Parameter Identification of Solar Photovoltaic Systems Using an Augmented Subtraction-Average-Based Optimizer

Abstract

Solar photovoltaic system parameter identification is crucial for effective performance management, design, and modeling of solar panel systems. This work presents the Subtraction-Average-Based Algorithm (SABA), a unique, enhanced evolutionary approach for solving optimization problems. The conventional SABA works by subtracting the mean of searching solutions from the position of those in the population in the area of search. In order to increase the search capabilities, this work proposes an Augmented SABA (ASABA) that incorporates a method of collaborative learning based on the best solution. In accordance with manufacturing, the suggested ASABA is used to effectively estimate Photovoltaic (PV) characteristics for two distinct solar PV modules, RTC France and Kyocera KC200GT PV modules. Through the adoption of the ASABA approach, the simulation findings improve the electrical characteristics of PV systems. The suggested ASABA outperforms the regular SABA in terms of efficiency and effectiveness. For the R.T.C France PV system, the suggested ASABA approach outperforms the traditional SABA technique by 90.1% and 87.8 for the single- and double-diode models, respectively. Also, for the Kyocera KC200GT PV systems, the suggested ASABA approach outperforms the traditional SABA technique by 99.1% and 99.6 for the single- and double-diode models, respectively. Furthermore, the suggested ASABA method is quantitatively superior to different current optimization algorithms.

1. Introduction

Because of the ongoing growth in power consumption and the swiftly diminishing supply of petroleum, coal and natural gas, scientists all over the world are forced to seek new sources of energy [1] through the use of small-scale distributed generation of electricity and energy storage facilities [2]. Their massive incorporation into the electrical network would raise the regulatory burden and have an impact on the main grid’s security operation [3]. A microgrid may operate independently or in conjunction with the main system. It is often made up of dispersed power sources, fuel cells, energy storage systems, and regulated loads [4]. In this regard, the mathematical representation of the microgrid components is of great importance to be modeled. Fuel cells were mathematically represented where SDBT [5], LST [6] and GTT [7] were adopted to adequately estimate the PEMFC parameters including the maximum current density, membrane thickness, membrane preparation parameter, parametric coefficient, cell connection resistance and empirical coefficient, respectively. Batteries were mathematically represented [8] where the generic equivalent circuitry modeling is built with a series of RC component parts designated by RC-model [9]. This comparable RC group is regarded as a relaxation portion to alleviate the difficulty of extracting the battery characteristics, whereas the state space model for battery dynamics was represented [10]. Also, TLSM were mathematically represented in [11], where a MGWT was developed to determine the shape design of the TLSM to maximize the operating force and minimize the flux saturation. The developed MGWT incorporated an outside archive with a predetermined size that is integrated for storing and retrieving Pareto optimal solutions.

Furthermore, the identification of solar photovoltaic system parameters is critical for effective performance management, design, and modeling of solar panel systems [12]. Significant progress has been achieved in understanding how the properties of PV systems operate by applying mathematical models of the PV technology during the past few decades. A large number of estimates match the reported current-voltage measurements collected from PV devices under all operating circumstances, and resemble the behavior of genuine PV cells [13]. In general, diode-based similar circuit models are employed. In light of this, the PVSD [14] and PVDD [15] varieties are particularly popular [3]. The current voltage (I-V) characteristics are similarly described by the three-diode design. Identical configurations involve the identification of five, seven, and nine parameter values [16].

Due to the nonlinear nature, non-convex form and multi-parameter properties of the PV framework make it difficult to determine its unrecognized variables, and many researchers are addressing this. In recent years, metaheuristic optimizing algorithms have been widely used to address the parameter estimation of PV systems, and they have demonstrated better performance than established methodologies. Because it is gradient-free, simple to create, and performs well, this approach is perfect for dealing with difficult optimization issues. These approaches of problem resolution include SDBT [17,18], PSO [19], sunflower optimization technique [20], MPA [21], social networking searching method [22], butterfly optimizing approach [23], bonobo optimizer [24], HBM [25], etc. In [26], an IGA with a non-uniform mutation was presented for retrieving features from PV models in a trustworthy, exact, and time-efficient way. On PVSD and PVDD systems, the IGA approach has been employed in this work. Nevertheless, it was not used in the three-diode PV configuration. GA [27] is strongly reliant on the initial PV parameter selection. If the starting settings are set improperly, the parameters generated from the subsequent changes will most likely decline to a locally optimum solution. As a result of the inaccuracy of the PV model derived parameters, the operational performance of the PV system is incorrect. In [28], a solar three-diode design model based on different PV systems has been employed to simulate an artificially generated hummingbird algorithm. The PV parameter optimizing method was handled in [29] due to the seagull optimizer’s extended exploration of the parameterized space throughout the migration cycle. Introducing a non-linear regulating component to adjust both local and global searches minimized the deterioration, but it was significantly limited to increasing the reliability and quality of the generated variables. In [23], insufficient study of butterfly optimization typically excludes promising locations in the parameter’s region, making accurate PV system knowledge difficult to obtain. In accordance with the aforementioned investigation, present-day extraction methods for tackling the parameters identifying the PV framework have multiple flaws, such as simply dropping in a local optimum, insufficient converge, and inappropriate exploration [29]. As a result, building a new effective approach for correctly representing the PV equivalent model and determining the undetermined parameters is hard and time-consuming.

Each year, hundreds of algorithms for optimization are proposed, adopted, and used in the domains of power system engineering, such as hydro generation scheduling [30], improving the power system operation [31], wind power uncertainty in distribution systems [32] optimal reactive-power dispatch [33], static Var compensator devices applications [34] and Efficiency Improvement of Distribution Systems [32]. Recently, a technique named SABA [35] has been presented where its fundamental premise is to update population members’ locations in the search space by deducting the average of searcher agents. This technique is beneficial since it can be easily applied to engineering applications and has minimal parameters that need to be changed. The results of the SABA were contrasted with more modern approaches and other existing techniques considering several benchmark models. The main contributions of this paper can be summarized as follows:

• An ASABA is provided in this study to improve searching capabilities.
• The suggested ASABA expands on the traditional SABA by incorporating a cooperative learning technique based on the leader response.
• The present research also develops the suggested ASABA for optimally obtaining the PV characteristics.
• Because of the unique properties of the ASABA approach, it emphasizes enhancing the electrical features of different types of comparable circuits of solar power systems that take into account the combination of PVSD and PVDD.
• The suggested ASABA is used for a variety of purposes in PV technologies, including two commercial RTC France PV panels and two Kyocera KC200GT PV modules. In addition, the suggested ASABA is statistically compared to previously documented optimization procedures in the literature.

2. ASABA for Parameter Identification of Solar Photovoltaic Systems

2.1. Proposed ASABA

The typical SABA technique works by subtracting the mean of the seeker participants’ position in the search space. The dimension that corresponds to the width of the area of search is equal to the number of control variables under consideration. They are established by the locations of the algorithm’s seeking solutions. As a consequence, each exploring option or person is mathematically portrayed as a vector that contains data about the control variables, as shown in Equation (1) [36]. Also, the range refers to the different permissible intervals of the control variables that may be stated as in Equation (2).

Every single one of the investigated solutions appears to be a feasible answer to the evaluated element. The perfect and poorest solutions are defined by the best and worst values generated for the goal function, based on the assessed values for the objective function. The SABA’s approach was founded on mathematical factors such as mean amounts, changes in searching indicative locations, and the sign of the variance between two measurable that are objective. The SABA’s approach for determining the arithmetic average is entirely unique since it depends upon a specific functional the “v-subtraction operator.” As a result, every solution with a vector component in the SABA community has been modified in accordance with Equation (3). In addition, Equation (4) represents the mathematical representation of the subtraction procedure for the two seeking options (Sa_i and Sa_k) derived from the SABA community.

The associated value of the objective is calculated and judged once every option vector is updated. Following that, according to (5), the newly constructed solution substitutes the old solution if the newly developed option has a higher objective rating. Furthermore, relying on the leader’s solution, the suggested ASABA contains a collaborative learning technique. In this sense, the average value of the v-subtraction operation determines the alteration in location of every seeking option (Sa_i) inside the area of search. The exploratory qualities are substantial and effective when employing this design. On the other hand, the exploitation seeking features need to be improved by additionally assisting the localized looking process surrounding the most promising location. To do this, an ASABA variant incorporates a method of collaboration to offer information for learning from the best solution vector, as shown in Equation (3) [35].

$$S{a}_i=LL+Range\times rand(1,Dim)\hspace{1em}i=1:Ns$$

(1)

$$Rang{e}_i=U{L}_i-L{L}_i\hspace{1em}i=1:Dim$$

(2)

$$\begin{array}{l}S{a}_{i,new}=\left\{\begin{array}{c}S{a}_i+\overrightarrow{z_i}\times \frac{1}{Ns}\left({\displaystyle \sum _{k=1}^{Ns}\left(S{a}_i-{}_{\upsilon }\mathit{Sa}_k\right)}\right)\\ S{a}_{BEST}+\overrightarrow{w_i}\times \left(S{a}_{R1}-S{a}_{R2}\right)\end{array}\right.\begin{array}{c}\hspace{1em}if CR\le rand(0,1)\\ Else\end{array}\\ \hfill i=1,2,\dots \dots \dots \dots ,Ns\end{array}$$

(3)

$$S{a}_i-{}_{\upsilon }\mathit{Sa}_k=sign\left(Fi{t}_i(S{a}_i)-Fi{t}_k(S{a}_k)\right)\left(S{a}_i-\overrightarrow{\upsilon }\odot S{a}_k\right)$$

(4)

$$S{a}_i=\left\{\begin{array}{c}S{a}_{i,new}\\ S{a}_i\end{array}\right.\begin{array}{c}if Fi{t}_{i,new}(S{a}_{i,new})\le Fi{t}_i(S{a}_i)\\ Else\end{array}$$

(5)

A choice probability (CR) of 50% is maintained to create an equilibrium between the exploratory features and the improved exploit qualities as stated in Equation (3). Figure 1 depicts the critical phases of the proposed ASABA. As displayed in the first step in the flowchart in Figure 1, there are three control parameters of the suggested ASABA algorithm which are: the choice probability), the population size (Ns) and the maximum number of iterations.

2.2. Electrical Circuit Representation of Solar Photovoltaic Systems

To show the I-V properties of solar PV frames, various electrically equivalent designs have been created. In practice, the PVDD and PVSD designs constitute the most commonly utilized analogous circuits. The PVDD is widely adopted to represent the attributes of solar cells. Figure 2 depicts the PVDD related circuitry. The key components in this model are two diodes, two resistors, and a current source. Kirchoff’s Current Law is used to mathematically depict the PVDD model’s load-current formula [37];

$$I=I_{ph}-I_{S1}\left[\exp \left(\frac{I{R}_S}{{\eta }_1{V}_{th}}+\frac{V}{{\eta }_1{V}_{th}}\right)-1\right]-I_{S2}\left[\exp \left(\frac{I{R}_S}{{\eta }_2{V}_{th}}+\frac{V}{{\eta }_2{V}_{th}}\right)-1\right]-\frac{I{R}_S}{R_{sh}}+\frac{V}{R_{sh}}$$

(6)

The thermal voltage of the PV system and can possibly be computed, as stated by the following [38]:

$$V_{th}=\frac{K_{B}T}{q_c}$$

(7)

This model’s seven unnamed parameters—I_Ph, I_S1, I_S2, R_Sh, R_S, η_1, and η₂—must be computed utilizing the PV panels’ I-V data.

On the other side, the PVSD is a much simpler design which is widely adopted to represent the attributes of solar cells as well. In this model, only one diode is considered, as displayed in Figure 3. Equation (8) depicts the output current using Kirchhoff Current Law, which can be calculated by employing the Shockley diode, as below [37];

$$I=I_{ph}-I_{S1}\left[\exp \left(\frac{I{R}_S}{{\eta }_1{V}_{th}}+\frac{V}{{\eta }_1{V}_{th}}\right)-1\right]-\frac{I{R}_S}{R_{sh}}+\frac{V}{R_{sh}}$$

(8)

This model’s five unnamed parameters—I_Ph, I_S1, R_Sh, R_S, and η₁—must be computed utilizing the PV panels’ I-V data.

The statistical evaluation in this paper was carried out using the subsequent formula, which was centered on the RMSE [38,39,40]:

$$RMSE=\sqrt{\frac{1}{PN}\left({\displaystyle \sum _{K=1}^{PN}{(I_{cal}^K(V_{\mathit{exp}}^K,x)-I_{\mathit{exp}}^K)}^2}\right)}$$

(9)

Figure 2. Circuit design of PVDD model [39].

Figure 2. Circuit design of PVDD model [39].

Figure 3. Circuit design of PVSD model [39].

Figure 3. Circuit design of PVSD model [39].

3. Results and Discussion

This section explores R.T.C France and the Kyocera KC200GT PV systems employing the proposed ASABA strategy as opposed to the traditional SABA strategy. The first investigation focuses on the R.T.C France system that works at 1000 w/m² of solar radiation and 33 °C of temperature. The second instance investigated is on the Kyocera KC200GT PV system, which is made up of 54 multi-crystalline cell series which are connected [39]. Furthermore, the proposed ASABA strategy is examined and used to parameter extracting difficulties across different PVSD and PVDD systems, for contrast to recent optimisation strategies previously described in the literature. The prescribed ASABA and regular SABA procedures are used with an identical amount of iterations of 1000 and solutions of 200.

3.1. First Case Study

First, the suggested ASABA and traditional SABA strategies are utilized with the PVSD framework to minimize the RMSE function goal for the RTC France silicon PV system.

Table 1. Obtained parameters by the proposed ASABA and standard SABA for RTC France considering PVSD design.

Parameters	Standard SABA	Proposed ASAB
IPh (A)	0.77096543	0.76077553
Rsh (Ω)	0.03535773	0.03637709
RS (Ω)	87.20679827	53.71852224
IS1 (A)	0.00000100	0.00000032
η1	1.60420121	1.48118359
RMSE	0.00993415	0.00098602

highlights their calculated parameter values.

From Table 1, the suggested ASABA produces a substantially lower RMSE value of 9.8602 × 10⁻⁴ than the traditional SABA approach, which obtains an RMSE value of 0.00993415. As a result, the suggested ASABA approach outperforms the traditional SABA technique by 90.1%. Also, Figure 4 displays the convergence patterns corresponding to the suggested ASABA and traditional SABA strategies. As shown, the suggested ASABA technique continues progress in reducing the RMSE objective, compared the traditional SABA method. As demonstrated in the zooming part of SABA in Figure 4c, the SABA technique becomes stuck very early at the 185th iteration at a local minimum RMSE value.

Additionally, for the PVSD, Figure 5 displays the obtained RMSE targets for the separate runs regarding both the suggested ASABA and traditional SABA strategies. As shown, great superiority is declared due to the suggested ASABA. It shows a very high improvement compared to the traditional SABA strategy with 90.1%, 90.1%, 93.6%, 94.2%, 94.7%, 94.9%, 94.9%, 94.9%, 95.0% and 92.9% for the different separate executions.

Table 2. Comparisons of the proposed ASABA and standard SABA for RTC France considering PVSD design.

Algorithms	RMSE	Algorithms	RMSE	Algorithms	RMSE
Proposed ASABA	0.00098602	TLBO [46]	9.8733 × 10−4	HSBA [41]	9.95146 × 10−4
Standard SABA	0.00993415	JAYA optimizer [47]	9.8946 × 10−4	ABC [44]	10 × 10−4
IGA [26]	9.8618 × 10−4	Improved DE [46]	9.89 × 10−4	Chaotic PSO [46]	13.8607 × 10−4
GWO [43]	75.011 × 10−4	Comprehensive Learning PSO [42]	9.9633 × 10−4	BBO with mutation [45]	9.8634 × 10−4
CSA [46]	9.91184 × 10−4

compares the proposed ASABA to several optimization techniques for the PVSD framework that have been published in the scientific community. It contrasts the RMSE score of the proposed ASABA approach to that of existing optimization techniques, such as HSBA [41], comprehensive learning PSO [42], GWO [43], ABC [44], BBO with mutation [45], TLBO [46], JAYA optimizer [47] and IGA [26]. As shown, the proposed ASABA surpasses other alternatives in locating the lowest RMSE.

Second, the suggested ASABA and traditional SABA strategies are utilized with the PVDD framework to minimize the RMSE function goal for the RTC France silicon PV system.

Table 3. Obtained parameters by the proposed ASABA and standard SABA for RTC France considering PVDD design.

Parameter	Standard SABA	Proposed ASABA
IPh (A)	0.765657	0.760775
Rsh (Ω)	0.025307	0.036528
RS (Ω)	75.35299	54.59428
IS1 (A)	9.14 × 10−7	3.53 × 10−7
η1	1.668141	1.999956
IS2 (A)	8.51 × 10−7	2.76 × 10−7
H2	1.684558	1.467738
RMSE	0.008073	0.0009835

highlights their calculated parameter values. Table 3 demonstrates that the proposed ASABA yields a significantly lower RMSE value of 0.0009835 than the standard SABA method, which yields an RMSE value of 0.008073. As a consequence, the proposed ASABA method surpasses the standard SABA method by 87.8%. Figure 6 also depicts the convergence patterns associated with the proposed ASABA and existing SABA techniques. As demonstrated, the proposed ASABA strategy continues to make progress in lowering the RMSE target when compared to the existing SABA method. The SABA approach becomes stuck relatively early at the 93rd iteration at a local minimum RMSE value, as seen in the zooming section of SABA in Figure 6c.

Furthermore, Figure 7 for the PVDD shows the achieved RMSE objectives for the various runs using both the recommended ASABA and standard SABA techniques. As demonstrated, considerable superiority is stated as a result of the recommended ASABA. It outperforms the classic SABA method by 87.8%, 93.9%, 95.1%, 95.7%, 96.2%, 96.6%, 96.9%, 97.0%, 97.3%, and 97.4% for the various individual executions.

Table 4. Comparisons of the proposed ASABA and standard SABA for RTC France considering PVDD design.

Algorithms	RMSE	Algorithms	RMSE	Algorithms	RMSE
Proposed ASABA	0.0009835	ABC [54]	1.28482 × 10−3	Flower pollination algorithm [49]	1.934336 × 10−3
Standard SABA	0.008073	Teaching–learning–based ABC [51]	1.50482 × 10−3	Cat swarm algorithm [48]	1.22 × 10−3
TLBO [55]	1.52057 × 10−3	Generalized oppositional TLBO [50]	4.43212 × 10−3	Comprehensive learning PSO [52]	1.3991 × 10−3
Sine cosine approach [53]	9.86863 × 10−4

compares the proposed ASABA to several optimization techniques for the PVSD framework that have been published in the scientific community. It contrasts the RMSE score of the proposed ASABA approach to that of existing optimization techniques, such as cat swarm algorithm [48], flower pollination algorithm [49], generalized oppositional TLBO [50], teaching–learning–based ABC [51], comprehensive learning PSO [52], sine cosine approach [53], ABC [54] and TLBO [55]. As shown, the proposed ASABA surpasses other alternatives in locating the lowest RMSE.

Figure 8 also depicts the PVDD model’s actual and projected P-V and I-V properties. The data generated by the proposed ASABA technique generally coincided with the experimental data, suggesting that the proposed ASABA methodology efficiently extracts the needed PV parameters.

3.2. Second Case Study

First, the suggested ASABA and traditional SABA strategies are utilized with the PVSD framework to minimize the RMSE function goal for the KC200GT PV system.

Table 5. Obtained parameters by the proposed ASABA and SABA for KC200GT PV module considering PVSD design.

Applied Technique	Standard SABA	Proposed ASABA
IPh (A)	8.209911	8.216767
Rsh (Ω)	0.00388	0.004826
RS (Ω)	83.12138	6.280213
IS1 (A)	8.64 × 10−7	2.62 × 10−8
η1	1.476826	1.212905
RMSE	0.070075	0.000637

highlights their calculated parameter values. Also, Figure 9 displays the convergence patterns corresponding to the suggested ASABA and traditional SABA strategies. Both Table 5 and Figure 9 shows that the suggested ASABA produces a substantially lower RMSE value of 0.000637 than the traditional SABA approach, which obtains an RMSE value of 0.070075. As a result, the suggested ASABA approach outperforms the traditional SABA technique by 99.1%. From Figure 9, it is observed that the standard SABA algorithm converges, approximately, 93 iterations, but the proposed ASABA algorithm takes 1000 iterations for convergence. Despite that, the standard SABA algorithm is always stuck in a local minimum, but the proposed ASABA has the ability to achieve lower objective scores.

Moreover, for the PVSD, Figure 10 displays the obtained RMSE targets for the separate runs regarding both the suggested ASABA and traditional SABA strategies. As shown, great superiority is declared due to the suggested ASABA. It shows a very high improvement compared to the traditional SABA strategy with 99.1%, 99.6%, 99.7%, 99.7%, 99.7%, 99.7%, 99.7%, 99.8%, 99.8% and 95.4% for the different separate executions.

Table 6. Comparisons of the proposed ASABA and standard SABA for KC200GT considering PVSD design.

Applied Technique	Standard SABA	Proposed ASABA	EVO [39]
RMSE	0.070075	0.000637	0.023069893
Applied technique	FPA [39]	HTS [39]	GO [39]
RMSE	0.011225773	0.01799763	0.008515347

compares the proposed ASABA to several optimization techniques for the PVSD framework that have been published in the scientific community. It contrasts the RMSE score of the proposed ASABA approach to that of existing optimization techniques, such as GO [39], FPA [39], and HTS algorithm [39] and EVO [39]. As shown, the proposed ASABA surpasses other alternatives in locating the lowest RMSE.

Second, the proposed ASABA and classic SABA techniques are combined with the PVDD framework to achieve the KC200GT system’s RMSE function aim.

Table 7. Obtained parameters by the proposed ASABA and standard SABA for KC200GT considering PVDD design.

Applied Technique	Standard SABA	Proposed ASABA
IPh (A)	8.16780	8.21618
Rsh (Ω)	0.00308	0.00488
RS (Ω)	100.00000	6.45712
IS1 (A)	0.00000	0.00000
η1	1.57325	1.24808
IS2 (A)	0.00000	0.00000
η2	1.54171	1.00000
RMSE	0.07577	0.00034

shows the values of their computed parameters. Figure 11 also shows the convergence trends for the recommended ASABA and standard SABA strategies. Table 7 and Figure 11 indicate that the proposed ASABA strategy yields a significantly lower RMSE value of 0.00034 than the classic SABA approach, which yields an RMSE value of 0.07577. As a consequence, the proposed ASABA method surpasses the standard SABA methodology by 99.6%.

Figure 11 shows that the conventional SABA algorithm remains constant for about 365 iterations from iteration 95 to 460. Not only that, but it also maintains constant for almost 500 iterations from iteration 460 until the end result. The suggested ASABA method, on the other hand, requires 1000 iterations to reach convergence. This observation illustrates that the current SABA algorithm is always locked in a local minimum, but the suggested ASABA can reach lower objective scores.

Moreover, for the PVDD, Figure 12 displays the obtained RMSE targets for the separate runs regarding both the suggested ASABA and traditional SABA strategies. As shown, great superiority is declared due to the suggested ASABA. It shows a very high improvement compared to the traditional SABA strategy with 99.6%, 99.3%, 99.3%, 99.3%, 99.4%, 99.3%, 99.3%, 99.2%, 98.5% and 88.5% for the different separate executions.

Table 8. Comparisons of the proposed ASABA and standard SABA for KC200GT considering PVDD design.

Applied Technique	Standard SABA	Proposed ASABA	EVO [39]
RMSE	0.07577	0.00034	0.02717656
Applied technique	FPA [39]	HTS [39]	GO [39]
RMSE	0.014006267	0.020515491	0.009049475

compares the proposed ASABA to several optimization techniques for the PVSD framework that have been published in the scientific community. It contrasts the RMSE score of the proposed ASABA approach to that of existing optimization techniques, such as GO [39], FPA [39], HTS [39] and EVO [39]. As shown, the proposed ASABA surpasses other alternatives in locating the lowest RMSE.

Figure 13 also depicts the PVDD model’s actual and projected P-V and I-V properties. The data generated by the proposed ASABA technique generally coincided with the experimental data, suggesting that the proposed ASABA methodology efficiently extracts the needed PV parameters.

Based on the above simulation results, the proposed ASABA algorithm is superior to SABA due to different features:

• However, the standard SABA algorithm converges approx. 100 iterations and the proposed ASABA algorithm takes 1000 iteration for convergence; the standard SABA algorithm is always stuck in a local minimum but the proposed ASABA has the ability to achieve lower objective scores.
• The proposed ASABA provides superior robustness over SABA with more than a 90% improvement based on the obtained RMSE for the different separate runs.

4. Conclusions

This work introduces the SABA as an innovative enhanced evolutionary technique for solving optimization problems. The classic SABA works by subtracting the solution-finding method from the population’s position in the search zone. This study describes an ASABA that incorporates a method of collaborative learning based on the best solution to increase search capabilities. According to the manufacturer, the suggested ASABA is used to efficiently estimate PV characteristics for two independent solar PV modules, RTC France and Kyocera KC200GT PV modules. Using the ASABA approach, the simulation findings improve the electrical characteristics of PV systems. The suggested ASABA outperforms the regular SABA in terms of efficiency and efficacy. Furthermore, the suggested ASABA approach statistically outperforms several current optimization techniques. Compared to four recently presented algorithms of GO, FPA, HTS and EVO, the proposed ASABA surpasses them in finding the lowest RMSE for both models of the PVSD and PVDD KC200GT PV module. Added to that, the proposed ASABA technique can efficiently extract the PV parameters where the extracted data are generally coincided with the experimental data. Given the great effectiveness of the proposed ASABA technique in the aforementioned PV parameter estimation concern, it is advised that the provided method be examined for adequacy in future attempts to tackle the power system operation [56,57,58] and control [59,60,61] optimization frameworks.