Bio-Inspired Optimization Algorithms Based

Design of Robust Controller for Single Machine

Power System Stabilizer

Manogna Bojugu1 , Satish Kumar Injeti1(B) , and Dasu Butti2

1 Electrical Engineering Department, NIT Warangal, Hanumakonda 506004, Telangana, India
drinjetisatishkumar@[Link]
2 Faculty of Electrical and Electronics Engineering Department, SRGEC (A),

Gudlavalleru 521356, A.P., India

Abstract. A power system stabilizer (PSS) is a device that provides additional

damping torque to the power system during the mitigation of oscillations caused
by various sorts of disturbances. Bio-Inspired Optimization Algorithms are used
to construct PSS on a Single Machine Infinite Bus (SMIB) system and two areas,
respectively, in this research. The study examines four different multi-machine
generator systems working under a variety of disturbances. The SMIB system
uses ISE as an objective function, while the Multimachine system uses Eigen
value shifting. Modified Heffron-Phillips (MHP) model is used in the SMIB sys-
tem to lower the system complexity and processing time. The SMIB system’s
response time is improved by integrating a PID controller with MPSS. Particle
swarm optimization (PSO), whale optimization (WO) and butterfly optimization
(BO) techniques are used to find the best parameters for PSS and PID. A thor-
ough evaluation of the MPSS, MPSS-PID, and other optimization algorithms’
performance is offered based on the simulation findings.

Keywords: Bio-inspired Optimization · Power System Stabilizer · SMIB · PID

controller · Dynamic Stability Improvement

1 Introduction
A variety of disruptions can occur in the power system, including faults, changes in
torque, and a sudden shift in load, all of which can impact the network’s power transmis-
sion capacity constraints and result in synchronism loss, system blackouts, and eventually
system failure. Low-frequency oscillations (between 0.2 and 3.0 Hz) caused by these
disturbances have a significant impact on the dynamic stability and performance of the
system. Power System Stabilizer reduces these oscillations by providing the necessary
damping torque.
In order to induce either +ve or −ve damping, the excitation control dV/dt is used in
the research of PSS and excitation control in [1]. Robust Control [2] and Artificial Intel-
ligence Techniques [3–5] have been used to tune PSS parameters, while the use of Eigen
value drift as an objective function in Robust Control has made the tuning mechanism

© The Author(s) 2023

V. E. Balas et al. (Eds.): AITEES 2022, AHIS 3, pp. 148–168, 2023.
[Link]
 Bio-Inspired Optimization Algorithms 149

flexible to changes in operating conditions. Complex and time-consuming approaches

like this necessitate a substantial amount of computation time. As a result, heuristic
algorithms such as Tabu Search [6], Simulated Annealing [7], Genetic Algorithms [8],
Particle Swarm Optimization [9–11], Differential Evolution [12], Artificial Bee Colony
[13–15] and Honey Bee Algorithm [16] are being considered in recent research to solve
complex problems. As our power system is highly non-linear coupled and operates under
a variety of operational situations, these approaches necessitate the use of an optimal
PSS design strategy. PSO stands out as a strategy with the capacity to address a wide
range of optimization issues. Modern optimization and Hybrid research on PSO and GA
has been increasing in recent years [17–19, 26]. By using an infinite bus voltage, all of
the strategies outlined above can be compared.
An HP model PSS design relies on data that may not be immediately available from
the framework’s respective generator. The PSS structure for such a system necessitates
extensive knowledge of system parameters external to the generating station. If the sys-
tem settings aren’t correct, it may be impossible to do an accurate PSS configuration.
Because consumer demand is constantly shifting, it’s challenging to develop a PSS that
can withstand these constantly shifting operational conditions (due to change in load
demand). Such a wide range of operating conditions makes conventional PSS parameter
design problematic, as complexity and computation time will increase. Researchers are
inspired by this to develop a PSS design that does not rely on external system data.
Additionally, PSS incorporates a PID controller to enhance the system’s dynamic per-
formance. The transformer secondary voltage of the generator is used instead of the
infinite bus voltage in the Modified Heffron Phillip Model [20, 21] for the design of PSS
parameters. This update allows PSS to be constructed in a way that takes into account
local information and data, rather of relying solely on data from the external system.
However, the PSS design process will become tedious and complex while working in a
variety of environments. Thus, Modified Power System Stabilizer (MPSS) is developed
to overcome the above drawback for the SMIB system. P. Kundur [22] had built up a
precise strategy for PSS tuning and executed on Ontario Hydro station. To check PSS
performance and robustness, the PSS designing technique is the extent to benchmark
Multi-Machine system (Two Area-Four Generator system) [23]. Many authors reported
great results on this particular engineering problem. From the literature, it has been
observed that the solution to this problem is still evolving because of the latest method-
ologies and evolution of novel optimization algorithms availability. So, in this paper an
attempt has been made to find a better solution through implementing efficient optimiza-
tion algorithms such as PSO, WOA and BOA to optimal parameter selection of power
system stabilizer in case of single as well as multi-machine two are power system.

2 System Modelling
2.1 Mathematical Model for MHP (for SMIB) Model

Modified Heffron-Phillips model is constructed in this study by using the transformer

secondary bus voltage as the reference instead of infinite bus voltage shown in Fig. 1(a).
SMIB system for MHP model is depicted in Fig. 1(a) using a single line diagram.
150 M. Bojugu et al.

Fig. 1(a). SMIB system for MHP model

This Modified mode yields the G-constants shown below

Vso Eqo sin δso Xq − Xd

G1 = + Vso sin δso (1)
Xq + Xt Xt + Xd
Xq + Xt
G2 = iqo (2)
Xt + Xd
Xt + Xd
G3 = (3)
Xd + Xt
Xd − Xd
G4 = Vso sin δso (4)
Xt + Xd
Xq Vdo Vso cos δso X  Vqo Vso sin δso
G5 =   + d  (5)
Xq + Xt Vto Xt + Xd Vto
Xt Vqo
G6 = (6)
Xt + Xd Vto
 
Eqo sin δso Xq − Xd Iqo cos δso
Gv1 =  +   (7)
Xq + Xt Xt + Xd
 
Xd − Xd cos δso
Gv2 = −   (8)
Xt + Xd
Xqo Vdo sin δso X  Vqo cos δso
Gv3 =   + d  (9)
Xq + Xt Xt + Xd Vto

We have derived the MHP model’s G-constants above, which are similar to the HP
model K-constants, except that they are assessed by taking as a reference at generator
side the secondary side of the step-up transformer, rather than the infinite bus voltage.
The MHP model’s block diagram is shown in Fig. 1(b).
 Bio-Inspired Optimization Algorithms 151

Fig. 1(b). MHP model’s block diagram

Fig. 2. (a) Structure of MPSS (b) Block diagram of PID-MPSS

2.2 Proposed PID-MPSS

A CSMIB system’s parameters might vary under normal operating conditions, and these
parameters are difficult to determine because they are not easily accessible. The P-I-D
controllers are well-known for their ability to improve the dynamic performance of power
systems. The SMIB system’s dynamic performance can be improved by integrating
MPSS with various algorithm-tuning based P-I-D controllers. The P-I-D-MPSS block
diagram can be seen in Fig. 2(b). PSO, WOA, or PSO-PID-MPSS are used to determine
the gain setting and parameter of MPSS in the next section, respectively.
152 M. Bojugu et al.

3 Problem Formulation
3.1 SMIB System

Integral Square Error (ISE) is the most accurate and reliable objective function for
assessing the dynamic performance of a power system. The following is an illustration
of the performance index:
t
J = ∫(ω)2 dt (10)
0

where t is the simulation time and ω is the change in speed. Based on the optimization
function J, the fitness function can be expressed as:

pss ≤ K pss ≤ K pss

K min max
(11)

K min
p ≤ K p ≤ K max
p (12)

K min
i ≤ K i ≤ K max
i (13)

K min
d ≤ K d ≤ K max
d (14)

3.2 Multi-machine System

The Eigenvalues of the test systems determine the goal function. It is here that an objec-
tive function is created in order to shift Eigenvalues into proper s-plane position. The
oscillating behaviour of the system is attributed solely to the use of gently damped
Eigenvalues in the construction of the objective function. Because the system’s oscil-
latory behaviour can only be explained by gently damped Eigenvalues, these are taken
into account while constructing the goal function. By using this objective function, only
the desired poles are regarded to be relocated to their new places.

J = J1 + c ∗ J2 (15)

Np  Np 
Where, J1 = (σo − σi )2 & J2 = (ςo − ςi )2 (16)
j=1 σi ≤σo j=1 ςi ≤ςo

Relative stability (σo ) and damping ratio I are used in this example, where Np is the
population size, I the ith Eigenvalue of the population, and σo is set as 0.3. Eigen value
is set to 0.15 as well. As shown in Fig. 3(a), if just J1 is taken into account, eigenvalues
will be located in the highlighted areas. If just J2 is taken into account, the eigenvalues
will be shifted to the area shown in Fig. 3(b). An objective function can be created by
combining two single objective functions J1 and J2 by assigning them a weighting factor
of c = 10. Instance No. 3(c).
 Bio-Inspired Optimization Algorithms 153

Fig. 3. Eigen control plane

4 Optimization Algorithms
The suggested objective function is optimised using PSO, WOA, and BO bio-inspired
algorithms. For solving engineering optimization issues, it has been found that PSO
(1995) and WOA (2016) well-proven optimization algorithms because of their advan-
tages such ease of implementation and understanding. Using the most recent version of
the BO optimization algorithm (2018), a comparison study was conducted to see if the
PSO and WOA had provided a better solution than the others. No need to memorise the
specific best placements of each agent, as the BO method has no memory requirements,
is a major benefit. It is a newly created optimization algorithm with a new means of
propagating information about an agent’s fitness via fragrance.

4.1 Particle Swarm Optimization Algorithm

Based on the social comparison between fish and birds, Particle Swarm is a biologically-
inspired programme. Within an initialization zone, the PSO algorithm generates random
placements for the particles. The velocities of particles can be set to zero or too small
random values in order to prevent them from leaving the search area during the first
iterations of the search algorithm. This technique uses Eq. 17 and Eq. 18 to iteratively
update the particle velocities and positions [24] during the main loop of the process.
   
k
Vk+1
i = w ∗ Vi + C1 R1 xbest,k
i − xki + C2 R2 Gbest,k
i − xki (17)

xk+1
i = xki + Vk+1
i (18)
k
the ith particle’s velocity vector V i at the kth iteration is given by, where each value
k
of V i must lie within the range. A particle’s current position vector is given by x I at
each iteration. The best position vector for that particle up to the kth iteration is given
by xbest,k
i and the best position vector for all particles up to that iteration is given by
(Gbest,k
i ). W is the weighing function or inertia weight factor.

4.2 Whale Optimization Algorithm

As a group, most whales spend much of their time together. As the largest humpback
whale in its kind, it’s also one of the most unique. The term for this kind of hunting
154 M. Bojugu et al.

is “bubble-net feeding.” Fish schools near the surface are a favourite hunting ground
for humpback whales. This foraging is done by blowing distinct bubbles in a circle,
as has been observed. In its hunting method, bubbles are related with upward spirals
and double loops. A humpback whale dives to a depth of about 12 m before rising and
swimming toward the surface in a meandering pattern around its food. Third, a catch
circle is used in conjunction with all three of the previous manoeuvres. Spiral bubble-net
feeding movement is theoretically simulated in order to do optimization [25].
In any optimization algorithm, initially, exploration takes place then exploitation.
The Eq. 19 and Eq. 20 represent the exploration stage of WOA.
 
D = C · Xrand − X(t) (19)
 
X(t + 1) = Xrand − A · D if A > 1 (20)

A = 2a · r − a (21)

C = 2r (22)

Coefficient vectors A and C are defined. In both the exploration and exploitation stages,
a is linearly decreased from 2 to 0 while r is a random vector in [0, 1]. Experimentation
begins in Eq. 20 when the vector |A| > 1 is used, and it ends when the vector |A| < 1 is
used. Models for two exploitation strategies are shown in this way.
 ∗
X (it) − A · D if p < 0.5
X(t + 1) = ∗ (23)
D · ebl · cos(2πl) + X (it) if p > 0.5
x*(t) − x*(t), where | | denotes the absolute value, is the position vector we’re interested
in; (·) represents an element-by-element multiplication; The distance between the prey
and the whales is (D ) |X − (t) – X − (t)| and indicates that the best solution so far, b
is the constant for shaping the logarithmic spiral, and the random integer in [1] is l. X*
should be modified in each iteration if there is a better option. Using Eq. 23, we can
see that when p is less than or equal to 0, we can use a shrinking encircling method of
exploitation or a spiral updating mechanism of exploitation. The WOA will come to an
end once the necessary conditions have been met.

4.3 Butterfly Optimization Algorithm

Developed by Sankalp Arora and Satvir Singh in 2018, the butterfly optimization (BO)
technique is a new meta-heuristic optimization algorithm. In order to create the BO
algorithm, we take into account the butterfly’s food-gathering and mating habits. But-
terflies rely on their keen sense of smell to locate food and a mate in the wild. Each
butterfly emits a distinct scent as it searches for food, and the strength of the scent is a
direct reflection of the quality and amount of food available in the immediate area. The
butterfly’s scent will go for a long way. As soon as the scent was detectable, additional
butterflies would fly towards it. To get to a suitable food supply, butterflies will have to
travel around in the real world in this manner.
 Bio-Inspired Optimization Algorithms 155

Every butterfly in the BO algorithm is viewed as a search agent. Each agent is

assigned a role and a distinct scent to go with it. As each agent’s ability to carry out
its objective function improves, so does its scent. Equation 24 provides a mathematical
model for the aroma.

f = cIa (24)

f is the perceived scent magnitude, I is the stimulus intensity, c is the sensor mode, and
an is the power exponent. The search agent’s or butterfly’s fitness is represented by the
letter I in the BO algorithm. BO algorithm control parameters c and a were analysed
in detail in [27] and are referred to in this article. If the best agent is chosen, all agents
will migrate to their new locations based on a switch probability “p” and the magnitudes
of all their perfumes, with the exception of one. P, the likelihood of switching between
local and global search options, determines which path the agent will take. Below are
the formulae for updating your position.
Perform a global search using Eq. 25 if rand < P
 
xid (t + 1) = xid (t) + r 2 ∗ gbest − xid (t) ∗ fi (25)

or local search using Eq. 26 if rand > P

 
xid (t + 1) = xid (t) + r 2 ∗ xjd (t) − xkd (t) ∗ fi (26)

where xjd (t) and xkd (t) are Butterflies from the same swarm in the solution space as Jth
and Kth and a random number in the range of [0, 1].
PSO, WOA, and BO algorithms can be implemented using the following detailed
procedures.

The first step is to set up the problem and algorithm settings.

Prior to implementing the algorithm, a number of parameters must be set up, includ-
ing population size, dimension, maximum number of iterations (itermax), and acceler-
ation constants c1, c2, as well as the probability switch P, power exponent a, and sensor
modality for PSO, WOA, A, and C and BO algorithms, respectively Set the upper and
lower bounds of variables to their default values (MPSS or PID parameters).
Step 2: Random generation of PID gains
⎡ 1 ⎤
x1 x21 · · · xd1 −1 xd1
⎢ x2 x22 · · · xd2 −1 xd2 ⎥
⎢ 1 ⎥
⎢ . . . . .. ⎥
X =⎢ .⎢ . . . . ⎥ (27)
. . . . ⎥
⎢ pop−1 pop−1 pop−1 pop−1 ⎥
⎣ x1 x2 · · · xd −1 xd ⎦
pop pop pop pop
x1 x2 · · · xd −1 xd
j  
xi = xmin,i + xmax − xmin,i ∗ rand() (28)
156 M. Bojugu et al.

PID gains, i.e., the jth population of the ith parameter, are created randomly between
the limits as x (max,i ) and x (min,i ) are the ith parameter limits, and rand() is a random
number in between 0 and 1. d is the number of decision variables.
 
Soln = Kp , Ki , KD . . . . (29)

Soln symbolises a swarm of particles in the PSO algorithm. In each particle, there
is a PID gain. Soln is a collection of search agents in WOA, BO.
Step 3: Fitness evaluation (Objective function).
Equation 10 and Eq. 15 should be used to calculate the fitness value for each initial
solution, and the gbest solution should be recorded in the case of BO, X should be
recorded in the case of the WOA algorithm,
Step 4: Set iteration count = 0.
Step 5: During this step, PSO, WOA, and BO algorithms begin their evolutionary process.
Update the number of iterations by one.
Step 6: Use Eq. 17 to update particle velocities, and then Eq. 18 to update particle position
for the PSO algorithm.
In WOA algorithm update X using Eq. 23.
Figure 4 depicts a thorough flow chart for implementing optimization methods.
For BO algorithm, calculate the fragrance fN for each agent or butterfly using Eq. 24
and then perform a global search and local search as follows.
If rand < probability P perform global search using Eq. 30.
 
solndN (t + 1) = solndN (t) + r 2 ∗ gbest − solndN (t) ∗ fN (30)

If rand > probability P perform a local search using Eq. 31

 
solndN (t + 1) = solndN (t) + r 2 ∗ solndj (t) − solndk (t) ∗ fN (31)

where solndj (t) and solndk (t) are A random value between [0, 1] is chosen as the rth element
of the solution space and used as the initialization vector.
Step 7: Fitness evaluation (Objective function).
Each initial solution’s fitness value is calculated using Eq. 10 and Eq. 15; the best
solution is recorded for each solution in BO and X in [Link], xbest,k
i , Gbest,k
i
for the PSO algorithm.
Step 8: Stopping criterion.
The computation is finished and the results are printed if the iteration count surpasses
the specified maximum. Steps 7 through 10 must be repeated if this is not the case.
 Bio-Inspired Optimization Algorithms 157

Fig. 4. Flow chart for implementing optimization methods.

5 Simulation Results and Discussion

At four different loading circumstances, PSS is tested in two cases (i.e. two different
generator parameters) for the SMIB system: higher loading, nominal loading, weak
158 M. Bojugu et al.

loading and lead power factor loading. For each of the two scenarios, detailed system
information is provided in Appendix A. The proposed methods were tested on the MHP
model in the MATLAB environment to see how well they worked. All the proposed
algorithms were run for several times with Integral Square Error (ISE) was used as
an objective function to test each algorithm. The suggested stabilizer’s efficacy and
robustness are examined using two disturbances, which are a 10% step change in T m
and a 10% step change in V ref for all operational points in each example.

5.1 SMIB Case-1

The MPSS system’s performance is tested under four different operating situations using
the SMIB system, which uses G-constants that have been adjusted. Integral Square
Error (ISE) serves as the objective function for the PSO, WOA, and BOA optimization
algorithms. The usage of a stabiliser in a classical manner can reduce calculation time
and increase dynamic performance. Appendix A contains the generator’s SMIB system
data parameters. The dynamic performance of the system has been improved by utilising
a PID controller. Gain settings, i.e., Kp, Ki, and Kd, for PID controllers are optimised
using presented techniques. Table 1 lists the ranges of various PSS parameters. Table 2
is a list of the various operating points. Two types of disturbances are looked for in the
system: a 10% step change at Tm and a 10% step change at Vref.
The simulation results for PSO, WOA & BOA Algorithms for four operating points
for disturbances of 10% step change at Tm for SMIB with MPSS and MPSS-PID

Table 1. Range of various PSS parameters (SMIB)

S. No. Range for Operating Points

1 2 3 4
Kpss 1 & 50 1 & 50 1 & 50 1 & 50
T 1 , T2 , 0.001 & 1 0.001 & 1 0.001 & 1 0.001 & 1
T 3 , T4
KP 0.1 & 12 0.1 & 12 0.1 & 12 0.1 & 12
KI 0.1 & 12 0.1 & 12 0.1 & 12 0.1 & 12
Kd 0.01 & 5 0.01 & 5 0.01 & 5 0.01 & 5

Table 2. SMIB Case-1 Operating conditions

Operating Points Xe Pt (p.u.) Qt (p.u.)

1 0.3-Higher Loading 1.1 0.5
2 0.4-Nominal Loading 0.8 0.41
3 0.8-Weak Loading 0.4 0.1
4 0.4-Lead p.f. Loading 0.7 −0.2
 Bio-Inspired Optimization Algorithms 159

Fig. 5. Speed deviation for 10% step change at Tref for MPSS-SMIB Case-1 at Operating
Point 1

Fig. 6. Speed deviation for 10% step change at Tref for MPSS-PID-SMIB Case-1 at Operating
Point 1

are shown in Fig. 5 and Fig. 6 and disturbances of 10% step change Vref for SMIB
with MPSS and MPSS-PID are shown in Fig. 7 and Fig. 8 respectively for operating
condition 1. In each figure is having four curves namely: No MPSS (green), MPSS-PSO
(red), MPSS-WOA (blue) and MPSS-BOA (black). The gain settings are determined
by considering Integral Square Error (ISE) as an Objective Function. An algorithm has
been run for 50 iterations and optimised parameters for MPSS-PSO, MPSS-WOA and
MPSS-BOA with and without tuned PID Controller are given in Table 3 and Table 4.
160 M. Bojugu et al.

Table 3. Optimized parameters of MPSS with PSO, WOA and BOA for SMIB Case-1

Parameters PSO-MPSS WOA-MPSS BOA-MPSS

Operating Points
1 2 3 4 1 2 3 4 1 2 3 4
Kpss 10.234 10.354 10.014 10.128 10.098 10.074 10.014 10.129 10.154 9.9954 10.032 10.945
T1 0.9865 0.9025 0.9925 0.9854 0.9465 0.9654 0.0100 0.9965 0.978 0.9632 0.9547 0.9832
T2 0.393 0.0257 0.4249 0.2681 0.5222 0.3981 0.424 0.4468 0.1220 0.3981 0.1541 0.4468
T3 0.9865 0.0063 0.9925 0.9854 0.6105 0.9654 0.9951 0.9965 0.978 0.9632 0.9547 0.9832
T4 0.393 0.4622 0.4249 0.268 0.3713 0.3981 0.424 0.4468 0.1220 0.3981 0.1541 0.4468

Table 4. Optimized parameters of PID-MPSS with PSO, WOA and BOA for SMIB Case-1

Operating Points PSO-PID-MPSS WOA-PID-MPSS BOA-PID-MPSS

Optimization Parameters
Kpss Kp Ki Kd Kpss Kp Ki Kd Kpss Kp Ki Kd

1 20.354 4.654 8.114 0.105 20.095 5.095 9.0354 0.198 8.824 11.098 8.9325 0.832
2 19.654 4.001 7.981 0.1032 8.0147 10.62 8.354 0.852 5.025 5.0251 9.9654 2.2798
3 20.065 3.987 8.147 0.0541 20.659 5.032 9.521 0.0431 8.154 10.692 9.1467 0.9903
4 18.475 5.258 8.394 0.325 10.521 4.958 3.324 1.085 9.996 4.8254 3.325 0.9460

Table 5. Time Response Specification for SMIB Case-1 using PSO, WOA and BOA Algorithms

Operating PSO
Points With MPSS With PID-MPSS
Settling Peak Time Rise Time (tr ) Settling Peak Time Rise Time (tr )
Time (ts ) (tp ) Time (ts ) (tp )
1 4.4543 0.7600 0.0000000202 3.5022 0.6800 0.0000000245
2 6.1625 0.7900 0.0000001846 3.3695 0.7000 0.0000033627
3 9.364 0.8800 0.0000003727 3.3400 0.7700 0.0000459879
4 4.6780 0.8000 0.0000000497 3.8400 0.7300 0.0000126464
WOA
1 4.5399 0.7500 0.0000000174 3.1451 0.6700 0.0000000184
2 5.4530 0.8000 0.0000000467 3.2834 0.6900 0.0000643617
3 8.1825 0.9000 0.0000236033 3.1254 0.7700 0.0001250593
4 4.5326 0.8000 0.0000000632 3.5124 0.7400 0.0000000787
BOA
1 3.3344 0.7000 0.0000012152 1.2522 0.6400 0.0002973060
(continued)
 Bio-Inspired Optimization Algorithms 161

Table 5. (continued)

Operating PSO
Points With MPSS With PID-MPSS
Settling Peak Time Rise Time (tr ) Settling Peak Time Rise Time (tr )
Time (ts ) (tp ) Time (ts ) (tp )
2 2.9519 0.7000 0.0000094311 1.1904 0.6300 0.0002336126
3 2.6175 0.7000 0.0000052016 1.7873 0.6700 0.0006541999
4 4.3412 0.765 0.0001165013 1.6291 0.6300 0.0000587452

Fig. 7. Speed deviation for 10% step change at Vref for MPSS-SMIB Case-1 at Operating
Point 1

Fig. 8. Speed deviation for 10% step change at Vref for MPSS-PID-SMIB Case-1 at Operating
Point 1

Values presented in Table 5 shows the time response specifications i.e., rise time
(tr ), peak time (tp ) & settling time (ts ) for MPSS with and without PID controller tuned
using three bio-inspired algorithms such as PSO-PID, WOA-PID and BOA-PID. It has
been observed that system without PSS becomes unstable for variable operating points
162 M. Bojugu et al.

with introducing forced disturbances and when the fixed gain stabilizer is introduced,
the system becomes stable. For Higher Loading, the Settling time for MPSS-PSO is
improved from 4.4543 s to 3.5022 s with the use of PID Controller. For MPSS-WOA,
settling time is improved from 5.4530 s to 3.2834 s and for MPSS-BOA it is improved
from 3.3344 s to 1.2522 s. Peak time and Rise time also gets improved and the same
follows with other Loading conditions (Operating Points). Figure 9 and Fig. 10 shows
the comparison of convergence characteristics.

Table 6. SMIB Case-2 Operating conditions

Operating Points Xe Pt (p.u.) Qt (p.u.)

1 0.3-Higher Loading 1.0 0.2
2 0.4-Nominal Loading 0.8 0.41
3 0.8-Weak Loading 1.0 0.5
4 0.4-Lead p.f. Loading 1.0 −0.5

Fig. 9. Convergence Plots for operating condition 1 of MPSS-SMIB Case-1 with PSO, WOA and
BOA at Tref

Fig. 10. Convergence Plots for operating condition 1 of MPSS-PID-SMIB Case-1 with PSO,
WOA and BOA
 Bio-Inspired Optimization Algorithms 163

5.2 SMIB Case-2

The machine data and exciter data for this case is given Appendix A. In this case, the
performance of proposed stabilizer is connected with PID controller and is tuned with
PSO, WOA & BOA optimization algorithms. The Operating point for this case is listed
in Table 6. After several iterations, optimised parameters are given in Table 7 and Table 8.
Table 7 depicts the parameters for MPSS tuned with algorithms and Table 8 gives the
parameters for MPSS tuned with PID Controller. The gain settings are determined by
considering Integral Square Error (ISE) as an Objective Function.
In this case also the performance of Algorithms tested on four operating conditions
under the two disturbances i.e., 10% step change at Tm and 10% step change at Vref .
The simulation results for PSO, WOA & BOA Algorithm for disturbances of 10% step
change at Tm are shown in Fig. 11 and Fig. 12 and disturbances of 10% step change
Vref are shown in Fig. 13 and Fig. 14.
The results clearly show that speed deviation (oscillations) is much more reduced
with the PID controller with proposed algorithms and settled faster compared to the
conventional approach. Table 9 given below is for time response specifications for MPSS
tuned PSO, WOA and BOA Algorithms with & without PID Controller. Rise time (tr ),
Peak time (tp ) & Settling time (ts ) are compared.
In this case also, without MPSS system becomes unstable for all operating points.
And with MPSS system becomes stable. For 1st operating condition, settling time for
PSO-MPSS is improved from 4.4543 s to 2.0062 s with PID controller tuned PSO i.e.,
MPSS-PID-PSO. For WOA, settling time is shifted from 3.9005 s (MPSS-WOA) to
1.6692 s (MPSS-PID-WOA). For BOA, settling time is shifted from 3.9005 s (MPSS-
BOA) to 1.6691 s (MPSS-PID-BOA).

Fig. 11. Speed deviation for 10% step change at Tref for MPSS-SMIB Case-2 at Operating
Point 1
164 M. Bojugu et al.

Fig. 12. Speed deviation for 10% step change at Tref for MPSS-PID-SMIB Case-2

Fig. 13. Speed deviation for 10% step change at Vref for MPSS-SMIB Case-2 at Operating
Point 1

Table 7. Optimized parameters of MPSS with PSO, WOA and BOA for SMIB Case-2

Parameters PSO-MPSS WOA-MPSS BOA-MPSS

Operating Points
1 2 3 4 1 2 3 4 1 2 3 4
Kpss 2.354 2.246 6.241 10.08 40.21 39.45 35.73 40.025 2.032 2.032 6.52 10.214
T1 0.965 1.012 0.998 0.996 0.0198 0.895 1.001 0.1100 0.967 1.012 1.000 1.000
T2 0.176 0.207 0.315 0.347 0.0135 0.599 0.738 0.0284 0.1760 0.2272 0.361 0.3479
T3 0.965 1.012 0.998 0.996 0.0135 0.895 1.001 0.120 0.967 1.012 1.00 1.00
T4 0.176 0.207 0.315 0.347 0.0135 0.895 0.468 0.0284 0.1760 0.2272 0.361 0.3479
 Bio-Inspired Optimization Algorithms 165

Table 8. Optimized parameters of PID-MPSS with PSO, WOA and BOA for SMIB Case-2

Parameters PSO-PID-MPSS WOA-PID-MPSS BOA-PID-MPSS

Optimization Parameters
Kpss Kp Ki Kd Kpss Kp Ki Kd Kpss Kp Ki Kd

Operating Point 1 40.021 2.821 11.832 0.0054 40.321 2.594 11.521 0.5004 40.021 2.821 8.1134 0.200
Operating Point 2 5.041 5.00 10.54 2.41 40.21 2.510 7.00 0.200 5.014 5.21 10.33 2.012
Operating Point 3 6.0352 5.0014 8.064 2.0014 25.0012 3.625 0.5012 0.1501 6.021 8.002 10.021 0.200
Operating Point 4 6.512 5.014 8.654 2.014 40.021 2.001 14.52 0.151 6.021 10.35 10.120 2.001

Table 9. Time Response Specification for SMIB Case-2 using PSO, WOA and BOA Algorithms

Operating PSO
Points With MPSS With PID-MPSS
Settling Peak Time Rise Time (tr ) Settling Peak Time Rise Time (tr )
Time (ts ) (tp ) Time (ts ) (tp )
1 4.4543 0.6800 0.0000095464 2.0062 0.6400 0.0002281065
2 4.4056 0.7300 0.0000005871 3.2834 0.6900 0.0000643617
3 4.9278 0.7500 0.0000017379 2.4815 0.7100 0.0000636908
4 4.9946 0.7000 0.0000055255 3.1115 0.6400 0.0000508992
WOA
1 3.9005 0.7000 0.0000012152 1.6692 0.6200 0.0002258219
2 4.4046 0.7300 0.0000005868 2.6078 0.6700 0.0000744554
3 4.9115 0.7500 0.0000017379 1.8783 0.6900 0.0000042916
4 4.6073 0.7200 0.0000046539 2.6693 0.6400 0.0001855305
BOA
1 3.9005 0.7000 0.0000012541 1.2520 0.6400 0.0002973060
2 4.5846 0.7200 0.0000094311 1.1904 0.6300 0.0002336126
3 4.6715 0.7120 0.0000065201 1.7873 0.6700 0.0000465199
4 4.5539 0.7650 0.0000165013 1.6291 0.6300 0.0001187452

Also, it can be observed that peak time is reduced thereby improving the stability
of the system by using a PID controller. Rise time also increased that implies improved
response of the system. Settling time, Peak time & Rise time improves in each case
following all other operating conditions.
166 M. Bojugu et al.

Fig. 14. Speed deviation for 10% step change at Vref for MPSS-PID-SMIB Case-2 at Operating
Point 1

6 Conclusion

Two different generating parameters (i.e., generator parameters) for the SMIB system are
tested in this research to examine the system’s effectiveness and efficiency. To enhance
the power system’s dynamic stability, this research proposes the use of PID-MPSS, an
optimization algorithm based on three bio-inspired bio-inspirational algorithms. The
MPSS on the MHP model is constructed by taking the secondary bus voltage from the
generator side transformer rather than the infinite bus voltage reference. Adopting this
approach has the benefit of removing the reliance on external system data for local knowl-
edge on the generator side. The PID controller is used to find the best settings for MPSS
utilising the proposed algorithms. Also, this test system is tested in a variety of oper-
ational circumstances to ensure that the suggested approach works as expected. Using
simulation data, it can be determined that MPSS-PID outperforms other approaches in
terms of peak and settling time in every situation. BOA, however, is the best solution for
the MHP (SMIB) model since it requires the fewest number of parameter tuning rounds
and has the lowest error rate (ISE).
This paper proposes a PSS design method based on three bio-inspired optimization
algorithms for an interconnected power system. Three-phase fault at t = 10 s has been
successfully realised using the design technique on the aforementioned case studies.
BOA-PSS gives greater damping performance than PSO and WOA for all generating
scenarios. In comparison to the other two algorithms, the BOA’s Peak Overshoot, Settling
Time, and Peak Time have all been enhanced.

Acknowledgments. Authors are thankful to the authorities of the NIT Warangal.

Authors’ Contributions. All authors contributed equally.

 Bio-Inspired Optimization Algorithms 167

Appendix A
Case-I Machine Data
Xd = 0.1026; Xq = 0.658; = 5.67; H = 8; D = 0; fB = 60 Hz; Ke = 400; Te = 0.02 s;
Efdmax = 6 p.u; Efdmin = 6 p.u;

Case-II Machine Data

Xd = 1.6; Xq = 1.55; = 6; H = 5; D = 0; fB = 60 Hz; EB = 1 p.u; Ke = 200; Te =
0.05 s;
Efdmax = 6 p.u; Efdmin = 6 p.u;

References
1. Schleif, F.R., Hunkins, H.D.: Excitation control to improve powerline 2(6), 1426–1434 (1968)
2. Gupta, R.: Robust Nondynamic Multirate Output Feedback Technique based Power System
Stabilizers (2003)
3. Ghoshal, S.P., Chatterjee, A., Mukherjee, V.: Expert systems with applications bio-inspired
fuzzy logic-based tuning of power system stabilizer. Exp. Syst. Appl. 36(5), 9281–9292 (2009)
4. Chaturvedi, D.K., Malik, O.P., Fellow, L.: Neurofuzzy power system stabilizer 23(3), 887–894
(2008)
5. Khalil, A.A., Ahmied, N.M.: Optimal tunning of lead-lag and fuzzy logic power system
stabilizers using particle swarm optimization. Exp. Syst. Appl. 36(2), 2097–2106 (2009)
6. Abido, M.A.: A novel approach to conventional power system stabilizer design using tabu
search. Int. J. Electr. Power Energy Syst. 21(6), 443–454 (1999)
7. Abido, M.A.: Robust design of multimachine power system stabilizers using simulated
annealing. IEEE Trans. Energy Convers. 15(3), 297–304 (2000)
8. Kim, D.H.: Hybrid GA–BF based intelligent PID controller tuning for AVR system 11, 11–22
(2011)
9. Mostafa, H.E., El-sharkawy, M.A., Emary, A.A., Yassin, K.: Electrical power and energy sys-
tems design and allocation of power system stabilizers using the particle swarm optimization
technique for an interconnected power system. Int. J. Electr. Power Energy Syst. 34(1), 57–65
(2012)
10. Olivas, F., Amador-Angulo, L., Perez, J., Caraveo, C., Valdez, F., Castillo, O.: Comparative
study of type-2 fuzzy particle swarm, Bee Colony and Bat Algorithms in optimization of
fuzzy controllers (2017)
11. Abido, M.A.: Optimal design of power-system stabilizers using particle swarm optimization
17(3), 406–413 (2002)
12. Wang, Z., Chung, C.Y., Wong, K.P., Tse, C.T.: Robust power system stabiliser design under
multi-operating conditions using differential evolution 2(5), 690–700 (2008)
13. Karaboga, D., Basturk, B.: A powerful and efficient algorithm for numerical function
optimization: artificial bee colony (ABC) algorithm, 459–471 (2007)
14. Karaboga, D., Basturk, B.: On the performance of artificial bee colony (ABC) algorithm.
Appl. Soft Comput. 8, 687–697 (2008)
15. Mohammadi, M., Ghadimi, N.: Optimal location and optimized parameters for robust power
system, 1–17 (2014)
168 M. Bojugu et al.

16. Elazim, S.M.A., Ali, E.S.: Electrical power and energy systems optimal power system sta-
bilizers design via Cuckoo Search algorithm. Int. J. Electr. Power Energy Syst. 75, 99–107
(2016)
17. Shakarami, M.R., Davoudkhani, I.F.: Wide-area power system stabilizer design based on
Grey Wolf Optimization algorithm considering the time delay. Electr. Power Syst. Res. 133,
149–159 (2016)
18. Islam, N.N., Hannan, M.A., Shareef, H., Mohamed, A.: An application of backtracking
search algorithm in designing power system stabilizers for large multi-machine system.
Neurocomputing 237, 175–184 (2016)
19. Abdulkhader, H.K., Jacob, J., Mathew, A.T.: Fractional-order lead-lag compensator-based
multi-band power system stabiliser design using a hybrid dynamic GA-PSO algorithm (2018)
20. Gurrala, G., Sen, I.: A modified Heffron-Phillip’s model for the design of power system
stabilizers (2008)
21. Butti, D., Kumar, S., Rao, S.: Design of robust modified power system stabilizer for dynamic
stability improvement using particle swarm optimization technique. Ain Shams Eng. J. 10
(2019)
22. Kundur, P., Klein, M., Rogers, G.J., Zywno, M.S.: Application of power system stabilizers
for enhancement of overall system stability. IEEE Trans. Power Syst. 4(2), 614–626 (1989)
23. Dasu, B., Sivakumar, M., Srinivasarao, R.: Interconnected multi-machine power system stabi-
lizer design using whale optimization algorithm. Prot. Control Mod. Power Syst. 4, 2 (2019).
[Link]
24. Kennedy, J., Eberhart, R.: Particle swarm optimization. In: Proceedings of the IEEE
International Conference on Neural Networks, ICNN’95, vol. 11, no. 1, pp. 111–117 (1995)
25. Mirjalili, S., Lewis, A.: The Whale optimization algorithm. Adv. Eng. Softw. 95, 51–67 (2016)
26. Arora, S., Singh, S.: Butterfly optimization algorithm: a novel approach for global optimiza-
tion. Soft. Comput. 23(3), 715–734 (2018). [Link]
27. Injeti, S.K., Thunuguntla, V.K.: Optimal integration of DGs into radial distribution network in
the presence of plug-in electric vehicles to minimize daily active power losses and to improve
the voltage profile of the system using bio-inspired optimization algorithms. Prot. Control
Mod. Power Syst. 5(1) (2020)

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-
NonCommercial 4.0 International License ([Link]
which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any
medium or format, as long as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter’s Creative
Commons license, unless indicated otherwise in a credit line to the material. If material is not
included in the chapter’s Creative Commons license and your intended use is not permitted by
statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder.