An Experimental Study of Drift Caused by Partial Shading Using a Modified DC-(P&O) Technique for a Stand-Alone PV System

Abstract

There is tremendous potential in solar energy to meet future electricity demands. Partial shading (PS) and drift are two major problems that must be addressed simultaneously to achieve the maximum power point (MPP) of a stand-alone PV system, which are discussed in this paper. Both of these factors contribute to the voltage drop due to heavy steady-state oscillation. The partial shading and drift problem are associated with severe rapid changes in the insolation. A modified drift-control perturbation and observation DC-(P&O) approach was investigated using a low-cost programmable hardware solution, i.e., the ARM Cortex M4 32-bit Microcontroller (MC) (STM32F407VGT6), with efficient embedded programming and Waijung block sets for real-time solutions. The experimental setup was accomplished on a 40-watt solar panel. It was found that the proposed method had a significant impact on drift control during abrupt changes in current and voltage caused by shading effects, with the controller conversion efficiency of 80.39% and 94.48% with percentage absolute errors of 7.3 and 7.2 for cases with and without PS and drift, respectively.

1. Introduction

Traditional electricity generation relies mostly on coal-based power plants that produce carbon dioxide (CO₂), which, in turn, contributes to air pollution in the environment. Researchers have sought numerous alternative sources of energy. Renewable energy (RE) is being used as an alternate source of electrical power generation to address this issue [1]. India has made remarkable efforts in the last seven years to adopt energy through public centric energy policy. India’s goals have risen from 175 GW of installed renewable energy capacity by 2022 to 500 GW of non-fossil fuel-based capacity by 2030 [2]. In the previous 5 years, Asian countries’ non-fossil fuel electrical capacity has increased by 1.6 times to 158 GW, accounting for nearly 40% of total installed power output [3]. Photovoltaic-based power is becoming increasingly popular among renewable energy choices due to its plentiful availability and inexhaustibility [4]. Furthermore, conventional fuels still account for 84 percent of global energy use, and this figure is steadily decreasing. According to the literature, solar energy supersedes all other green energy sources, since solar technologies have advanced with new features. Partial shading and weather conditions are the main cause of drifting MPPT points. The impact of these factors is responsible for certain critical points due to heavy shading on solar panels [5]. The efficiency of a PV array system is determined by operating circumstances, as well as the quality of the solar cells and array design. The array can be different types of series and parallel combinations of solar cells [6]. Solar energy contributions to electric power networks have been rapidly growing in the last couple of years.

Proper planning is required before installing the panels. Due to the constraints of the area, it is inevitable for buildings and other nearby structures to obstruct the sunlight from reaching the panels [7]. The difference in electrical characteristics between unshaded and shaded panels can lead to significant losses. The difference in the illumination intensities between the two different regions affects the power output. This issue is a crucial factor that can lead to a reduction in the power output due to the operation in different operating conditions [8]. Besides the location of the PV array, other factors such as the level of shading and the position of the panel are also taken into account to reduce the power output [9].

Partial shading conditions (PSCs) are also important factors that can affect the performance of a PV system [10]. The reduction in the photon-generated current due to the PS of a PV module can be attributed to the intensity of the light. Usually, series connections of PV modules are more affected by PSCs than parallel connections. The most common type of connection scheme is the series–parallel. These connections are mainly dependent on the region’s asymmetrical or symmetrical shading pattern [11].

The DC–DC converter’s duty cycle was changed to allow the user to track the MPPT. This eliminates the issue of partial shading. To achieve the optimal performance of the DC–DC converter, the pulse signal of the device must be controlled by a power switch. Unfortunately, when the PV modules are partially shaded, their performance deteriorates [12].

The maximum power of the solar module may be tracked with P–V and I–V characteristics. The highest point of the curve indicates the maximum power point (MPP), as represented in Figure 1 and Figure 2.

1.1. Background, Related Research, and Motivation

Blaabjerg et al. [13] discussed various designs of PV cells and suggested their use based on the size of PV systems and provided detailed analyses of changes in temperature or irradiance in terms of the nonlinear characteristics of a PV array. Based on a closed-loop reference model, Blaabjerg et al. demonstrated how to build and obtain a novel real-time output characteristic of a PV array emulator. Robles-Campos et al. [14] and Shams et al. [15] focused on the impact of partial shade (PS) circumstances with the bypass diodes technique. The authors studied the existing research on bypass diode topologies and proposed a suggested model, which was thoroughly examined and validated with PSCAD/EMTDC and MATLAB.

Baba et al. [16] discussed MPP evolutionary algorithms for controllers in terms of efficiency, performance, modernity, complexity, and tracking speed. Teng et al. [17] and Obukhov et al. [18] emphasised extracting maximum power (MP) in real time, with a considerable step size of control variables. The authors also discussed the challenges of MPP, i.e., the high cost of hardware implementation and its critical points.

Valenciaga et al. [19] and Kumar et al. [20] discussed the unique MPP approach of the second-order sliding mode gradient (SOSM) server. The authors also suggested the traditional PI controller for a closed-loop control with fast-tracking dynamics. Dallago et al. [21] presented the D-MPPT technique, in which the breadboard was used to test the hardware, with a tracking efficiency is 78.8%. Endo et al. [22] provided a comparison between two hardware, i.e., between a microarchitectural simulator and ARM Cortex MC. They suggested the ARM Cortex MC has a low absolute error of nearly 7% in comparison with others.

1.2. Research Gap

In recent years, solar cell technology has proved to be promising for meeting the necessary energy requirements. Nevertheless, the world is still facing many challenges such as partial shading (PS), battery management systems (BMS), and grid compatibility [23]. In the past year, several researchers have provided a plethora of studies in the literature on the shading phenomenon, with different topologies. Past algorithms have the ability to extract maximum power during shading conditions in the presence of drift [15,24]. However, the authors found that more research is required for reducing steady-state oscillations and fast-tracking MPP. Therefore, they attempted to address this problem with experimental validation in this paper.

1.3. Contribution and Case Description

• In this research, a modified fast-tracking DC-(P&O) approach was used to facilitate the determination of MPP at a certain position;
• A 40 W PV panel consisting of 36 cells, with 4 rows each having 9 cells with a DC load, was used to prepare an experimental setup on which the PS effect of 30% was observed;
• In this study, it was found that modified DC-(P&O) can control and trace MPP during darkened conditions with the help of slope $\left(\frac{dP}{dV}\right)$ position on PV characteristics. If $\frac{dP}{dV}>0$, then MPP will be on the left, and otherwise on the right;
• The proposed study provides a solution to drift when rapid changes occur in the environment.

1.4. Modelling of Single-Diode Model with a Boost Convertor Using Modified DC-(P&O)

The equivalent diagram with a single diode of a PV model [25] is represented in Figure 3, according to which PV is connected with a diode (D) and shunt resistance (R_sh) in parallel and resistance in series (R_s).

Figure 4a,b illustrate, respectively, the ON and OFF operations of a boost converter using an IGBT switch. When the switch is ON, the current passes through the inductor and the switch. When the switch is OFF, the current passes through the load.

When the switch is always ON, the duty ratio can be calculated using Equation (1) as follows:

$$\frac{V_o}{V_{in}}=\frac{1}{1-D}$$

(1)

where D is the duty ratio, $V_o$ is the output voltage, and $V_{in}$ is the input voltage.

Load power is calculated by Equation (2).

$${}_{P_L = V_L \ast I_L = \frac{V_L{}^2}{R_L}}$$

(2)

where $P_L$ is the load power, $V_L$ is the load voltage, $I_L$ is the load current, $R_L$ is the load resistance, and $R_L$ can be calculated by Equations (3)–(5) as follows:

$$R_L=\frac{V_L}{I_L}$$

(3)

$$R_L=\frac{\frac{V_{PV}}{(1-D)}}{I_{PV}(1-D)}$$

(4)

$$R_L=R_{PV}\frac{1}{{(1-D)}^2}$$

(5)

The efficiency of solar cells can be defined by using Equation (6).

$${\eta }_{pv}=\frac{V_L{I}_L}{V_{pv}{I}_{pv}}=\left(\frac{1}{{(1-D)}^2}\right)\frac{R_{pv}}{R_L}$$

(6)

where ${\eta }_{pv}$ is the efficiency of PV cell, $V_{pv}{I}_{pv}$ are the voltage and current of PV cell, $R_{pv}$ is input impedance, which is varied based on the change in duty cycle.

We know that the condition of maximum power extraction is $R_{pv}$ = $R_L$.

Then, PV current corresponds to maximum power and can be written using Equation (7) as follows:

$$I_{pv}=\frac{V_{pv}}{{(1-D)}^2{R}_L{\eta }_{pv}}$$

(7)

Photovoltaic current I_pv can be calculated by applying KCL in Figure 3, and obtained using Equation (8) as follows:

$$I_{pv}=I_{sc}-I_d-I_{sh}$$

where the value of $I_d$ and V_ph can be expressed by

$$I_d=I_0\left[\exp (\frac{V_{pv}}{n{V}_T})-1\right],V_{ph}=I_{sh}\times R_{sh}$$

where $I_d$ is the current across the diode, $I_0$ is the saturation current, and V_ph is the voltage across the diode.

$$I_{pv}=I_{sc}-I_o\left[\exp (\frac{V_{pv}}{n{V}_T})-1\right]-I_{sh}$$

(8)

where $I_{sc}$ is short-circuit current (SCC), $n$ is the ideality factor,$V_T$ is the terminal voltage, and $I_{sh}$ is the shunt current, which can be written using Equation (9).

$$I_{sh} \mathrm{shunt} \mathrm{current}=\frac{(V_{pv}+I_{pv}{R}_s)}{R_{sh}}$$

(9)

By putting the I_sh from Equation (9) into Equation (8), I_pv can be represented by Equation (10) as follows:

$$I_{pv}=I_{sc}-I_o\left[\exp (\frac{V_{pv}}{n{V}_T}-1)\right]-\frac{(V_{pv}+I_{pv}{R}_s)}{R_{sh}}$$

(10)

From Equations (7) and (10), we obtain

$$\frac{V_{pv}}{{(1-D)}^2{R}_L{\eta }_{pv}}=I_{sc}-I_o\left[\exp (\frac{V_{pv}}{n{V}_T}-1)\right]-\frac{(V_{pv}+I_{pv}{R}_s)}{R_{sh}}$$

(11)

Putting the value of $I_{pv}$ in Equation (11), we obtain Equation (12) as follows:

$$V_{pv}\left(\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}+\frac{1}{R_{sh}}+\frac{R_s}{R_{sh}{(1-D)}^2{R}_L{\eta }_{pv}}\right)=I_{sc}-I_o\left[\exp (\frac{V_{pv}}{n{V}_T}-1)\right]$$

(12)

$$\frac{V_{pv}}{{(1-D)}^2{R}_L{\eta }_{pv}}+\frac{(V_{pv}+\frac{V_{pv}}{{(1-D)}^2{R}_L{\eta }_{pv}}{R}_s)}{R_{sh}}=I_{sc}-I_o\left[\exp (\frac{V_{pv}}{n{V}_T}-1)\right]$$

(13)

$$V_{pv}\left(\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}+\frac{1}{R_{sh}}+\frac{R_s}{R_{sh}{(1-D)}^2{R}_L{\eta }_{pv}}\right)=I_{sc}-I_o\left[\exp (\frac{V_{pv}}{n{V}_T}-1)\right]$$

(14)

By considering equally, $\exp (\frac{V_{pv}}{n{V}_T})$ to $1+\frac{V_{pv}}{n{V}_T}$, we have

$$V_{pv}\left(\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}+\frac{1}{R_{sh}}+\frac{R_s}{R_{sh}{(1-D)}^2{R}_L{\eta }_{pv}}\right)=I_{sc}-I_o\left[1+\frac{V_{pv}}{n{V}_T}-1\right]$$

(15)

$$I_{sc}=V_{pv}\left(\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}+\frac{1}{R_{sh}}+\frac{R_s}{R_{sh}{(1-D)}^2{R}_L{\eta }_{pv}}+\frac{I_o}{n{V}_T}\right)$$

(16)

Solving Equation (16) with total irradiance Q leads to

$${\left.V_{pv}\right|}_Q=\frac{{\left.I_{sc}\right|}_Q}{\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}\left(1+\frac{R_s}{R_{sh}}\right)+\frac{I_o}{n{V}_T}+\frac{1}{R_{sh}}}$$

(17)

Similarly, using Equations (7) and (17), we obtain

$${\left.I_{pv}\right|}_Q=\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}\frac{{\left.I_{sc}\right|}_Q}{\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}\left(1+\frac{R_s}{R_{sh}}\right)+\frac{I_o}{n{V}_T}+\frac{1}{R_{sh}}}$$

(18)

After taking the derivative of Equation (17),

$$\frac{d{V}_{pv}}{dQ}=\frac{(I_{sc,n}+{\alpha }_1{\Delta }_T)\frac{1}{Q_n}+{\alpha }_1\frac{Q}{Q_n}\frac{dT}{dQ}}{\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}\left(1+\frac{R_s}{R_{sh}}\right)+\frac{I_o}{n{V}_T}+\frac{1}{R_{sh}}}\succ 0$$

(19)

After taking the derivative of Equation (18),

$$\frac{d{I}_{pv}}{dQ}=\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}\frac{(I_{sc,n}+{\alpha }_1{\Delta }_T)\frac{1}{Q_n}+{\alpha }_1\frac{Q}{Q_n}\frac{dT}{dQ}}{\frac{1}{{(1-D)}^2{R}_L{\eta }_{pv}}\left(1+\frac{R_s}{R_{sh}}\right)+\frac{I_o}{n{V}_T}+\frac{1}{R_{sh}}}\succ 0$$

(20)

The factor $d{I}_{ }$ will enhance the MPP due to the rapid changes in the insolation as per the algorithm introduced in Equation (20).

2. Block Diagram of Workflow

The proposed workflow is shown in Figure 5 using a block diagram to facilitate its understanding for the readers.

3. Modified DC-(P&O) Mitigation Technique

As per findings of the literature [19,20,21,22,23], it is difficult to track the MPP without drifting because of the shedding effect (SE) and abrupt changes in the surroundings. The proposed modified DC-(P&O) technique allows MPP to be tracked without drifting, which was not feasible in traditional P&O.

A three-level operational approach is used to obtain the appropriate results [26], which are dependent on the two critical parameters of perturbation time (T_a) and step size (ΔD). The flowchart of this approach is shown in Figure 6.

In Figure 7, the slope $\left(\frac{dP}{dV}\right)$ is +ve on the left and -ve on the right of MPP in traditional P&O procedures [27]. The duty cycle is performed to track MPP with the help of +ve and −ve slopes.

Figure 6. The flowchart of the modified DC-(P&O) technique.

Figure 6. The flowchart of the modified DC-(P&O) technique.

Figure 7. Three-level operation of P&O technique with slope changes $\left(\frac{dP}{dV}\right)$.

Figure 7. Three-level operation of P&O technique with slope changes $\left(\frac{dP}{dV}\right)$.

3.1. Steady-State Operation (SSO) of Traditional P&O Procedure

Figure 7 shows a three-level operation of the traditional P&O approach, in which the point is relocated from 1 to 2, and the choice must be made based on the values of dP and dV at P2. According to the traditional equations $dP=\left(P_2-P_1\right)>0$ and $dV=\left(V_2-V_1\right)>0$, the method reduces the duty cycle and advances the operating point 2 to P3. At P3, As $dP=\left(P_3-P_2\right)<0$ and $dV=\left(V_3-V_2\right)>0$, the algorithm raises the duty cycle, and the operating point returns to P2. At P2, $dP=\left(P_3-P_2\right)>0$ and $V=\left(V_2-V_1\right)<0$ approaches zero; hence, the algorithm raises the duty cycle [28], and the operating point returns to P1. At P1, As $dP=\left(P_3-P_2\right)<0$ and $dV=\left(V_3-V_2\right)<0$ at P3, the algorithm reduces the duty cycle, and the operating point returns to P2.

3.2. Steady-State Operation (SSO) of Proposed Modified DC-(P&O) Procedure without Drift

It is very difficult to obtain MPP with precise I–V characteristics represented in Figure 8a, according to which insolation increases or decreases rapidly due to cloudy days or PS conditions [29]. This problem is also responsible for drift, which can occur from any of the three steady-state points (SSP) depicted in Figure 8b, in which the interval of perturbation time is represented by (Ta). The power also increases or decreases when $dP>0$ due to the disturbance in insolation. From Equations (19) and (20), it is clearly shown that the variance in the temperature is related to the change in insolation proportionally, i.e., $\left(\frac{dT}{dQ}>0\right)$. To validate the condition $\left(\frac{d{V}_{pv}}{dQ}>0\right)$ and $\left(\frac{d{I}_{pv}}{dQ}>0\right)$, the numerators and denominators should be positive quantity for a larger value of $R_{sh}$ and minimum value for $R_s$. By introducing $d{I}_{ }$ into the algorithm, as shown in Figure 6, the drift problem may be addressed. Points are shifted (from point 3 to point 4) by increasing the insolation. In Figure 8, $\left(dI=I_4\left(\alpha T_a\right)-I_2\left(\left(\alpha -1\right)T_a\right)\right)>0$ is represented in Figure 8a, and $\left(dP=P_4\left(\alpha T_a\right)-P_2\left(\left(\alpha -1\right)T_a\right)\right)>0$ and $\left(dV=V_4\left(\alpha T_a\right)-V_2\left(\left(\alpha -1\right)T_a\right)\right)>0$ are at the same time represented in Figure 8b. The role of $dI$ is more important for moving from point 4 to point 7, as shown in Figure 8b. For $dI>0$, duty increases and helps to uplift the MPP for SSP.

4. Simulation Model for Partial Shading

The PV panel used in this study has 36 PV cells in the series, and 3 bypass diodes are connected, as shown in Figure 9. During a shaded situation, the bypass diode is enabled, allowing current to bypass through the shaded cells and reducing current losses. The three-bypass diode partitions the PV module into three vertical regions, each of which can hold up to 25 cells for a 36-cell module, as shown in Figure 10. The specifications of the solar panel are described in Table 1.

Table 1. W solar panels.

S.No	Parameter Description	Values at 1000 W/m2 on STC
1	Number of cells (NS)	36
2	Maximum power (PMAX)	40 Watt
3	Voltage at PMAX	17.18 W
4	Current at IMAX	2.33 A
5	Open-circuit voltage (VOC)	21.37 V
6	Short-circuit current (ISC)	2.5 A
7	Temperature coefficient of VOC	−0.2775 V
8	Temperature coefficient of ISC	0.0051 A
9	$\mathrm{Slope} \left(\frac{dV}{dI}\right)$ at VOC	−0.68 V
10	Brand energy gap (EG)	1.12
11	Ideality factor (n)	1.2
12	Shunt resistance (RSH)	25 Ω
13	Series resistance (RS)	0.0065 Ω
14	Ambient temperature (TA)	25 °C

Table 1. W solar panels.

S.No	Parameter Description	Values at 1000 W/m2 on STC
1	Number of cells (NS)	36
2	Maximum power (PMAX)	40 Watt
3	Voltage at PMAX	17.18 W
4	Current at IMAX	2.33 A
5	Open-circuit voltage (VOC)	21.37 V
6	Short-circuit current (ISC)	2.5 A
7	Temperature coefficient of VOC	−0.2775 V
8	Temperature coefficient of ISC	0.0051 A
9	$\mathrm{Slope} \left(\frac{dV}{dI}\right)$ at VOC	−0.68 V
10	Brand energy gap (EG)	1.12
11	Ideality factor (n)	1.2
12	Shunt resistance (RSH)	25 Ω
13	Series resistance (RS)	0.0065 Ω
14	Ambient temperature (TA)	25 °C

Figure 9. A 40-watt solar panel.

Figure 9. A 40-watt solar panel.

Figure 10. Shading-based simulation model of 40-Watt solar panel with bypass diode.

Figure 10. Shading-based simulation model of 40-Watt solar panel with bypass diode.

The P–V and I–V characteristics of a photovoltaic string display peak points, as shown in Figure 11, which reflects uniform irradiation. That peak is known as the global peak, representing the highest output during partial conditions on different irradiance values, i.e., 270 W/m², 480 W/m², and 600 W/m². The highest peak is represented on different insolation levels, as shown in the characteristics in Figure 11.

Table 1 shows the detailed specifications of solar panels used in our simulation studies at 1000 W/m² STC. The simulation analysis can easily be understood with the results of a previous study by the authors [6]. The values are written in Table 1, which are very useful for performing simulations as well as for experimental studies.

Simulations and experimental studies were performed on a DC–DC boost converter in the proposed research by using the parameters shown in

Table 2. Parameters of DC–DC boost converter.

S.No	Name of the Parameter	Values/Specification
1	Vin	12–22 V
2	Vout	36 V
3	$f_{PWM}$	20 Khz
4	Lboost	5 mH
5	Cboost	2200 µF
6	Diode	RHRP30120
7	Relay	5 V
8	Pout	40 W
9	IGBT	1200 V, 25 A (KGT25 N120 N)

.

5. Simulation Model

Figure 12a,b show the simulation model with the proposed technique, based on power simulator (PSIM) and MATLAB, respectively. A 40-watt PV panel was used as input for the experimental studies, with the given parameters shown in Table 1 and Table 2. The simulation waveform of output power (Po) and maximum power (P_max) are shown in Figure 13a, and cell voltage (V_cell) and load voltage (V_L) at 270 W/m² are shown in Figure 13b. The output power reached the maximum power within 0.11 s, with PS and drift shown in Figure 13a. The authors performed this experiment in the lab with artificial light (halogen lamp) and modulated the light with the help of an autotransformer, at an irradiance of 270 W/m², which was measured with a lux meter.

The controller results and parameters are shown in Figure 13c and

Table 3. Steady-state parameters of boost convertor.

S.No.	Name of Parameter	Tuned Condition	Blocking Condition
1.	Rise time	$1.99 \times {10}^{-5}$	$1.4 \times {10}^{-5}$
2.	Settling time	0.00012 s	0.000148 s
3.	Overshoot	25%	32%
4.	Peak	1.25	1.32
5.	Gain margin	$6.29 \times {10}^4$	$8.68 \times {10}^4$
6.	Phase margin	43 °C	38 °C
7.	Closed-loop stability	Stable	Stable

. In terms of stability, the authors found that, based on the results of the gain and phase margins, the system was under stable conditions, i.e., $2\sim 10 \mathrm{dB}$ and $30\sim 60°$, respectively, for the closed-loop system, where overshoots, peak time, rise time, and settling time were under tuned range with the convertor [30].

6. STM 32 ARM Cortex M4 Microcontroller

In the proposed study, an STM 32 ARM Cortex Microcontroller was used instead of an MPPT controller, due to its specialty in affordability, fast-tracking, quick processing, and simple handling features. This MC provides a favourable solution for inefficient conversion in stand-alone PV systems. In this experiment, an STM 32 ARM Cortex M4 Microcontroller was used. The performed test was set up for easy and quick modifications using the DC-(P&O) approach to locate the real data and to trace MPP rapidly for a stand-alone PV system. Rapid control improved its performance with the drift control effect and PS effect. This study proposes low-cost, easier coding to make handling controls and support configurable hardware solutions. The 32-bit ARM Cortex M4 (STM32F407VGT6) was utilised with the Waijung block set in MATLAB [31].

Hardware Descriptive View

The model contains three 12-bit A/D converters, with a total of 24 channels, as well as two 12-bit D/A converters and a DMA controller with FIFOs for GPU. On the MC board, 16-bit and 32-bit timers were used, with a total of 17 timers (12 for 16-bit and 2 for 32-bit timers, with 168 MHz). There are 15 comports on the proposed device. The suggested MC has tolerances of 5 V, which are better than conventional MC [32]. The main blocks are as follows:

(a)

Target setup: It consists of sampling time, MC unit and compiler, etc. For a given MC unit, this remains the same for all developed models. Here, we had to configure only the target device, as shown in Figure 14a;

(b)

Controller operation: The pulse and sine generator block, and other Simulink blocks can be used to create control logic blocks or code for any customizable applications. These blocks define a model’s control logic and are used to generate an application’s control signals/firing pulses. Control blocks for closed-loop applications can be designed using ADC and DAC, as shown in Figure 14a,b.

Figure 14. MPPT tracking under partial shading: (a) detailed block set and parameter values of 32-bit ARM Cortex M4 (STM32F407VGT6) Microcontroller; (b) implementation of DC-(P&O) method in MATLAB with code generated in (a).

Figure 14. MPPT tracking under partial shading: (a) detailed block set and parameter values of 32-bit ARM Cortex M4 (STM32F407VGT6) Microcontroller; (b) implementation of DC-(P&O) method in MATLAB with code generated in (a).

7. Experimental Setup

The proposed algorithm/simulation was validated by setting up the experiment as shown in Figure 15. Artificial light (halogen lamp) was used to illuminate the panel. Using the PV panel’s rating, the PV was linked to the DC–DC boost converter. Sensors sense input voltage and current. The ARM controller was utilised to record the gate pulse using gate terminals or a digital storage oscilloscope (DSO) to sense the output unit. With the aid of proportional integrated derivative (PID), a direct memory access (DMA) controller offered extra benefits, with FIFO support, for this user-friendly procedure. A data logger (DL) was used to capture all data.

8. Results and Discussion

The proposed algorithm was tested for the various parameters which are affected by PS and drift, with rapid changes in atmospheric conditions. The experiment tested for a step-change in insolation level from 270 W/m² to 480 W/m² at 0.11 s. Perturbation time (Ta) and perturbation size (ΔD) were 20.001 milliseconds and 0.01, or 1%, respectively. In Figure 16a, the three-level operation of the algorithm is clearly shown at 63% duty cycle. The closed-loop action is shown in Figure 16b, according to which a Po of 45 V was obtained with the input of 23 V in the boost converter. The parameters obtained were as follows: V_pv was 17.03 V, I_pv was 0.38 A, V_out was 45 V, and I_out was 0.13 A. Real-time P–V and I–V curves are shown in Figure 16c,d without PS and drift, and with PS and drift, respectively. The hardware structure is shown in Figure 16e.

The comparison of hardware and simulation results of the modified DC-(P&O) approach for conditions without PS and drift, and with PS and drift is shown in

Table 4. Experimental validation and comparison of hardware and simulation results at 270 W/m².

S.No	Name of the Parameters	Hardware Results of Modified DC-(P&O) without PS and Drift	Simulation Results of Modified DC-(P&O) without PS and Drift	Hardware Results of Modified (DC-P&O) with PS and Drift	Simulation Results of Modified DC-(P&O) with PS and Drift
1	Voc	18.50 V	18.50 V	18.50 V	18.50 V
2	Isc	0.60 A	0.60 A	0.60 A	0.60 A
3	Vmax	15.56 V	16.62 V	13.47 V	14.2 V
4	Imax	0.486 A	0.43 A	0.47 A	0.53 A
5	Duty ratio	63%	63%	63%	63%
6	Vout	45 V	39.4 V	34.45 V	35.53 V
7	Iout	0.123 A	0.149 A	0.119 A	0.123 A
8	MPP reaching time	0.11 s	0.11 s	0.11 s	0.117 s
9	Pout	5.5 W	5.9 W	4.1 W	4.4 W
10	Real time (Ppv)	5.8 W	6.22 W	5.1 W	5.24 W
11	Controller efficiency	94.48%	97.39%	80.39%	83.9%
12	% Error calculation for modified DC-(P&O) without PS and drift				7.2
13	% Error calculation for modified DC-(P&O) with PS and drift				7.3

. This table comprises many important parameters such as Pout, controller efficiency, and duty ratio (D). The results of the proposed algorithm on hardware and simulation environment were tested with and without PS and drift conditions. The findings of the proposed approach showed that the used microcontroller had percentage absolute errors of 7.2 and 7.3, with 94.48% efficiency without PS and drift condition, and 80.39% efficiency with PS and drift condition, respectively.

9. Conclusions

On the basis of the experimental study and analysis presented in this paper, the following conclusions were drawn:

• The proposed technique performed well in the condition of having PS and drifts, with the power output of 4.1 W at 30% PS and 270 W/m² insolation levels;
• The used converter yielded a 34.45 V output at 30% PS and 270 W/m² insolation level, which is more than expected;
• This experimental study confirmed that the proposed technique performed well under rapid changes in the environment;
• The proposed technique facilitated the tracking of the MPP at 63% within 0.11 s;
• The controller used for the experimental study worked on open-loop (OL) and closed-loop (CL) controls to obtain real-time P–V and I–V curves, as shown in Figure 16c,d;
• The conversion efficiency of this controller was 94.48% without PS and drift, and 80.39% with 30% PS, both of which are more than the data mentioned in the literature survey;
• The proposed technique is better suited for low-voltage-based stand-alone PV systems;
• The proposed approach yielded a percentage absolute error within the hardware limits.