque.
The bloc AACONTROL LAWAA is the bloc which produces the control vector.
The block AAdq-abc-PWM-InverterAA is the block of the inverter controlled by a PWM circuit based on sinusoidal modulation technique.
It is observed that the mechanical speed, the rotor flux, the load torque and the rotor time constant track the reference values accurately.
With or without load torque the robustness of the proposed approach is Figure 2.
Simulation scheme of adaptive robust vector control based on Backstepping ADAPTIVE ROBUST CONTRL:ROTOR FLUX
ADAPTIVE ROBUST CONTRL:ROTOR SPEED
ROTOR SPEED
ROTOR FLUX
Wmec Psiref Psirq Psird t.
Figure 3.
Rotor flux (W.
IJECE Vol.
No.
August 2017 : 1983 Ae 1993 Wref t.
Figure 4.
Rotor speed .
IJECE
ISSN: 2088-8708
ADAPTIVE ROBUST CONTROL:ROTOR TIME CONSTANT
ADAPTIVE ROBUST CONTRL:LOAD TORQUE
TL-est
LOAD TORQUE
ROTOR TIME CONSTANT
Tr-est t.
Figure 5.
Rotor time constant t.
Figure 6.
Load torque (N.
PERSPECTIVE of PRACTICAL IMPLEMENTATION The IFO control based on classical PI is widely used due to the simplicity of its implementation and the linearity of its steady state torque-slip characteristics .
, .
, .
The novel approach proposed in this paper was tested in MATLAB/SIMLINK environment, and the obtained results proved its feasibility.
This technique has a good dynamic performance but is difficult to implement compared to the first one.
This is due to the complexity of choosing appropriate controllers gains.
To validate experimentally these results the next study will be consecrated to develop a Rapid Control Prototyping bench.
The latter includes:
An induction machine that will be controlled by the developed approach.
A DC motor that will be used to provide load torque.
A three phase IGBT inverter controlled by sinusoidal PWM operating at 10 kHz.
A TMS320C40 digital signal processor connected to PC bus.
A set of A/D and D/A converters boards.
A PC used to develop, debug and download control programs.
Stator currents and rotor speed are measured by using Hall effect sensors and shaft encoder The flux can not be measured and must be estimated from accessible measured variables.
The frequency of rotor flux is deduced from the fifth equation of the dynamic model.
The control law, adaptation laws and rotor flux estimation algorithms, will be written in C code and downloaded to the DSP in order to implement the new vector control approach, and to estimate the actual values of rotor time constant, load torque, rotor flux and its spatial position.
CONCLUSION
Simulation work, carried out in this work, demonstrates that the controller performance is Moreover, using the adaptive Backstepping, the robustness of the proposed control scheme is Combining Backstepping and indirect field oriented control techniques constitutes a new approach for high performance control drives of induction machine.
The recursive on-line estimation predicts and tracks quickly and accurately the desired value of rotor time constant and load torque.
REFERENCES