Available online at www.sciencedirect.

com

ScienceDirect
Procedia Technology 25 (2016) 775 – 784

Global Colloquium in Recent Advancement and Effectual Researches in Engineering, Science and
Technology (RAEREST 2016)

H-Infinity Loop-Shaping Controller for Load Frequency Control of

a Deregulated Power System
Arlene Davidson Ra*, Dr. S. Ushakumarib
a
Research Scholar, Dept. of Electrical Engineering, College of Engineering Trivandrum, India
b
Professor, Dept. of Electrical Engineering, College of Engineering Trivandrum, India

Abstract

Load frequency control is an essential component of Automatic Generation Control of power systems. Deregulated power
system isa highly complex and uncertain system because of multiple bilateral transactions taking place. This necessitates the use
of advanced robust controllers for load frequency control. A decentralized load frequency controller using H-infinity based loop-
shaping method is suggested for a two area deregulated non-reheat thermal power system in this paper. The connections between
each control area with the rest of the system as well as the contracts existing in a deregulated system are treated as input
disturbance signals to achieve decentralization. Dynamic responses for three contract cases of operation in a deregulated system
using the H-infinity controller are obtained.
© 2016
© 2015TheTheAuthors.
Authors.Published
Published byby Elsevier
Elsevier Ltd.
Ltd. This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of the organizing committee of RAEREST 2016.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of RAEREST 2016
Keywords:Deregulation, load frequency control, H-infinity control, loop-shaping, unilateral, bilateral, contract violation

1. Introduction

Automatic Generation Control (AGC) is one of the significant problems in electric power system design and
operation. Any deviation in frequency with sudden load perturbation can directly influence power system operation
and system reliability. A non-nominal frequency results in a lower quality of the delivered electrical energy. Too
low frequencies lead to vibrations damaging steam turbines which in the worst case may have to be disconnected. A
large frequency deviation can cause an unstable condition for a power system. The main objectives of load
frequency control (LFC) in an interconnected power system include maintaining system frequency at nominal value
and minimizing unscheduled tie-line power flows between neighbouring control areas[1].
Electric power industry is currently undergoing transition from a vertically integrated structure to a horizontally
integrated structure with the main objectives of bringing down the price of electrical energy, improving the service
standards and developing a competitive market environment. In this new framework, generation, transmission and
distribution are handled by separate entities termed as Generation Companies (GENCOs), Transmission Companies
(TRANSCOs) and Distribution Companies (DISCOs). Since there are several GENCOs and DISCOs in the
deregulated structure, any DISCO has the freedom to make a contract with any GENCO in any control area for
purchase of power.All such transactions have to be cleared through an impartial entity called an Independent System

*Corresponding author Tel : +91-9400818313; Email id :arlenerosaline1@gmail.com

2212-0173 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of RAEREST 2016
doi:10.1016/j.protcy.2016.08.172
776 R. Arlene Davidson and S. Ushakumari / Procedia Technology 25 (2016) 775 – 784

Operator (ISO). The ISO has to control a number of ancillary services, one of which is AGC.
Several methods have been proposed to design load frequency controllers for deregulated power systems. The
operational structures resulting from deregulation are highlightedin [2]. A ramp following controller is designed for
a deregulatedsystemin [3]. Donde et al. [4] have taken into account the effect of bilateral contracts in modelling the
system and simulation are done considering bilateral contracts and contract violation. In [5], LFC synthesis problem
is formulated as a mixed H2/H∞ static output feedback control problem to obtain a desired PI controller. A
decentralized Neural Network controller for LFC in a deregulated power system is proposed in [6]. Genetic
algorithm is used for optimization of integral gains and bias factors in AGC for a three area power system after
deregulation in [7]. Genetic algorithm optimization techniques are employed for tuning PID controller gains in [8]
for a four area deregulated power system. Optimal output feedback control and reduced order observer are made use
of in [9] for a simplified model of the deregulated system.Two degree of freedom internal model control ( IMC )
method is used to tune decentralized PID type load frequency controllers in [10]for a deregulated environment.A
load following controller in deregulated scenario has been designed in [11]. Structured singular value method is
used for robustness analysis in [12]. Fractional Order PID controller is applied to AGC of multi area thermal system
with reheat turbines under deregulated environment in [13].Optimal output feedbackcontroller is used in [14] for
LFC in deregulated environment for multi-source combination of hydro, reheat thermal and gas generating units in
each control area.Optimal load frequency controller for a two area non-reheat thermal deregulated power system
using genetic algorithm is given in [15].
Being a highly complex and uncertain system because of the multiple bilateral transactions taking place, a fixed
controller will not be able to take care of the uncertainties in the system and so robust controller design is essential
for LFC in the restructured power scenario. Literature review suggests that only a very few works have been done
on application of robust controllers for load frequency control on deregulated power systems. It is also seen that H-
infinity controller based on loop-shaping has not yet been attempted in deregulated power systems. This paper deals
with the design of H-infinity loop-shaping controller for two-area deregulated non-reheat thermal power system.
Section 2 details the modelling of the deregulated power system. Section 3concerns the design and application of
H-infinity controller design based on loop-shapingfor the same system.Section 4.1 and Section 4.2 give the
controller design for Area 1 and Area 2 respectively. The dynamic response of the two area deregulated power
system with the application of the mentioned controller is given in Section 4.3. Section 5 gives the summary of the
work presented.

2. LFC In A Deregulated Power System

The LFC in a deregulated power market should be designed to accommodate all possible transactions such as
Poolco-based transactions, bilateral transactions and a combination of these two transactions. In bilateral transaction,
any DISCO has the freedom to have a contract with any GENCO in its own and other control areas whereas in
Poolco-based transactions, GENCOs participate in LFC of their own control areas only. In a competitive electricity
market, Poolco and bilateral transactions can take place simultaneously. The schematic block diagram of a two-area
deregulated power system consisting of two GENCOs and two DISCOs in each area is given in Fig. 1. In order to
meet the Poolco-based and bilateral transactions, a DISCO Participation Matrix ( DPM ) is used. The number of
rows of DPM is equal to the number of GENCOs and the number of columns equal to the number of DISCOs. Each
entry of DPM is defined as contract participation factor. Thus ‘cpfkl’ is contract participation factor between
kthGENCO and lthDISCO and indicates the fraction of the total load power contracted by DISCO ‘l’
fromGENCO‘k'. For a two area system with two GENCOs(GENCO1, GENCO2 ) and two DISCOs( DISCO1,
DISCO2 ) in area 1 and two GENCOs ( GENCO 3, GENCO 4 ) and two DISCOs ( DISCO3,DISCO 4)in area 2,
DPM is given by

(1)
 R. Arlene Davidson and S. Ushakumari / Procedia Technology 25 (2016) 775 – 784 777

Fig.1. Schematic block diagram of a two area deregulated power system

The sum of all entries in a column of DPM is unity. ie.,

∑ = 1; for l = 1,2,.....Nd (2)

where ‘Ng’ is the total number of GENCOs and‘Nd’is the total number of DISCOs. The generation of each GENCO
must track the contracted demands of DISCOs in steady state. The expression for contracted power of kth GENCO
with DISCOs is given by

∆ , = ∑ ∆ , ; fork = 1,2,.....Ng (3)

Where∆Pgc,k is the contracted power of kthGENCO and ∆ , is the total load demand of lthDISCO. Thescheduled
steady state power flow on the tie-line is expressed as the difference of total power exported from GENCOs in
control area 1 to DISCOs in controlarea 2 and total power imported by DISCOs in control area 1 from GENCOs in
control area 2.

∆ , = ∑ ∑ ∆ , − ∑ ∑ ∆ , (4)

The tie-line power error is defined as

∆ , =∆ , − ∆ , (5)

At steady state, tie-line power error, ∆ , vanishes as the actual tie-line power flow reaches the scheduled
power flow. This error signal is used to generate the respective Area Control Error ( ACE ) signal as in the
conventional power system.
778 R. Arlene Davidson and S. Ushakumari / Procedia Technology 25 (2016) 775 – 784

= ∆ + ∆ , (6)

= ∆ + ∆ , (7)

where = − where Pr1, Pr2 are the rated area capacities of area 1 and area 2 respectively.

The total load of the kth control area ∆ , is expressed as the sum of the contracted and uncontracted load demand
of the DISCOs of the kth control area.

∆ , =∑ ∆ , + ∆ , (8)

where ∆ , is the contracted load demand of the kthDISCO and ∆ , represents the uncontracted load demands of
DISCOs in ktharea.

3. Design of H-Infinity Loop-Shaping Controller

The objectives of H-infinity controller synthesis include ensuring the stability of systems in the face of uncertainties
in the system referred to as robust stability. In the control design for uncertain systems, it is necessary to know the
level of performance once stability is ensured. This is called as robust performance. The term ‘loop-shaping’ refers
to adjustment of frequency response of whole system within certain bounds so as to ensure sufficient robust
performance and robust stability [16].

Consider Gi(s) as a linear time invariant model for a given control area i with the following state space model:

̇ = + + (9.1)

= + (9.2)

where the suffix ‘i’ is used to represent the variable for a particular control area, is the plant coefficient matrix,
is the disturbance matrix, is the control matrix, is the output matrix, is the feedforward matrix, Xi is
the state variable vector, wi is the disturbance vector, uiis the control vector and yi is the measured output vector. The
H-infinitycontroller for the linear time invariant system Gi(s) with the state space realization given in (9.1-9.2) is to
find a matrix K, given by u = Kyi, such that the resulting closed loop system is internally stable and the H-infinity
norm from w to z is smaller than γ, a specified positive number, ie.,

‖ ‖∞ ˂ γ (10)

For a given plant G and controller K, the closed-loop performance objectives are given by
. ( + ) termed ‘Sensitivity, S’ which is the gain from output disturbance to controller input, or the gain
from reference signal to tracking error.
B. (( + ) ) which is the transfer function from input disturbance to plant output. The reciprocal of this
term indicates the maximum permissible additive controller perturbation for closed-loop stability.
. ( ( + ) ) which is the transfer function from output disturbance to controller output. The reciprocal of
this term represents the maximum allowable additive plant perturbation for closed-loop stability.
D. ( ( + ) ) termed ‘Complementary Sensitivity, T’ which is the transfer function from controller input
disturbance to plant output and also the same as transfer function from control input to output. The reciprocal of this
term represents the maximum permissible multiplicative plant perturbation for closed-loop stability. A. and B.are
robust performance objectives and are required to be small at low frequency while C. and D. are robust stability
objectives which are required to be small at high frequency [19]. According to the mentioned properties, in H-
infinity loop-shaping method, open loop singular value shaping is done.
 R. Arlene Davidson and S. Ushakumari / Procedia Technology 25 (2016) 775 – 784 779

The different steps in the computation of the controller are listed below :
1. Choose a desired loop-shape whose transfer function is given by Gd whose performance bound and
robustness bound are as in Fig.2.
2. Conversion of Gdtothe form in which the singular values of the nominal plant are shaped to give the
desired open-loop shape, Gd. The shaped plant can be expressed as Gs=GW, where W is a prefilter and G
is the nominal plant transfer function, so that the singular values ofGdapproximates the singular values of
Gs . Methodsfor computing W for loop-shaping within a required frequency rangeis given in [17] and
[18].
3. Computation of optimal loop-shaping controller by using Glover-McFarlane [19] normalized coprime
factor stabilization method whereby the optimal controller has the property that the singular value plot of
the shaped loop Ls=G*W*Ks, where Ks is the controller, matches the target loop shape Gd optimally,
roughly to within plus or minus 20*log(γ) db. The value of γ is an indicator of the accuracy to which the
optimal loopshape matches the desired loop shape and acts as an upper bound for the resonant peak
magnitude of the closed loop transfer function. ‘γ’ gives a good indication of robustness of stability to a
wide class of unstructured plant variations, with values in the range 1< γ<3 corresponding to satisfactory
stability margins for most typical control system designs.

Fig. 2. Desired loop-shape in Loop-shaping

4.Simulation Results

For the two-area deregulated power system having two GENCOs and two DISCOs in each area whose parameters
are mentioned in Appendix A, the above loop-shaping method was applied using the Robust Control Toolbox in
Matlab [20]. The connections between each control area with the rest of the system as well as the possible contracts
existing in a deregulated system are treated as input disturbance signals to achieve decentralization.The dynamic
responses for different target loop-shapes have been observed and the best target loop-shape was selected as Gd=1/s
for both the areas. The state space model of the two control areas of the mentioned deregulated power system is
provided in Appendix B.

4.1.Area 1
Fig. 3. (a) shows the singular value plot (S V Plot) of Area 1 withcontroller. It shows that the loop-gain has
improved tremendously which is essential to give good performance as far as reference tracking and disturbance
rejection are concerned. In the lower half of Fig. 3.(a), the S V Plot of open-loop gain is approximately the same as
reciprocal of SV plot of sensitivity function and in the lower half (below 0 dB line) the SV Plot of complementary
sensitivity function matches that of the open-loop gain with controller. This is expected because sensitivity function
780 R. Arlene Davidson and S. Ushakumari / Procedia Technology 25 (2016) 775 – 784

becomes approximately equal to the inverse of open-loop gain with controller for values of maximum singular
value of open-loop gain with controller>> 1. Also, if minimum singular value of open-loop gain with
controlleris<<1, the complementary sensitivity function approximately equals the open-loop gainwith controller.
The controller transfer function designed for Area 1 is given in (11). Fig.3.(b) shows the step response of closed
loop system of Area 1 with H-infinity loop-shaping controller (of order 10 ).

Fig. 3. (a) Singular value plot of Area 1 with controller; (b) Step response of Area 1 with controller

Transfer function of H-Infinitycontroller:

8.002e08 s + 9.856e12 s + 4.056e16 s + 5.614e19 s + 1.58e21 s + 1.456e22 s 5.618e22 s +
1.605e23 s + 2.693e23 s + 1.098e23 / (s + 2.459e04 s + 2.521e08 s + 1.378e12 s +
4.244e15 s + 6.984e18 s + 4.828e21 s + 6.725e22 s + 1.852e23 s + 1.099e23 s ) (11)
The value of for this controller was 1.4149. This shows the accuracy factor within which the designed controller
has approached the desired loop gain.
 R. Arlene Davidson and S. Ushakumari / Procedia Technology 25 (2016) 775 – 784 781

4.2.Area 2
Fig. 4. (a) shows the singular value plot (S V Plot) of Area 2 without controller. It shows that the loop-gain has
improved which is essential to give good performance as far as reference tracking and disturbance rejection are
concerned. It shows that the performance is good for low frequencies (indicated by inverse S V Plot of sensitivity
function ) and offers good robustness at high frequencies as is indicated by the singular value plot of complementary
sensitivity function. The controller transfer function designed for Area 2 is given in (12).
Transfer function of H-Infinity controller :
1.061e09 + 1.307e13 + 5.379e16 + 7.451e19 + 2.2e21 + 2.118e22 + 8.009e22 +
2.016e23 + 3.137e23 s + 1.331e23 /( + 2.459e04 + 2.521e08 + 1.379e12 + 4.246e15 +
6.991e18 + 4.838e21 + 7.475e22 + 2.125e23 + 1.333e23 s) (12)
The value of for this controller was 1.4149. This shows the accuracy factor within which the loop-gain using the
designed controller has approached the desired loop gain. The tracking performance of the system was observed
using unit step command and the response showed good performance as shown in Fig. 4.(b).

Fig. 4. (a) Singular value plot of Area 2 with controller;(b) Step responseof Area 2 with controller

4.3. Dynamic Performance of Deregulated Power System

Time domain simulations were carried out in Matlab for the deregulated power system whose system parameters are
given in Appendix A considering 10% load demand on each DISCO ie., ∆ , = ∆ , = ∆ , = ∆ , =
782 R. Arlene Davidson and S. Ushakumari / Procedia Technology 25 (2016) 775 – 784

0.1 . MW.
The three cases of deregulated power system operation include
i. Unilateral – GENCOs participate in LFC of their own control areas only. It is assumed that each DISCO has a
total load demand of 0.1 pu MW. For the DPM considered in Appendix C, Eqn. (3)of Section 2 gives the calculation
of steady state power output of GENCOs calculated on the basis of contracts with DISCOs. As explained in Section
2, cpfs given in DPM gives the value of the contracts that DISCOs make with GENCOs. Accordingly the values for
GENCO 1 and GENCO 2 are 0.1 pu MW each. The corresponding simulations for frequency deviation for the two
areas and power deviation of GENCOs are given in Fig. 5.(a) and Fig. 5.(b) respectively.
ii.Bilateral – A DISCO in an area is free to have power contract with any GENCO in any control area. For the DPM
considered in Appendix C, the steady state power output of GENCOs 1,2,3,4 are 0.105 pu MW, 0.045 pu MW,
0.195 pu MW and 0.055 pu MW respectively, calculated as per Eqn. (3) of Section 2. The corresponding
simulations for frequency deviation for the two areas and power deviation of GENCOs are given in Fig. 6.(a) and
Fig. 6.(b) respectively.
iii.Contract violation – DISCOs in an area may have excess uncontracted power demand. This uncontracted load
must be supplied by GENCOs in the same area according to their respective ACE participation factors (apfs). It is
assumed that there is an excess demand of 0.1 pu MW in area 1. So GENCOs of area 1 share the excess demand in
proportion to their apfs. For the DPM considered in Appendix C, the steady state power output of GENCOs 1,2,3,4
are 0.180 pu MW, 0.07 pu MW, 0.195 pu MW and 0.055 pu MW respectively, calculated as per Eqn. (3) of Section
2. The simulation results for frequency deviation for the two areas and power deviation of GENCOs are given in
Fig. 7.(a) and Fig.7.(b) respectively.

Fig.5.(a) Frequency deviation – Unilateral case; (b) Power generation of GENCOs _ Unilateral case

Fig. 6.(a) Frequency deviation – Bilateral case; (b) Power generation of GENCOs _ Bilateral case
 R. Arlene Davidson and S. Ushakumari / Procedia Technology 25 (2016) 775 – 784 783

Fig. 7.(a) Frequency deviation – Contract violation case; (b) Power generation of GENCOs _ Contract violation case

4. Conclusions
In this paper a decentralized H-infinity loop-shaping controller is proposed for a two-area deregulated non-reheat
thermal power system. The simulation results show that the closed loop performance objectives are achieved by
loop-shaping the open-loop gain using this controller. The results show good dynamic performance with respect to
reference tracking and disturbance attenuation at low frequencies and robustness against uncertainty especially at
high frequencies. Thus it may be concluded that this control strategy is a good control scheme for LFC problem in
deregulated power systems. A future work which can be pursued is to simulate uncertainties and see the effect of
robust stability and robust performance of the controller.

Appendix A
System Data
Kp1 = Kp2 = 127.5 Hz/pu MW
Tp1 = 25 s ; Tp2 = 31.25 s
R1 = 3 Hz/puMW ; R2 = 3.125 Hz/pu MW ;
R3 = 3.125 Hz/puMW ; R4 = 3.375 Hz/pu MW
B1 = 0.532 ; B2 = 0.495
Tg1 = 0.075 s ; Tg2 = 0.1 s ; Tg3 = 0.075 s ;
Tg4 = 0.0875 s
Tt1 = 0.4 s ; Tt2 = 0.375 s ; Tt3 = 0.375 s ;
Tt4 = 0.4 s
Appendix B
State Space Model :
First Area
−0.04 −5.1 5.1 0 5.1 0
⎡ 0.44 −5.1 0 0 0 0 0
0 0 0 0 0 ⎤ ⎡
⎢ 0 ⎥ 0 −1 0 0 0⎤ ⎡ 0⎤
0 −2.5 2.5 0 0 ⎢ ⎥ ⎢ ⎥
0 0 0 0 0⎥ 0
=⎢ ⎥; =⎢ ; = ⎢ ⎥; [0.532 1 0 0 0 0]
⎢−4.44 0 0 −13.33 0 0 ⎥ ⎢ 0 0 0 13.33 0⎥ ⎢ 10 ⎥
⎢ 0 0 0 0 −2.67 2.67⎥ ⎢ 0 0 0 0 0⎥ ⎢ 0⎥
⎣ 0 0 0 0 10⎦ ⎣2.5⎦
⎣ −3.2 0 0 0 0 −10⎦
;
=[0 0 −1 0 0]
Second Area
−0.03 −4.08 4.08 0 4.08 0 −4.08 0 0 0 0
⎡ 0.44 0 0 0 0 0 ⎤ ⎡ 0 −1 0 0 0 ⎤
⎢ ⎥ ⎢ ⎥
0 0 −2.67 2.67 0 0 ⎥ 0 0 0 0 0 ⎥
=⎢ ; =⎢ ;
⎢−4.27 0 0 −13.33 0 0 ⎥ ⎢ 0 0 0 13.33 0 ⎥
⎢ 0 0 0 0 −2.5 2.5 ⎥ ⎢ 0 0 0 0 0 ⎥
⎣−3.39 0 0 0 0 −11.43⎦ ⎣ 0 0 0 0 11.43⎦
784 R. Arlene Davidson and S. Ushakumari / Procedia Technology 25 (2016) 775 – 784

0
⎡ 0 ⎤
⎢ ⎥
0 ⎥
=⎢ ; = [0.495 1 0 0 0 0]; ; =[0 0 1 0 0]
⎢ 6.67 ⎥
⎢ 0 ⎥
⎣5.714⎦

Appendix C
Unilateral Bilateral Contract violation
DPM 0.6 0.4 0 0 0.5 0.25 0 0.3 0.5 0.25 0 0.3
0.4 0.6 0 0 0.2 0.25 0 0 0.2 0.25 0 0
0 0 0 0 0 0.25 1 0.7 0 0.25 1 0.7
0 0 0 0 0.3 0.25 0 0 0.3 0.25 0 0

References
[1] P. Kundur.Power System Stability and Control. McGraw Hill Inc; 2003
[2]Richard D Christie, Anjan Bose. Load Frequency control Issues in Power System Operations After Deregulation. In : IEEE Transactions on
Power systems, Vol.11, No.3, August 1996
[3] Bjorn H Bakken, Ove S Grande.Automatic Generation Control in a Power System. In: IEEE Trans. On Power Systems. November 1998
Vol.13, No.4.
[4] VaibhavDonde, M A Pai, Ian A Hiskens. Simulation and Optimization in AGC Systems after deregulation. In : IEEE Transactions on Power
Systems. August 2001 Vol.16, No.3.
[5] Bevrani H, Mitani Y, Tsuji K. Bilateral based robust load frequency control.In : Energy Conversion and Management 2004; 57:2297-2312
[6] Shayeghi H, Shayanfar H A,O P Malik. Robust decentralized neural networks based LFC in a deregulated power system.In: Electric Power
systems Research 2007 ;77 : 241-251
[7] Demiroren A, Zeynelgil H L. GA application to optimization of AGC in three-area power system after deregulation.In: Electrical Power and
Energy Systems 2007; 29: 230-240
[8] Praghnesh Bhatt, Ranjit Roy, S P Ghoshal. Optimized multi area AGC simulation in restructured Power systems. In: Electrical Power and
Energy Systems 2010 ;32: 311-322
[9]ElyasRakhshani, JavadSadeh. Practical viewpoints on load frequency control problem in a deregulated power system.In: Energy Conversion
and Management. 2010 ;51: 1148-1156
[10] Wen Tan, HongxiaZhang , Mei Yu. Decentralized load frequency control in deregulated environments.In: Electrical Power andEnergy
Systems 2012; 41: 16-26
[11] Rajesh Joseph Abraham, D Das, AmitPatra. Load following in a bilateral market with local Controllers.In: Electrical Power and Energy
Systems 2011;33: 1648 – 1657
[12] Wen Tan, Hong Zhou.Robust analysis of decentralised load frequency control for multi-area power systems. In: Electrical Power and Energy
Systems 2012; 43: 996-1005
[13] SanjoyDebbarma, Lalit Chandra Saikia, NidulSinha. AGC of a multi-area thermal system under deregulated environment using a non-
integer controller.In: Electric Power Systems Research 2013; 95 : 175-183
[14] K P Singh Parmar, S Majhi, D P Kothari.LFC of an interconnected power system with multi-source power generation in deregulated power
environment. In : Electrical Power and Energy Systems 2014 ;57 : 277-286
[15] Arlene Davidson R, Dr.S. Ushakumari. Optimal Load Frequency Controller for a Deregulated Non-Reheat Thermal Power System.In :
Proceedings of the IEEE International Conference on Control Communication and Computing (ICCC 2015)
[16] Zhou, K., and J.C. Doyle. Essentials of Robust Control. NewYork: Prentice-Hall; 1998
[17] Le, V.X., and M.G. Safonov. Rational matrix GCD's and the design of squaring-down compensators—a state space theory., IEEE Trans. On
Automatic Control.March 1992; AC-36(3):384–392
[18] Chiang, R.Y., and M.G. Safonov. H∞ synthesis using a bilinear pole-shifting transform.In: AIAA J. Guidance, Control and Dynamics
September–October 1992; 15(5):1111–1115
[19] Glover, K., D. McFarlane. Robust stabilization of normalized coprime factor plant descriptions with H∞-bounded uncertainty. In : IEEE
Trans. On Automatic Control. August 1992; AC-34(8):821–830
[20] Robust Control Toolbox, Matlab R 2015b