DOI: 10.

54392/irjmt25118
RESEARCH ARTICLE

Power Quality Improvement in a Grid Connected Solar-PV System

with Adaptive Neuro-Fuzzy Controller Based UPQC
Chilakapati Lenin Babu a, *, T. Gowri Manohar a
a Department of EEE, Sri Venkateswara University, Tirupati-517502, India
* Corresponding Author Email: [Link]@[Link]
DOI: [Link]
Received: 20-07-2024; Revised: 04-01-2025; Accepted: 21-01-2025; Published: 30-01-2025

Abstract: Power quality (PQ) problems are frequently encountered in power grids, and these problems become more
severe as the number of devices increases. Therefore, one of the primary responsibilities of utility systems remains
to supply people with electricity by standard sinusoidal voltage and current configurations at appropriate amplitudes
and frequencies adjoining the point of common connection (PCC). This research examines a solar-photovoltaic (PV)
coordinated unified power quality controller (UPQC) to enhance the PQ in the distribution grid network. This paper
proposes a UPQC control method relied on the Adaptive Leakage Least Mean Square (AL_LMS) algorithm. This
approach uses both the Adaptive Neural Fuzzy Inference System (ANFIS) controller and Proportional-Integral (PI)
controller to compare between the serial and shunted active converter switching of UPQC. The technique switches
the UPQC's shunt and series voltage source converters (VSCs) to obtain a reference signal through iterative weight
update. Through iterative weight update, the method obtains a reference signal by switching the shunted and serial
VSCs of the UPQC. Therefore, the PQ disturbances such as load voltage drop and ripple and harmonic distortion of
the grid’s current are abridged after the control technology is applied to the solar PV-UPQC. Conferring to the
simulation results, the proposed ANFIS control AL_LMS algorithm stands the utmost operative in improving the PQ
after applied to the solar-PV sourced UPQC, while complying with the IEEE-519 standards. The work was done in
using MATLAB/Simulation software.
Keywords: Renewable Energy Source, Power Quality, UPQC, Voltage Source Converter, ANFIS Control

1. Introduction voltage sag & swells, harmonics and flickering can cause
disturbances, damage delicate equipment and reduce
As traditional energy sources are increasingly the effectiveness of power systems. Therefore, the use
depleted and have a undesirable impact on the of PQ enhancement methods (PQE) has become
environment, the usage of renewable energy sources increasingly popular as a way to reduce over voltages
(RES) for instance solar-photovoltaic (PV) systems and and power outages. Various tools and procedures have
also wind power systems for electricity generation is been proposed and employed to progress the stability
becoming increasingly popular. Researchers in and performance of power systems and to solve PQ
academia and industry are now strongly focusing on the problems [4]. Innovative technological options for
network interactions of these RES to improve system improving voltage regulation, as well as stability and
performance. RES are more popular than traditional instantaneous apparent control of dynamic power and
energy sources due to their cost-effectiveness, reactive power by the nominal frequency, such as
environmental friendliness, and ease of use [1]. On the FACTS (Flexible AC Transmission System) devices, are
other hand, a large number of control equipment, presented in [5]. Other technologies use unique power
statistics processing equipment, RES, and subtle power devices that offer a range of options including parallel
electronic equipment (PEE) have been added to the and also series recompense of active power and
electricity distribution network in recent decades, reactive power towards refining PQ, controlling voltage
resulting in a sharp increase in electricity demand, which and performing at substantial and harmonic frequencies
is accompanying with a substantial increase in PQ in both steady-state and active situations [6]. One
problems in the distribution network [2]. PQ is crucial for possible solution to the voltage and current PQ problems
ensuring the proper operational of connected devices of distribution systems is the UPQC, a dedicated power
and the overall resilience of decentralized systems. device that can be applied to the PCC distribution
Therefore, power and voltage quality are critical. The network. Through voltage regulation and mitigation of
authors in [3] described how PQ complications such as frequent PQ complications such as voltage sags, swell,
Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 264
Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025
harmonics, etc., its main intention is to guarantee technique provides better concert in both steady-state in
continuous improvement of PQ. It has a common DC bus addition dynamic scenarios, and the weighted parameter
and consists of two parts: Series (-SeAPF) and Shunt (- shift does not go out of bounds, which makes the
ShAPF) active power filters. The SeAPF compensates proposed AFIS controller a better choice than the
for voltage correlated anomalies, including sags and traditional PI controller. In this way, the reference signal
swells, whereas the ShAPF controls the DC bus voltage for switching the VSC is estimated quickly and
and dispels current interrelated problems, for e.g., accurately to alleviate the PQ problem, in so doing
reactive power (Q) demand and harmonics [7]. enlightening the concert of the UPQC provided by the
solar PV.
According to [8], RES, exclusively solar-
photovoltaic (PV) systems, are increasingly integrated
into the grid, bringing many benefits to sustainable 2. Model System Description
development and low carbon emissions. Quite a lot of
rewards could be achieved by combining solar-PV Figure 1 schematically shows the grid
systems to a UPQC. Initially, solar-PV systems can connection method of the proposed system paradigm,
function by way of clean RES, lowering the demand for which combines UPQC with solar photovoltaic system.
conventional fossil fuel power generation and A 3-phase series injection transformer (Tse) and a
encouraging a more sustainable supply of electricity. coupling inductor (Lse) are used to connect the UPQC
Secondly, according to the authors [9, 10], photovoltaic series converter elements in sequence through the
systems can dynamically support active power, allowing supply grid network. The shunt-VSC element is
the UPQC to stabilize the grid and regulate the voltage connected in competition with the load point through the
underneath diverse load conditions. In order to switch interfaced inductor (Lsh), and the two compensators are
the shunt compensator or series compensator, an connected through a capacitor (Cdc). The solar
appropriate UPQC control structure must be conditioned photovoltaic module is connected to the DC capacitor
to engender the source current signal and the load through a reversing diode. The high frequency voltage
reference voltage signal. Some studies have used components instigated by VSC switching can be
traditional time-domain methods, such as the removed by adding ripple filters (Rr and Cr). A diode
symmetrical current component theory (ISCT), the bridge rectifier (3-Φ nonlinear load) near the system is
synchronous reference frame method (SRF or d-q-0), connected for PQ problem analysis.
the reactive power theory (IRPT or p-q) and the sliding
mode control [11-14]. Due to the low efficiency of the 3. Methodologies for UPQC
low-pass filter, these usual algorithms developed based
on these theories do not work well when the load is In this case, the VSC of the UPQC is shunted,
unbalanced. Several frequency-domain techniques maintaining the DC bus voltage at a significant level
have been implemented using fast Fourier transform and while reducing the harmonics of the mains source
wavelet transform [15], although as shown in [16], these current. It also helps the PCC in providing the required
techniques require higher computational efficiency, use reactive power. Connecting the VSC in series reduces
more memory and have slower elastic response. As the harmonics, dips and ripples of the load voltage. As
mentioned in [17], control strategies based on explained in the next section, an appropriate UPQC
computational software techniques (like Neural control structure essential be conditioned to provide the
Networks) remain also castoff to classify and distinguish load reference voltage signal besides the source current
PQ problems. Built on the Leaky LMS adaptive filter signal towards switching the serial and shunted
algorithm, research [18] proposed a method to reduce compensators, correspondingly.
the THD of solar photovoltaic systems. By using the
proposed filter, the harmonic content of the input signal
3.1. Adaptive Leaky LMS (AL_LMS) Method
is significantly reduced and sent to the load. Therefore,
the combination of modern power electronic equipment The AL_LMS algorithm [20, 21] stands a
and control methods provides long-term help in solving modified standard LMS approach for adaptive filtering
PQ problems [19]. applications. The goal of adaptive filtering is to
determine the ideal set of filter constants or weights that
Here, the focal aim is to solve the PQ problem
minimalize the odd differences between the expected
caused by 3-Φ nonlinear loads in the integration of the
and the filter’s outputs. At this time, a leakage factor is
proposed solar-PV power supply UPQC hooked on the
added toward the AL_LMS algorithm to control the filter
grid system. To approximate the references of voltage
coefficient update rate, thus balancing stability and
and current when the series and shunt compensators
tracking speed [22]. The following equations (1) and (2)
work together, a control method created on the AL_LMS
explain the mathematical equations related to AL_LMS
algorithm and by distinctive PI and AFIS controllers is
and the aforementioned estimator [23].
provided toward deriving the reference signals used to
jointly drive the UPQC parallel and series VSCs. The W(t + 1) = [1 − µλ(t)]. W(t) + µE(t). X(t) (1)
adaptive leakage constraint features of the AL_LMS
Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 265
Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025

Figure 1. Illustration of a suggested model for a solar-photovoltaic-fed UPQC system

Y(t) = W T (t). X(t) (2) R(t) = [β ∗ R(t − 1)] + [(1 − β) ∗ E(t) ∗ E(t − 1)] (8)
Where X(t): On time step ‘t’, the input Here,
vector.
R(t): E(t) & E(t-1) Autocorrelation
W(t): Weighted vector of filter quantities by
β: An exponential weighting parameter that
the time pace ‘t’
regulates averaged valued time so that 0 < β < 1
µ: Step-size or erudition rate, governing
α and ρ : Convergence governing time
the adaption rate
constraints so that 0 < α < 1 and ρ > 0.
λ(t): The leakage factor, which can differ at
Using the AL_LMS algorithm, the solar PV
each time pace ‘t’ depending on system’s
power supply of the UPQC in this study can be
circumstances
adaptively controlled to generate correct reference
E(t): The instant erroneous signal at time signals for equipment operation, as shown in Figure 2.
pace ‘t’, given as anticipated output Measurements for successfully correcting PK problems.
The above method calculates the difference amongst
‘D(t)’ minus forecast output ‘Y(t)’ i.e.
the measured and expected values of these constraints.
E(t) = D(t) − Y(t) (3) The reference signal is then adaptively modified to
remove the PQ error, which is represented by the
The total function Fc(t) is stated by
difference between these two values. This method uses
Fc (t) = E 2 (t) + [λ(t). W T (t). W(t)] (4) the erroneous signal and the learning rate (μ) towards
computing the adjustment rate to reliably modify the
Weights are updated such that they achieve a
reference signal. This program generates variable
negligible value. The equations (5) - (8) are used to get
reference signals to control the movement of the UPQC
the weight update expressions
[24]. This process is carried out through a feedback loop,
W(t + 1) = W(t)[ 1 − 2µ(t) ∗ λ(t)] + [2µ(t) ∗ E(t) ∗ X(t)] where the UPQC incessantly adjusts its control signal in
(5) response to changing electrical conditions within the
system, which includes producing suitable
λ(t + 1) = λ(t) − [2µ(t) ∗ ρ ∗ E(t) ∗ X T (t) ∗ W(t − 1)] (6)
compensation signals to resolve power quality
µ(t + 1) = [α ∗ µ(t)] + [λ(t) ∗ R2 (t)] (7) complications.
Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 266
Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025

Figure 2. Proposed AL_LMS Flow Chart

vsa vsb vsc

3.2 Proposed Schematic for Shunt-VSC vpa =
Vsp
; vpb =
Vsp
;vpc =
Vsp
(10)
The proposed controller diagram for parallel
The combination of quadrature and in-phase
VSCs is shown in Figure 3 below. In this way, a stable
unit templates of voltage are defined by below equation
sinusoidal grid/source current can be generated while
(11).
controlling the DC link’s voltage. First, measure the
instantaneous phase voltages (vsa, vsb, and vsc) of −vpb +vpc √3 1
vqa = ; vqb = vpa + (vpb − vpc ) ; vqc =
each phase at the PCC. The PCC voltage’s (Vsp) phase √3 2 2√3
√3 1
amplitude being then premeditated as shown in equation − vpa + (vpb − vqc ) (11)
2 2√3
(9).
The PV arrangement and power grid affords the
Vsp =
2 2
√ (vsa + 2
vsb + 2 )
vsc (9) required active power toward the load through the
3 control mechanism. In addition, the grid maintains a
The source voltages in-phase unit templates are stable DC link’s voltage, thereby eliminating the inherent
described by the equation (10) losses of UPQC.

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 267

Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025

Figure 3. Control Schematic Diagram of proposed Shunt-VSC

Through the determination of minimizing the R pa (t) = [β ∗ R pa (t − 1)] + [(1 − β) ∗ Epa (t) ∗ Epa (t − 1)]
error, the observed capacitor voltage at DC link (Vdc) is (18)
correlated by the expected DC link’s voltage (Vdcref)
obtained from the MPPT solar PV output to generate an The Error estimate can endure as;
error signal (Edc), which is then sent to the PI /ANFIS Epa (t) = iLa (t) − vpa (t) ∗ Wpa (t) (19)
controller. Therefore, the required loss component The load current of phase "a" is stated as iLa.
(Wloss) is given by equations (12) and (13) and written Phases "b" and "c" of the parallel VSC, as well as the
as active component weights (Vpb and Vpc), leakage
Edc (t) = Vdcref (t) + Vdc (t) (12) coefficients (λpb and λpc), step sizes (μpb and μpc),
Wloss (t + 1) = Wloss (t) + K p ∗ [Edc (t + 1) − Edc (t)] + error autocorrelation (Rpb, Rpc), and prediction errors
(Epb, Epc) can be modified at any time, just like phase
K i ∗ Edc (t + 1) (13)
"a". Equation (20) therefore gives the weighted average
The equation (14) specifies how much of active of the fundamental active components (Wpave).
current provided by a solar-photovoltaic (WPV) array. Wpa (t)+ Wpb (t)+Wpc (t)
Wpv = ∗
2 PPV
(14) Wpave = (20)
3
3 Vsp
The formula for calculating the active
Now, using the weight updates basic formula
components of the reference network’s currents (ispa,
found in equivalences (5) - (8) above, the key elements
ispb, ispc) and the total active component weight (Wsp)
of the active load current meant to phase "a" can be
is:
inscribed as;
Wsp = Wpave + Wloss − WPV (21)
Wpa (t + 1) = Wpa (t)[ 1 − 2µpa (t) ∗ λpa (t)] + [2µpa (t) ∗
Epa (t) ∗ vpa (t)] (15) ispa = Wsp ∗ vpa ; ispb = Wsp ∗ vpb ; ispc = Wsp ∗ vpc (22)

λpa (t + 1) = λpa (t) − [2µpa (t) ∗ ρ ∗ Epa (t) ∗ vpa (t) ∗ Similarly, equation (23) includes updated
Wpa (t − 1)] (16) weightiness intended to the reactive current’s
components and is also used to calculate the reactive
µpa (t + 1) = [α ∗ µpa (t)] + [λpa (t) ∗ R2pa (t)] (17) constituent of the load current utilizing the vector voltage
patterns of the quadrature elements.

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 268

Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025
isqa = Wsq ∗ vqa ; isqb = Wsq ∗ vqb ; isqc = Wsq ∗ vqc (23) Additionally, these in-phase unit templates were
utilized to calculate the quadrature current unit
Herein case, the weighted average of the templates, which can be shown as
fundamental reactive components is alike to the
weighted total reactive components (Wsq). Therefore, −upb +upc √3 ∗ upa 1
uqa = ; uqb = + (upb − upc ); uqc =
√3 2 2√3
adding the components to the above equations (22) and
√3 ∗ upa 1
(23) will give the current drawn from the final reference − + (upb − upc ) (27)
2 2√3
source as shown:
The weighted load’s voltage in-phase
i∗ sa = ispa + isqa ; i∗ sb = ispb + isqb ; i∗ sc = ispc + isqc (24) component (Wvpa) of the phase ‘a’ be now determined
Henceforth, to alleviate the current harmonic by means of the basic weighted update expressions,
problems, the Hysteretic Current Controller (HCC) now given by the below equations (28) - (31).
works by integrating the obtained reference currents (i*sa, Wvpa (t + 1) = Wvpa (t)[ 1 − 2µvpa (t) ∗ λvpa (t)] +
i*sb and i*sc) by means of the measurement system [2µvpa (t) ∗ Evpa (t) ∗ upa (t)] (28)
source currents (isa, isb and isc) and the resulting errors.
λvpa (t + 1) = λvpa (t) − [ 2µvpa (t) ∗ ρ ∗ Evpa (t)] ∗
3.3. Proposed Schematic for Series Converter [ upa (t) ∗ Wvpa (t − 1)] (29)

Figure 4 shows the control scheme of the VSC µvpa (t + 1) = [α ∗ µvpa (t)] + [λvpa (t) ∗ R2vpa (t)] (30)
series. The objective of the controlling is to ensure that R vpa (t) = [β ∗ R vpa (t − 1)] + [(1 − β) ∗ Evpa (t) ∗
the load voltage has the appropriate amplitude and
Evpa (t − 1)]
sinusoidal waveform. As shown below, the load
reference voltage is assessed using the proposed (31)
AL_LMS technique. 3-Φ load’s currents (iLa, iLb and iLc) Error Estimate could be computed as
be detected in order to calculate their amplitude (ILp),
Evpa (t) = Vsa (t) − [upa (t). Wvqa (t)] (32)
whose unit samples are given by equations (25) and
(26). Where, Vsa : source’s voltage of phase ‘a’. At
each point of the VSC series, the same method can be
2
ILp = √ (i2La + i2Lb + i2Lc ) (25) used to adjust the phases "b" and "c", the phase weight
3
components (Vvpb and Vvpc), the leakage factors (λvpb
iLa iLb iLc
upa = ; upb = ; upc = (26) and λvpc), the step sizes (μvpb ) and μvpc), the error
ILp ILp ILp
autocorrelations (Rvpb and Rvpc) and the prediction
errors (Evpb and Evpc). Equation (33) below is used to
find the phase components weighted average (Wvpave).

Figure 4. Control Scheme for Proposed Series-VSC

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 269
Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025
Wvpa (t)+ Wvpb (t)+ Wvpc (t) Finally, by means of relating the reference
Wvpave = = WLp (33)
3 voltage with the calculated load’s voltage, this variable is
The elements of the measured in-phase transferred towards the PWM generator in providing
reference load’s voltage are as follows: switching pulsations intended for the series VSC. In the
future, multiple UPQC components will be implemented
VLpa = WLp ∗ upa : VLpb = WLp ∗ upb : VLpc = WLp ∗ upc
to improve the PQ allegations.
(34)
Alike to this, the AL_LMS system is also
3.4 Adaptive Neuro-Fuzzy Inference System
proposed to estimate the weights of the quadrature
(ANFIS) Control
components of load’s voltage (Wvqa, Wvqb and Wvqc).
The following equation (35) represents the weighted Another neural network-based system is the
average reactance component. Adaptive Neuro-Fuzzy Inference System (ANFIS) [25],
Wvqa (t) + Wvqb (t) + Wvqc(t) combines advantages of the Fuzzy Logic (FL) and
Wvqave = (35) Neural Networks (NN). Control systems that incorporate
3
redundant mechanisms and neural networks are often
Now consider the loss component (Wqr) in this
used to increase their effectiveness and efficiency. The
case which is written as equation (36) to regularize the
weight of the neural network ranges from -1 to 1 and is
load voltage.
accustomed for modifying the constraints of the Fuzzy
Wqr (t + 1) = Wqr (t) + K pq ∗ [Eac (t + 1) − Eac (t)] + Inference System (FIS), as shown in Figure 4. A
K iq ∗ Edc (t + 1) (36) standard fuzzy rubrics set was produced by means of
the FIS using real voltage and current constraints like
By comparing the measured load voltages (VLa, input variables and compensatory voltage and current
VLb and VLc) with the nominal load’s voltage (VL*), an signals like output variables, as shown in Table 1. The
signaling error is engendered and sent to the PI/ANFIS interpretation of these factors by the system is crucial
controller. The following equation shows the ability of the since incorrect tuning may result in either overfit or
regulator to ultimately maintain the desired amplitude of underfit data.
the load voltage.
ANFIS model parameters can be varied using a
2 2 2 2 hybrid approach combining Least Squares (LS) with
VLm = √ (VLa + VLb + VLc ) (37)
3
Mean Square Error (MSE) methods. Reducing training
Eac (n) = VL∗ (n) − VLm (n) (38) error is the goal of the training process [26]. During the
The equation (39) for the total quadrature training process, the NN and FIS parameters are
weighted voltage component (W Lq) associated with the adjusted to minimize the difference amongst the actual
load is represented by compensation voltage and current and the predicted
compensation voltage and current. A set of validation
WLq = Wvqave + Wqr (39) rules is used to assess the precision of the trained AFIS
The entire quadrature component of the load’s algorithm in approximating compensation voltage and
reference voltage is described by the following equation current requirements under different operating
(40). conditions. After training and verification, the ANFIS
VLqa = WLq ∗ uqa ; VLqb = WLq ∗ uqb ; VLqc = WLq ∗ uqc system is incorporated into the control system to
establish the necessary relationships between ANFIS
(40)
outputs (such as control signals) and AFIS inputs (such
Subsequently, the substantial load’s voltages
as errors and error changes) to integrate PI controller
references for each in the 3-phase as given:
and replace with ANFIS controller as shown in Figures 5
∗ ∗ ∗
VLa = VLpa + VLqa ; VLb = VLpb + VLqb ; VLc = VLpc + VLqc and 6.
(41)

Table 1. Fuzzy Rules Table Set

e ce NB NS Z PS PB
NB NB NB Z NS Z
NS NB NS NS Z PS
Z NS NS Z PS PS
PS NS Z PS PS PB
PB Z PS PS PB PB

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 270

Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025

Figure 5. ANFIS Model Structure

Figure 5. Series-VSC ANFIS Control Strategy

Figure 6. Strategy in Shunt-VSC by ANFIS Control

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 271
Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025
4. Results and Discussion 4.1 Series VSC’s Performance Analysis for
Voltage Sag and Swell conditions
As shown in Figure 7, the solar photovoltaic
compatible UPQC arrangement is modelled in 4.1.1 with Custom PI Controller
MATLAB/Simulink software, and at that time the
Using the AL_LMS algorithm based on a
proposed AL_LMS algorithm is used in combination with
traditional PI controller, the figure below shows the
traditional PI and AFIS controller algorithms.
performance of the UPQC series VSC under voltage
The following table 2 lists the parameters that drop and ripple voltage. By connecting VSC UPQC in
were used to simulate the system that is being series, the voltage drop of each unit at t=0.1 seconds is
discussed. smoothed by 0.5, and the voltage jump of each unit at
t=0.3 seconds is smoothed by 1.2 to obtain the rated
load voltage.

Table 2. System Parameters

Parameters Values
Supply Grid Voltage (Vs) 415V (L-L)
Coupling inductors (Lse and Lsh) 1.5mH
DC-link Capacitor (Cdc) 8000μF
DC-link Voltage (Vdc) 700 V
Gains of Shunt VSC KP = 2 & KI = 0.1
Gains of Series VSC KP = 4 & KI = 800
Solar-PV MPPT (voltage and current) 700V, 60A
Non-linear load 15KW, 7.2KVA

Figure 7. MATLAB/Simulink diagram of a projected Solar-PV integrated UPQC System

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 272

Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025

Figure 8. PI Controller based UPQC Compensation to Voltage Sag and Swell Conditioning

Figure 9. %THD of Load’s voltage via custom PI control

Figure 9 shows that using a typical PI controller More efficient recompense of voltage sag and
mounted on UPQC, the load voltage %THD is 2.41%. In swell is achieved with expected ANFIS driven AL_LMS
addition, as shown below, inspections are performed process by the enhanced %THD performance.
through the ANFIS controller to improve power quality. Performance analysis of parallel VSCs under nonlinear
load conditions shows a measured load voltage %THD
of 0.94%, compared to %THD of 2.41% by a custom PI
4.1.2 With ANFIS controlled AL_LMS algorithm
controller.
Using the proposed AL_LMS model for ANFIS
control, the performance components of the UPQC
series under voltage drop and swell conditions are
4.2 Evaluation of Shunt-VSC Performance to
shown in the figure 10. The UPQC Series VSC improves Non-linear Load Circumstances
voltage to rated load voltage, resulting to a 0.5 P.U. The following description illustrates the
voltage drop at t=0.1s and a 1.2 P.U. voltage increase at effectiveness of parallel VSCs in the existence of non-
t=0.3s. linear loads.

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 273

Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025

Figure 10. Voltage Sag and Swell conditioning via ANFIS controller

Figure 11. Proposed ANFIS control AL_LMS model showing THD percentage of load voltage
This control technology involves the grid and given away in Figure 14, the parallel VSC UPQC can
solar photovoltaic arrays working together to meet the also compensate for the nonlinearity caused through the
power demand for load generation and maintain high DC system source current, maintaining a persistent DC bus
voltages. The results of the ANFIS and PI control voltage at t=0.04 s despite the fact generating a
techniques are shown in the figure below. sinusoidal grid current.
ANFIS controller achieves 0.98% THD in
4.2.1 Use of custom PI controlled AL_LMS algorithm grid/source current, while typical PI controller achieves
3.49%, as shown in Figure 15. In Table 3, both
The line/source current is affected by harmonics
traditional PI and ANFIS controlled based on AL_LMS
generated by the nonlinear load before t=0.04 s, as
method is presented for %THD of supply/grid current
shown in Figure 12, resulting in a sinusoidal line current.
due to nonlinear loads and also infers load’s voltage
As shown in Figure 13 below, the harmonic equation
THD due to sag and swell. By comparing AL_LMS
shows that the resulting current source has a THD of
algorithm based on proposed ANFIS controller with PI
3.49%.
controller, PQ is improved.

4.2.2 Use of ANFIS control-based AL_LMS algorithm 4.3 Assessment of Power flows in the system
By comparing the proposed ANFIS control with The following figures 16 and 17 illustrates the
the AL_LMS algorithm by PI control, the differences in active and reactive power flows in a system meeting the
source/grid and load’s currents are further minimized. As load demand by both the Solar-PV system and Grid
supply system.
Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 274
Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025

Figure 12. Shunt VSC performance toward non-linear loads using a custom PI controller

Figure 13. Source’s current %THD with custom PI control

Table 3. %THD assessment of proposed UPQC controller methods

% THD with Proposed AL_LMS Algorithm
Parameters Types of Controller
With PI Controller [22] With Fuzzy Controller [23] With ANFIS Controller
Load Voltage
(For Sag and Swell) 2.41% 2.29% 0.94%

Grid/Source Current
(For Non-Linearities) 3.49% 3.38% 0.98%

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 275

Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025

Figure 14. Presentation of ANFIS controlled Shunt-VSC toward Non-linear loads

Figure 15. Proposed ANFIS control for source’s current THD

From the figure 16, it is observed that the Solar- It can be seen from figure 17 that the required
PV power is meeting the required load demand while the reactive power is met by the UPQC resulting in the lower
remaining power is fed to the main grid, resulting in the burden on the main grid system.
effective utilization of energy resource.

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 276

Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025

Figure 16. Active power flows of the system

Figure 17. Reactive power flows of the system

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 277

Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025
5. Conclusion and Future Scope [7] R. Satish, B. Pydi, S. Balamurali, S.R. Salkuti,
A.Y. Abdelaziz, S. Feleke (2023) A
Herein the study, an adaptive control method Comprehensive Power Quality Mitigation Tool.
relied on the AL_LMS algorithm thru the ANFIS Challenges and Mitigation Techniques,
controller is anticipated to solve the power quality Singapore: Springer Nature Singapore.
problem of solar photovoltaic power supply UPQC in [Link]
decentralized networks. The system performance can [8] K. Jha, A.G. Shaik, A Comprehensive Review of
be evaluated by comparing the %THD of the ANFIS Power Quality Mitigation in the Scenario of Solar
controller with that of a conventional PI controller under PV Integration into Utility Grid, E-Prime -
the conditions of nonlinear load, voltage sags and Advances in Electrical Engineering, Electronics
swells. In these cases, better results are achieved by and Energy, 3, (2023) 100103.
using the AL_LMS algorithm, which is based on the [Link]
ANFIS controller and generates appropriate reference
[9] N.K. Agarwal, A. Saxena, A. Prakash, A. Singh,
currents for parallel and series VSC of UPQC. The load A. Srivastava, A. Baluni, Review on Unified
voltage is efficaciously upheld at the nominal rated Power Quality Conditioner to Mitigate Power
voltage even during sags and surges. In addition, the Quality Problems, IEEE, 2nd Global Conference
THD percentage of the source’s current remains for Advancement in Technology (2nd GCAT)
significantly reduced and becomes sinusoidal for the (2021),IEEE, Bangalore, India.
nonlinear load current. Moreover, the proposed method [Link]
helps in utilizing solar-PV energy to meet the load active 86
power demand, in so doing tumbling the power grid’s [10] P. Ray, [Link], S.K. Dash, Power Quality
burden. With the intention of enlightening the power Enhancement and Power Flow Analysis of a PV
quality of the system, UPQC can adopt a multi-level Integrated UPQC System in a Distribution
converter topology and might incorporate optimization Network. IEEE Transactions on Industry
techniques to improve the regulation of the PI controller Applications, 58(1), (2022) 201–211.
in the further research. [Link]
[11] A.M. Diab, S. Bozhko, [Link], M. Rashed, G.
References Buticchi, Zhuang Xu, S. S. Yeoh, Chris Gerada,
G. Michael, Fast and Simple Tuning Rules of
[1] M.K.G. Deshmukh, M. Sameeroddin, D. Abdul, Synchronous Reference Frame Proportional -
M. Abdul Sattar, Renewable energy in the 21st Integral Current Controller. IEEE Access, 9,
Century: A Review. Materials Today: (2021) 22156–70.
Proceedings, 80, (2021) 1756–1759. [Link]
[Link] [12] S. Ahmad, S. Mekhilef, H. Mokhlis, (2020) DQ-
[2] S. Alam, F.S. Al-Ismail, A. Salem, M.A. Abido, axis Synchronous Reference Frame based P-Q
High-level Penetration of Renewable Energy Control of Grid Connected AC Microgrid, IEEE
Sources into Grid Utility: Challenges and International Conference on Computing, Power
Solutions. IEEE Access, 8, (2020) 190277– and Communication Technologies (GUCON),
190299. IEEE, Greater Noida, India.
[Link] [Link]
1080
[3] M.H.J. Bollen (2000) Understanding Power
Quality Problems: Voltage Sags and [13] T. Geetha, S. Chitra. Coordination Control of
Interruptions. Wiley-IEEE Press, New York. Microgrid Using Sliding Mode Controller.
International Research Journal of
[4] L.B. Chilakapati, T.G. Manohar, Microgrid Power Multidisciplinary Technovation, 1(3), (2019) 57-
Quality Enhancement with Adaptive Control 64. [Link]
Strategies: A Literature Survey. Przegląd
Elektrotechniczny, 1(3), (2024) 115–119. [14] M. Bajaj, A.K. Singh, Grid Integrated Renewable
[Link] DG Systems: A Review of Power Quality
Challenges and State‐of‐the‐art Mitigation
[5] N.G. Hingorani, L. Gyugyi, (2000) Understanding Techniques. International Journal of Energy
FACTS: Concepts and Technology of Flexible AC Research, 44(1), (2020), 26–69.
Transmission Systems. Wiley-IEEE Press. New [Link]
York.
[15] V. Thiriburasundari, R. Shanmugalakshmi.
[6] M. Ebeed, S. Kamel, F. Jurado, S.H.E. Abdel Identification of Power System Disturbances
Aleem, A.F. Zobaa (2023), Modeling Combined Using Wavelet Transform. International
Shunt/Series FACTS in Power Flow Solutions: A Research Journal of Multidisciplinary
Comprehensive Review. Modernization of Technovation, 1(3), (2019) 65-72.
Electric Power Systems. Springer International [Link]
Publishing. [Link]
18996-8_1 [16] N. Singh, M. A. Ansari, M. Tripathy, V.P. Singh,
Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 278
Vol 7 Iss 1 Year 2025 Chilakapati Lenin Babu & T. Gowri Manohar /2025
Feature Extraction and Classification Techniques Technologies, (3rd ICEPE) (2021), Shillong,
for Power Quality Disturbances in Distributed Meghalaya, India.
Generation: A Review. IETE Journal of [Link]
Research, 69(6), (2023) 3836–385. 514
[Link]
[25] U.K. Renduchintala, C. Pang, K.M. Tatikonda, L.
[17] R. Abdulkader, H.M.A. Ghanimi, [Link], Yang, ANFIS‐Fuzzy Logic based UPQC in
[Link], W. El-Shafai, M.M. Fouda, M.H. Aly, Interconnected Microgrid Distribution Systems:
D. Swaminathan, S. Sengan, Soft computing in Modeling, Simulation and Implementation. The
Smart Grid with Decentralized Generation and Journal of Engineering, 2021(1), (2021) 6–18.
Renewable Energy Storage System Planning. [Link]
Energies. 16(6), (2023) 2655.
[26] H. Kenjrawy, C. Makdisie, I. Houssamo, N.
[Link]
Mohammed, New Modulation Technique in
[18] N. Devena, G. Gengi, B.K. Vijayammal, K.S. Smart Grid Interfaced Multilevel UPQC-PV
Reghuraman, D. Gunaselvam, J. Balaraman, Controlled via Fuzzy Logic Controller.
Improvement of Power Quality Using Variable Electronics, 11(6), (2022) 919.
Leaky Least Mean Square Controller with [Link]
Adaptive Shunt Active Filter for Non-Linear
Loads. AIP Conference Proceedings, Annapolis,
2857(1), (2023) 020046. Authors Contribution Statement
[Link] Chilakapati Lenin Babu: Original draft, writing and
[19] R. Jadeja, A. Ved, T. Trivedi, G. Khanduja (2020) formatting, methodology and conceptualization,
Control of Power Electronic Converters in AC software modeling of the system and T. Gowri Manohar:
Microgrid. in N. Mahdavi Tabatabaei, E. Kabalci, Research Supervision, Methodology, Reviewing and
and N. Bizon (eds), Microgrid Architectures, Suggestions.
Control and Protection Methods, Springer
International Publishing. Funding
[Link] The authors declare that no funds, grants or any other
[20] L. Alhafadhi, J. Teh, (2022). Power Quality support were received during the preparation of this
Enhancement in Stand-Alone PV System Using manuscript.
Leaky LMS Adaptive Algorithm. International
Conference on Robotics, Vision, Signal Competing Interests
Processing and Power Applications, Springer, The authors declare that there are no conflicts of interest
Singapore, 449–454 [Link] regarding the publication of this manuscript.
981-16-8129-5_69
[21] A. Bag, B. Subudhi, P.K. Ray, An Adaptive Data Availability
Variable Leaky Least Mean Square Control The data supporting the findings of this study can be
Scheme for Grid Integration of A PV System, obtained from the corresponding author upon
IEEE Transactions on Sustainable Energy, 11(3), reasonable request.
(2020) 1508–1515.
[Link] Has this article screened for similarity?
[22] L.B. Chilakapati, T.G. Manohar, Control Yes
Strategies for Enhancing Power Quality with
Unified Power Quality Conditioner in a Solar-PV
About the License
Integrated Utility System. International Journal of
Experimental Research and Review, 35, (2023) © The Author(s) 2025. The text of this article is open
1–15. access and licensed under a Creative Commons
[Link] Attribution 4.0 International License.
[23] C. Lenin Babu, T. Gowri Manohar, An Improved
Power Quality in a Renewable Energy-based
Microgrid System Using Adaptive Hybrid UPQC
Control Strategy. International Journal of
Experimental Research and Review, 36, (2023)
217–231.
[Link]
[24] P. Ray, P.K. Ray, Design and Control of PV-
UPQC using Variable Leaky LMS based
Algorithm for Power Quality Enhancement. IEEE,
3rd International Conference on Energy, Power
and Environment: Towards Clean Energy

Int. Res. J. Multidiscip. Technovation, 7(1) (2025) 264-279 | 279