Harmonic Minimization In Multilevel Inverters By Using PSO

ACEEE Int. J. on Control System and Instrumentation, Vol. 02, No. 03, October 2011

Harmonic Minimization In Multilevel Inverters By
Using PSO
1
N. Prashanth, 2B. Kumar, 3J. Yadagiri, 4A.Dasgupta
1
PG Scholar,Electrical Engg., KIIT University Bhubaneswar , India
prashanthkumar0228@gmail.com
2
Lecturer, Electrical Engg, IGIT Sarang Bhubaneswar,India,
brijesh_elect@yahoo.co. in
3
Assistant. Professor Electrical Engineering KIIT University Bhubaneswar, India
giri055@gmail.com
4
Professor, Electrical Engineering KIIT University Bhubaneswar, India
adasguptafel@kiit.ac.in

Abstract— Harmonic Elimination in a multilevel inverters is exist, are required. The theory of resultants of polynomials
an optimization problem which is solved by applying particle and the theory of symmetric polynomials has been suggested
swarm optimization (PSO) technique. The derived equation to solve the polynomial equations obtained from the
for the computation of total harmonic distortion (THD) of the
transcendental equations[4]. A difficulty with these
output voltage of the multilevel inverter is used as the objective
approaches is that for several H-bridges connected in series,
function in PSO algorithm. The objective function used is to
reduce the THD of the multilevel inverter and obtain the the order of the polynomials become very high thereby
corresponding switching angles with the elimination of making the computations of the solutions of these
possible lower order harmonics. In this paper a pseudo code polynomials very complex. Optimization technique based on
based algorithm is proposed to deal with inequality constraints Genetic Algorithm (GA) was proposed for computing
which will helps in accelerating the optimization process. The switching angles for 7-level Inverter. The implementation of
proposed method is applied for seven level cascade inverter to this approach requires proper selection of certain parameters
eliminate the 5th and 7 th order harmonics to reduce the total such as population size, Mutation rate etc, thereby its
harmonic distortion .This proposed PSO algorithm is effective
implementation becomes also difficult for higher level inverters
in reducing the total harmonic distortion corresponding the
[2]. To circumvent the above problems, in this paper we
range of modulation index. The simulation results shows that
the proposed PSO method is indeed capable of obtaining presented the application of PSO for solving these equations
higher quality of solutions to eliminate 5 th and 7 th order and obtain the harmonic distortion lesser compare to other
harmonics and to reduce the total harmonic distortion of 7- techniques which are used in SHE method. This paper
level cascade inverter presents the analysis of 7-level cascaded inverter with three
H-bridges, solution of three transcendental equations using
Keywords— selective harmonic elimination(SHE),particle PSO and elimination of lower order harmonics such as 5th
swarm optimization(PSO),Total harmonic distortion(THD) and 7th order with minimum harmonic distortion and obtain
desired fundamental voltage.
I. INTRODUCTION
Multilevel inverters have drawn tremendous interest in II. CASCADED MULTILEVEL INVERTER
high power applications. It synthesize a desired output voltage The cascade multilevel inverter consists of a series of H-
from several levels of dc voltages as Inputs. By taking bridge (single-phase full-bridge) inverter units. As stated
sufficient number of dc sources, a nearly sinusoidal voltage above, the general function of the multilevel inverter is to
waveform can be synthesized[1] synthesize a desired voltage from several separate dc sources
To control the output voltage and reduce the harmonic (SDCSs) such as solar cells, fuel cells, ultra capacitors, etc.
distortion several methods have been presented. In those Fig.1 shows a single-phase structure of a seven level cascade
the mainly using methods are sinusoidal pulse width inverter with SDCSs. Each SDCS is connected to a single-
modulation and selective harmonic elimination method (SHE). phase full-bridge inverter and can generate three different
The SHE method is a switching technique for which voltage outputs, +Vdc, 0 and -Vdc. This is accomplished by
transcendental equations characterizing harmonics are solved connecting the dc source to the ac output side by using
to compute Switching angles. It is difficult to solve the SHE different combinations of the four switches and the ac output
equations as these are highly nonlinear in nature and may of each level’s full-bridge inverter is connected in series such
produce simple, multiple, or even no solutions for a particular that the synthesized voltage waveform is the sum of all of the
value of modulation index. Iterative numerical techniques has individual inverter outputs.
been implemented to solve the SHE equations producing only The number of output phase (line-neutral) voltage levels
one solution set, and even for this a proper initial guess and in a cascade multilevel inverter is then 2S+1, where S is the
starting value of modulation index for which the solutions
© 2011 ACEEE 1
DOI: 01.IJCSI.02.03.6


number of dc sources. For example the phase voltage output inverter, the fundamental voltage in terms of switching angles
of seven level inverter as shown in Fig.2, there are three is given in equation(2).
switching angles and three non-linear transcendental
4Vdc
equations, where ‘s’ is the number of H-bridges connected in V1  cos(1   cos( 2 )  cos( 3 )); … (2)
cascade per phase[3]. Among s number of switching angles, 
generally one switching angle is used for fundamental voltage The Modulation index is defined as the ratio of fundamental
selection and the remaining (s-1) switching angles are used voltage (V1) to the maximum obtainable fundamental voltage
to eliminate certain predominating lower order harmonics. In (V1max). The maximum fundamental voltage is obtained when
three-phase power system with isolated neutral, triplen all the switching angles are zero and the V1max is given as
harmonics are cancel out automatically, and only non-triplen
odd harmonics are present. 4Vdc
V1max  3  ( ) …. (3)

Then the modulation index (M) is given as

V1
M 
 4Vdc  …… (4)
3 * ( )
  
The 7-level cascaded inverter requires three H-bridges. The
non linear equations which are used to finding the switching
angles and desired fundamental voltage of 7-level inverter
are equations(5),(6)and(7).

cos(1 )  cos( 2 )  cos( 3 )  3M ......(5)
cos(51 )  cos(5 2 )  cos(5 3 )  0......(6)
cos(71 )  cos(7 2 )  cos(7 3 )  0......(7)
These are the non linear transcendental equations for
Fig.1.seven level cascaded inverter eliminating lower order harmonics such as 5th and 7th order
and get desired fundamental voltage[7]. For the given values
of M (from 0 to 1), it is required to get complete and all
possible solutions of these equations for determining the
switching angles and lower THD.

IV. PARTICLE SWARM OPTIMIZATION
Particle Swarm Optimization (PSO) refers to a relatively
new family of algorithms that may be used to find optimal
solutions to numerical and qualitative problems[6]. PSO was
introduced by Russell Eberhart and James Kennedy in 1995
inspired by social behavior of birds flocking or fish schooling.
It is easily implemented in most programming languages and
Fig.2 output voltage of 7-level cascade inverter has proven to be both very fast and effective when applied
to a diverse set of optimization problem.
III. HARMONIC MINIMIZATION ANALYSIS In PSO, the particles are “flown” through the problem
space by following the current optimum particles. Each par-
The Fourier series expansion for the above stair case wave ticle keeps track of its coordinates in the problem space, which
form is given bellow are associated with the best solution (fitness) that it has
  cos n1   cos n2    achieved so far. This implies that each particle has memory,
Van(t)   4Vdc
  sin nt   …… (1) which allows it to remember the best position on the feasible
n 1,5,7
n
 cos n   
 3 
search space that has ever visited. This value is commonly
The fundamental voltage is obtained from the calculated
called Pbest . Another best value that is tracked by the par- -
switching angles α1, α2, α3… αn. from equation(1). ‘n’ represents
the order of the harmonics and the switching angles are equal ticle swarm optimizer is the best value obtained so far by any
to the number of dc-sources. It is required to find the particle in the neighborhood of the particle. This location is
switching angles in the range of 0 to π/2 considering 5th and commonly called Gbest. The basic concept behind the PSO
7th order phase voltage to zero. For seven level cascade technique consists of change in the velocity (or accelerat
© 2011 ACEEE 2
DOI: 01.IJCSI.02.03.6


ing) of each particle toward its Pbest and Gbest posi- V. THE PROPOSED PSO ALGORITHM

tions at each time step. This means that each particle tries to The minimization of THD in multilevel inverters is
modify its current position and velocity according to the achieved by using this PSO algorithm because of its simple
distance between its current position and Pbest , and the in nature and easy to implement, computational efficient. The
procedure for getting optimized results the objective function
distance between its current position and Gbest . The posi- is taken as the THD equation is as follows
tion and velocity vectors of the ith particle of a d-dimensional
search space can be represented as 2 2 2
(V 2  V 3  ...V n)
X i  ( xi1 , xi 2 ,.............xid ) %THD   100 …. (11)
V 1
and Vi  (vi1 , vi 2 ,............vid ) respectively. On the basis The proposed PSO algorithm is given as following steps:
of the value of the evaluation function, the best previous Step1: create the random initial population size of switching
position of a particle is recorded and represented angles by considering their limitation is 0 to π/2
Step2: initialize the velocity, Pbest, Gbest, iteration count for
as Pbest i  ( Pi1 , Pi 2 ,....................Pid ) . If the g th par- computing switching angles
ticle is the best among all particles in the group so far, Step 3: update the iteration count
Step 4: update the velocity and position according to the
it is represented as Gbest  Pbest g  ( Pg1 , Pg 2 ,..Pgd ) .
equations (8) & (9) for moving of particles in search space
The particle tries to modify its position using the current Step 5: evaluate the fitness or objective function by using
velocity and the distance from Pbest and Gbest . The modi- equation (11)
fied velocity and position of each particle for fitness evalua- Step 6: update the values of Pbest and Gbest
tion in the next iteration are calculated using the following Step 7: Is criterion achieved then go for next step otherwise
equations repeat the step3 to step6 for best solution
Step 8: select the best solution of fitness value.
v id1`  w  v id`  c1  rand 1  ( Pbest id  x id )
k k k
In this process the maximum iterations are 100 and the
k population size is 20 and using the decreasing inertia function
 c2  rand 2  (Gbest gd  xid ) …… (8) with initial weight of 0.9 and final weight of 0.4. This proposed
algorithm is mainly used to minimize the total harmonic
x id1  x id  v id1`
k k k
……..(9)
distortion for selected initial random switching angles. For
Here w is the inertia weight parameter, which controls the getting lower value of total harmonic distortion and
global and local exploration capabilities of the particle. c1 , c 2 minimization of lower order harmonics the objective function
is taken as a total harmonic distortion equation.
are cognitive and social coefficients, rand 1 and rand 2 For every iteration the value of the THD is updated for
are random numbers between 0 and 1. For the proposed their suitable best values by changing the velocity and
position of the current particles. Similarly the values of
method, c1  2, c 2  2. A large inertia weight factor is used switching angles also updated for optimum values to get
during initial exploration and its value is gradually reduced lower value of THD
as the search proceeds. The concept of time-varying inertial
weight (TVIM) is given by. VI. COMPUTUTIONAL RESULTS
itermax  iter For seven level cascade inverter it is required to eliminate
w  ( wmax  wmin )   wmin …… (10) the 5th and 7th order harmonics and get the minimization in
itermax
total harmonic distortion. The Fig.4 shows the comparison
w max  0 . 9 ; w min  0 .4 of Modulation Index Vs THD for 7-level Inverter in PSO
method. From that we observed that where the Modulation
Where, itermax is the maximum number of iterations. Index is Increases then the total harmonic distortion is
decreases to minimum value. The modulation index value is
varies for the linear range from 0.1 to 0.9. The THD value is
decreased from 21 to 5. The lowest THD value is occurred at
0.9 modulation index value. The Fig.5 shows the obtained
switching angles corresponding with their Modulation index
values. These results have been shown by using the curve
fitting tool. The switching angles (α1, α2, α3 ) are varies within
the range of 0 to 900 only.

Fig.3.concept of modification of searching point

© 2011 ACEEE 3
DOI: 01.IJCSI.02.03.6


of 0.9. The remaining methods having the value of THD are
above 10 at the higher modulation index. The proposed algo-
rithm has the advantage to overcome the drawbacks of the
programming methods
This proposed method is the fast convergence and it is
easy to implement to get optimum values of desired output.
Previous methods work has the complexity nature of solving
the selective harmonic elimination equations and gives the
slower convergence. This method uses the initial guess
Fig.4 comparison of THD Vs Modulation Index for PSO values and the objective function, to get the optimum values
of switching angles and total harmonic distortion at the same
time. The previous work of basic PSO has done the work on
taking the objective function as to get the optimum values of
switching angles [6]. The proposed PSO method eliminates
the lower order harmonics and minimizes the harmonics up to
39th order of harmonics.

VII. HARDWARE IMPLEMENTATION
The Hardware Implementation of 7-level cascade
Fig.5 comparison of switching angles Vs Modulation Index multilevel inverter requires three single phase full bridge
The Fig.6 shows the Harmonic analysis of 7-level cascade inverters and for each bridge requires four IGBT’s .The output
multilevel inverter for elimination of lower order harmonics of these three bridges is connected in series manner such
such as 5th and 7th order. The arrow marks shown in Fig.6, that getting the desired stair case output voltage waveform
shows the elimination of lower order harmonics in 7-level for given input DC Sources. The output voltage wave form is
cascaded inverter. The Fig.7 shows the comparison of approximately equal to the desired sinusoidal waveform. In
this hardware implementation, we have used the PIC16F877A
microcontroller programme to generate the gating signals to
drive the IGBT’s. In this PIC micro controller three ports are
used for generating the gating signals
A) The Components Required:

Fig.6 Harmonic Analysis for 7-level Cascade Inverter

B) Experimental setup:
The Fig.8 shows the experimental setup of 7-level
cascaded multilevel inverter. In this the three H-bridges are
used to construct the single phase 7-level cascaded inverter
and each bridge is supplied by 5V as shown .The gate driving
circuits are used to triggering the IGBT’s. In this hardware
implementation the 12 gate drive outputs are taken from gate
drive circuit and given to each IGBT’s of 7-level CSMLI.
The PIC microcontroller is used to generate the triggering
pulses for each switching device. This PIC microcontroller
Fig.7 Comparison of THD Vs switching angles of different selective programme was loaded in to PIC16F877A chip by using the
harmonic elimination methods
IC-Prog software. After loading the programme in to PIC chip
THD vs. switching angles of different programming meth- the output of this chip is connected to the Gate driving
ods and it also verify that the THD value of PSO method is circuits for giving gate pulses to all switching devices.
less compare to the other programming methods. The value
of THD for proposed method of PSO is 5 at modulation index

© 2011 ACEEE 4
DOI: 01.IJCSI.02.03.6


whereas %THD of simulation is 15.37%. The total harmonic
distortion is different for both simulation and hardware due
to blanking time of 7 micro sec provided to the H-bridges.

Fig.11 output voltage waveform of single phase 7-level CSMLI

Fig.8 Experimental setup of single phase 7-level CSMLI

C) Observations:
The input of the microcontroller is 5V and generation of
triggering pulses from microcontroller is shown in Fig.9.

Fig.12 FFT analysis of output voltage, single phase 7-level CSMLI

Fig.9 Triggering pulses from the PIC micro controller

Fig.13 Comparison of harmonics spectra for Hardware and
simulation results (Modulation Index 0.9)

VIII. CONCLUSION
The selective harmonic elimination method can used to
Fig.10 Output Voltage Waveform of 7-level CSMLI from CRO eliminate the lower order harmonics and reduce the total
The given input DC-voltage of each bridge is 5V, thus the harmonic distortion. The PSO method gives the lower %THD
total input voltage of 7-level inverter is 15V.The FFT analysis compared to the other classical Iterative methods. In recent
of output voltage waveform is shown above Fig.12. The years the strategy for elimination of harmonics in multilevel
output voltage waveform obtained in the CRO is shown in Inverters by using PSO method has been done by taking
Fig.10 and this waveform is loaded in the computer through better switching angles as the objective function. In this work
software called SW206 signal analyzer to perform the FFT lower order harmonics have been eliminated by using the
analysis. Fig.12 shows the FFT analysis and the amplitudes equation for THD as the objective function and have given
of their corresponding frequencies. Fig.11 shows the output better results in minimization of THD for up to the 39th order
voltage waveform of single-phase 7-level inverter using of harmonics.
SW206 signal analyzer. The comparison of harmonic patterns
between simulation and hardware for the modulation index
of 0.9 is shown in Fig.13. The %THD of the hardware is 27.30%
© 2011 ACEEE 5
DOI: 01.IJCSI.02.03.6


REFERENCES [4] John N. Chiasson, Leon M. Tolbert, “Elimination of Harmonics
In Multilevel Inverters Using the theory of Symmetric polynomial
[1] F. Z. Peng and J. S. Lai, “Multilevel converters—A new breed and Resultant” Proceedings of the 42nd IEEE Conference on
of power converters,” IEEE Trans. Ind. Applicant., vol. 32, pp. Decision and Control
509–517, May/June 1996. Maui, Hawaii USA, December 2003
[2] Zhong Du, Leon M. Tolbert, John N. Chiasson “Reduced [5] H.Taghizadeh and M. Tarafdar Hagh, “Harmonic Elimination
Switching Frequency Computed PWM method for multilevel Of Multilevel Inverters Using Particle Swarm Optimization”
converter control” Power Electronics Specialists Conference, 2005. Industrial electronics, 2008, ISIE 2008. IEEE International
PESC ’05. IEEE , 2005 , Page(s): 2560 – 2564 . symposium on Industrial Electronics. 18th November 2008.
[3] Jagdish Kumar, Biswarup Das, Pramod Agarwal, “Selective [6] Jagdish Kumar, Biswarup Das, Pramod Agarwal “Harmonic
Harmonic Elimination for Multilevel Inverters” by Fifteenth Reduction Technique For a Cascade Multilevel Inverters” ,
National Power Systems Conference (NPSC), IIT Bombay, International Journal of Recent Trends in Engineering , 2009
December 2008

© 2011 ACEEE 6
DOI: 01.IJCSI.02.03.6

Harmonic Minimization In Multilevel Inverters By Using PSO

More Related Content

What's hot (20)

Similar to Harmonic Minimization In Multilevel Inverters By Using PSO (20)

More from IDES Editor (20)

Recently uploaded (20)

Harmonic Minimization In Multilevel Inverters By Using PSO