Inter national J our nal of P o wer Electr onics and Dri v e System (IJPEDS) V ol. 16, No. 3, September 2025, pp. 1711 ∼ 1720 ISSN: 2088-8694, DOI: 10.11591/ijpeds.v16.i3.pp1711-1720 ❒ 1711 Optimization of ANN-based DC v oltage contr ol using h ybrid rain optimization algorithm f or a transf ormerless high-gain boost con v erter Mohcine Byar, Abdelouahed Abounada Research T eam on Control and Ener gy Con v ersion, Department of Electrical Engineering, F aculty of Science and T echnology , Sultan Moulay Slimane Uni v ersity , Beni Mellal, Morocco Article Inf o Article history: Recei v ed Jan 14, 2025 Re vised May 13, 2025 Accepted May 25, 2025 K eyw ords: High-g ain boost con v erter Neural netw ork control Rain optimization algorithm Solar photo v oltaic systems V oltage re gulation ABSTRA CT This paper introduces an adapti v e v oltage re gulation technique for a transformerless high-g ain boost con v erter (HGBC) inte grated within standalone photo v oltaic systems. A neural netw ork controller is trained and ne-tuned using the rain optimization algorithm (R O A) to achie v e impro v ed dynamic beha vior under v ariable solar conditions. The proposed R O A-ANN frame w ork continuously updates the duty c ycle to ensure output v oltage stabi lity in real time. V alidation w as carried out using MA TLAB–OrCAD co-simulation under multiple scenarios. Comparati v e results highlight superior performanc e of the R O A-ANN controller in terms of con v er gence speed, o v e rshoot minimization, and steady-state response, outperforming con v entional PID and ANN-based methods. This is an open access article under the CC BY -SA license . Corresponding A uthor: Mohcine Byar Research T eam on Control and Ener gy Con v ersion, Department of Electrical Engineering F aculty of Science and T echnology , Sultan Moulay Slimane Uni v ersity Beni Mellal 23000, Morocco Email:mohcinebyar2013@gmail.com 1. INTR ODUCTION The increasing en vironmental and economic challenges associated with fossil fuel consumption ha v e accelerated the global shift to w ard rene w able ener gy systems [1]. Among these, solar photo v oltaic (PV) technology has g ained prominence due to its scalability , ease of deplo yment, and lo w en vironmental impact. Despite these benets, uctuations in solar irradiance introduce po wer v ariability that undermines v oltage stability and o v erall system ef cienc y [2]. T o mitig ate such issues, adv anced po wer conditioning and adapti v e control techniques are required to ensure stable operation under changing en vironmental conditions [3]. In PV systems, DC-DC boost con v erters (BCs) are commonly emplo yed to step up the inherently lo w v oltage output of PV modules to le v els appropriate for standalone or grid-connected use [4]. While traditional BCs perform adequately for moderate v oltage requirements, their ef cienc y signicantly declines at high v oltage g ain dem ands. This is primarily due to increased duty c ycles and associated switching losses, which ele v ate stress on components and reduce o v erall performance [5]. T o address these shortcomings, HGBCs ha v e been introduced based on adv anced congurations such as coupled i nductors, switched capacitor netw orks, and v olt age multiplier circuits [6], [7]. These topologies achie v e ele v ated output v oltages while operating at moderate duty c ycles, thereby minimizing component stress. Nonetheless, maintaining v oltage stability under dynamic operating conditions—particularly with J ournal homepage: http://ijpeds.iaescor e .com Evaluation Warning : The document was created with Spire.PDF for Python.
1712 ❒ ISSN: 2088-8694 uctuating irradiance—continues to pose a signicant control challenge that demands adapti v e re gulation strate gies. Proportional inte gral deri v ati v e (PID) controllers are widely utilized in po wer electronic applications due to their straightforw ard implementation and ef fecti v e performance under stable condi tions [8]. Ho we v er , their static g ain structure limits adaptability to rapidly changing inputs, often resulting in o v ershoot, prolonged settling times, and steady-state errors in nonlinear systems such as PV -po wered HGBCs [9]. Additionally , PID controllers typically require manual tuning, making them less suitable for en vironments subject to frequent or unpredictable uctuations. T o enhance control e xibility and precision, ANNs ha v e emer ged as a viable alternati v e to con v entional methods. Due to their abilit y to capture comple x nonlinear mappings between system v ariables, ANNs are well-suited for v oltage re gulation in po wer con v erter applications [10]. Prior research has demonstrated that ANN-based controllers can signicantly impro v e both transient and steady-state beha vior in HGBCs [11]. Nonetheless, their ef fe cti v eness hea vily depends on the netw ork’ s architecture and h yperparameter conguration, which directly inuence training outcomes and control accurac y . T o o v ercome these lim itations, researchers ha v e emplo yed metaheuristic algorithms to automat ically ne-tune ANN parameters [12]. Among these, rain optimization algorithm (R O A) has demonstrated strong capabilities in a v oiding local optima and achie ving rapid con v er gence during training [13]. Inspired by the natural o w of raindrops across terrain surf aces, R O A has been ef fecti v ely applied in optimizing neural netw orks for v arious engineering applications [14]. Recent studies indicate that combining R O A with ANN signicantly enhances controller performance, particularly in terms of response time and output accurac y [15]. This paper proposes an ANN control strate gy optimized using R O A to re gulate the output v oltage of a transformerless HGBC in P V -based applications. The objecti v e is to enhance t racking accurac y , minimize o v ershoot, and achie v e f aster dynamic response under v arying irradiance le v els. R O A is emplo yed to adjust the ANN’ s wei ghts and biases during the training phase, resulting in a more rob ust and adapti v e control system . The ef fecti v eness of the proposed approach is v eried through a co-simulation en vironment combining MA TLAB and OrCAD, which enables accurate v alidation of both the control logic and hardw are beha vior . Performance is e v aluated ag ainst con v entional ANN and PID controllers to demonstrate the achie v ed impro v ements in realistic PV operating scenarios. The rest of this paper is structured a s follo ws: Section 2 presents the proposed methodology , co v ering con v erter modeling and R O A-ANN controller design. Section 3 pro vides simulation results and a comparati v e e v aluation of controller performance across dif ferent scenarios. Section 4 summarizes the main ndings and outlines directions for future research. 2. METHODOLOGY This section presents the conguration of a standalone PV system emplo ying a trans formerless high-g ain boost con v erter (HGBC) re gulated by a R O A-optimized ANN controller . The control objecti v e is to maintain a stable output v oltage despite uctuations in irradiance. The R O A-ANN controller computes the duty c ycle in real time based on system feedback. The complete control structure is sho wn in Figure 1. Figure 1. Synoptic diagram of the co-simulated PV -fed HGBC system Int J Po w Elec & Dri Syst, V ol. 16, No. 3, September 2025: 1711–1720 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Po w Elec & Dri Syst ISSN: 2088-8694 ❒ 1713 2.1. Model of the transf ormerless high-gain boost con v erter High-g ain DC-DC con v erters play a crucial role in PV applications by stepping up the inherently lo w v oltage of PV sources to le v els com patible with connected loads or in v erters [16]. The transformerless HGBC considered i n this study enhances v oltage g ain while reducing circuit comple xity and po wer losses. Eliminating the need for b ulk y magnetic components such as transformers not only reduces the o v erall size and cost of the system b ut also impro v es its ef cienc y and reliability under uctuating en vironmental conditions. As illustrated in Figure 2, the HGBC topology comprises three inductors ( L 1 , L 2 , L 3 ), tw o capacitors ( C 1 , C o ), tw o diodes ( D 1 , D o ), and three switching de vices ( S 1 , S 2 , S 3 ). The con v erter is designed to operate in continuous conduction mode (CCM), which f acilitates consistent ener gy o w and stable DC output, e v en under v ariable irradiance conditions [17]. The con v erter operates in tw o distinct phas es. During the rst phase, when switches S 1 , S 2 , and S 3 are turned ON, i n duct or L 1 is ener gized directly from the PV source, while inductors L 2 and L 3 are char ged via capacitor C 1 . At this stage, diodes D 1 and D o are re v erse-biased, ef fecti v ely i solating the load and maintaining the v oltage le v el across C o . In the second phase, once the switches are turned OFF , inductor L 1 releases its stored ener gy to capacitor C 1 through diode D 1 , while inductors L 2 and L 3 dischar ge through diode D o to supply the output stage. This switching pattern enables high v oltage g ain at moderate duty c ycles, which helps minimize conduction losses and e xtends the operational lifespan of the con v erter . Applying the v olt-second balance principle, the v oltage g ain of the con v erter is gi v en by (1). V o = (1 + D ) (1 − D ) 2 V in (1) Where D represents the duty c ycle. This relationship demonstrates that the con v erter can attain substantial v oltage amplication without requiring e xcessi v ely high duty ratios, thereby reducing component stress and enhancing system reliability . Figure 2. T opology of the proposed transformerless HGBC 2.2. Structur e of the R O A-optimized ANN contr oller T o maintain stable output v oltage from the transformerless HGBC under uctuating en vironme ntal conditions, an ANN-based adapti v e control approach is adopted. While con v entional PID controllers of fer simplicity and ease of deplo yment, their limited adaptability in nonlinear and time-v arying systems mak es them less ef fecti v e under rapidly shifting irradiance conditions [18], [19]. In contrast, ANNs possess the capability to approximate comple x nonli near relationships and adjust dynamically to input v ariations, making them highly suitable for v oltage re gulation in PV applications. The implemented ANN is structured as a feedforw ard netw ork designed to generate the optimal duty c ycle D for the HGBC. It processes four input signals: reference v oltage ( V r ef ), output v oltage ( V out ), PV v oltage ( V pv ), and the v oltage error dened as e = V r ef − V out . These inputs enable the netw ork to e v aluate the system’ s real-time condition and mak e adapti v e control decisions accordingly [20]. Sigmoid acti v ation functions are emplo yed in the hidden layers due to their ability to capture nonlinear relationships [21], while the output layer produces a normalized v alue of D that directly controls the switching of the con v erter . The training dataset w as obtained through MA TLAB–OrCAD co-simulation by e v aluating the system under v arious irradiance le v els and v oltage scenarios. A supervised learning approach w as emplo yed, utilizing Optimization of ANN-based DC volta g e contr ol using hybrid r ain optimization algorithm ... (Mohcine Byar) Evaluation Warning : The document was created with Spire.PDF for Python.
1714 ❒ ISSN: 2088-8694 the backpropag ation algorithm to minimize the mean squared error (MSE) during the training process. M S E = 1 n n X i =1 D i − ˆ D i 2 (2) Where D i and ˆ D i represent the tar get and predicted duty c ycles, respecti v ely . T o enhance con v er gence performance and a v oid entrapment in local optima, R O A is incorporated into the ANN training process [22], [23]. This algorithm is used to optimize critical h yperparameters—including learning rate, neuron allocation, and initial weight settings—that are typically chosen heuristically in standard ANN congurations. Inspired by the natural o w dynamics of raindrops, R O A f acil itates a balanced search between global e xploration and local e xploitation within the solution space, as sho wn in Figure 3 [24]. Figure 4 illustrates the training performance of the con v entional ANN compared to the R O A-optimized ANN. The R O A-enhanced model demonstrates f aster con v er gence and achie v es a lo wer nal error . Its adapti v e learning capability helps pre v ent stagnation during training and supports better generalization, which is particularly important in PV systems subject to signicant en vironmental v ariability . The progression of critical h yperparameters throughout the training process is depicted in Figure 5. R O A adapti v ely modies the learning rate and adjusts the neuron count, contrib uting to enhanced training stability and impro v ed performance [25]. In comparison, x ed h yperparameter settings in con v entional ANN training can result in suboptimal con v er gence beha vior or increased risk of o v ertting. The training w orko w that incorporates R O A into the ANN optimization process is illustrated in Figure 6. R O A starts by initializing a population of candidate solutions and progressi v ely renes them based on their mean squared error (MSE) performance. Through this iterati v e mechanism, the ANN controller is guided to w ard generating optimal duty c ycle v alues suitable for dynamic PV operating conditions. Figure 3. Con v er gence of R O A vs. con v entional optimization during ANN training Figure 4. T raining performance comparison between con v entional ANN and R O A-optimized ANN Int J Po w Elec & Dri Syst, V ol. 16, No. 3, September 2025: 1711–1720 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Po w Elec & Dri Syst ISSN: 2088-8694 ❒ 1715 By inte grating the learning ability of ANN with the adapti v e search ef cienc y of R O A, the proposed control approach deli v ers precise tracking, rapid con v er gence, and st rong rob ustness under uctuating irradiance conditions. These attrib utes establis h it as a reliable solution for adv anced DC v oltage re gulation in PV -po wered HGBC systems. Figure 5. Ev olution of learning rate and neurons per layer: R O A vs. con v entional ANN Figure 6. W orko w of R O A-inte grated ANN training 3. RESUL TS AND DISCUSSION This section pro vides a detailed e v aluation of the tr ansformerless HGBC system managed by the R O A-optimized ANN controller . The performance assessment is carried out through a MA TLAB–OrCAD co-simulation en vironment, enabling the observ ation of both control beha vior and circuit-le v el dynamics. The analysis emphasizes output v oltage re gulation, transient response characteristics, and controller rob ustness across four operating scenarios representati v e of real-w orld PV v ariations. The e v aluated PV system is congured to supply 2.2 kW using a 4S-2P module arrangement, deli v ering 128 V at its maximum po wer point (MPP). This v oltage is stepped up to 650 V by the HGBC, operating at a switching frequenc y of 100 kHz. K e y system parameters are pro vided in T able 1, while the simulation scenarios used for performance e v aluation are outlined in T able 2. The performance of the R O A-ANN controller is compared ag ainst con v entional ANN and PID Optimization of ANN-based DC volta g e contr ol using hybrid r ain optimization algorithm ... (Mohcine Byar) Evaluation Warning : The document was created with Spire.PDF for Python.
1716 ❒ ISSN: 2088-8694 controllers across the dened test scenarios. F our graphical results illustrate the dynamic responses under dif ferent conditions, follo wed by a comprehensi v e summary in T able 3, which quanties k e y performance metrics for each control strate gy . T able 1. Standalone solar system parameters P arameter V alue PV array T otal po wer output ( P P V ) 2204.16 W Conguration 4S-2P V oltage at MPP ( V mp ) 128 V Current at MPP ( I mp ) 17.22 A Daily ener gy production 9.2 kWh/day Annual ener gy production 3365 kWh/year High-g ain boost con v erter Input v oltage ( V in ) 128 V Output v oltage ( V out ) 650 V Switching frequenc y ( f s ) 100 kHz T able 2. Simulation scenarios for controller performance e v aluation Scenario Irradiance (W/m 2 ) Reference v oltage (V) S1: Constant irradiance 1000 650 S2: Constant irradiance, v ariable reference 1000 500, 550, 600, 650 S3: V ariable irradiance 300, 500, 800, 1000 650 S4: V ariable irradiance, v ariable reference 300–1000 500–650 T able 3. Controller performance summary across all scenarios Performance metric R O A-ANN ANN PID Rise time (s) < 0.1 0.6 > 1.5 Settling time (s) < 0.3 0.7 > 1.5 Ov ershoot (%) < 2 4–7 10–15 T racking noise V ery lo w Moderate High Response to ref. shift Immediate Acceptable Ov ershoot + delay Response to irradiance drop Rob ust Mild sag Ripple + drift Handling dual v ariation Excellent Acceptable Unstable Figure 7 sho ws the system’ s dynamic beha vior under scenario S1, which represents standard test conditions with constant irradiance and reference v oltage. The R O A-ANN controller e xhibits a rapid transient response, achie ving a rise time bel o w 0.1 s and maintaining zero o v ershoot—highlighting its high re gulation accurac y and system stability . This performance stems from t he R O A ’ s ability to conti n uous ly optimize the controller parameters in real time. In comparison, the con v entional ANN requires approximat ely 0.6 s to stabilize due to its stati c training limitations. The PID controller performs poorly in this scenario, e xhibiting a 12% o v ershoot and persistent oscillations be yond 1.5 s, mainly due to its x ed g ain conguration and limited adaptability to the HGBC’ s nonlinear dynamics. Scenario S2, illustrated in Figure 8, e v aluates the controllers’ ability to adapt to v arying reference v oltages. The R O A-ANN demonstrates precise tracking of each reference shift with minimal steady-state error and rapid con v er gence, emphasizing its strong general ization capability and responsi v eness to dynamic inputs. While the con v entional ANN maintains reasonable accurac y , it e xhibits noticeable delays during signicant transitions—such as from 500 V to 600 V—re v ealing the limitations of of ine-trained architectures. The PID controller performs inadequately in this scenario, sho wing o v ershoots of up to 15% and e xtended settling times, indicati v e of its restricted adaptability under time-v arying operating conditions. Scenario S3, sho wn in Figure 9, e xamines system performance under v arying irradiance le v els with a x ed reference v oltage. The R O A-ANN controller consi stently deli v ers stable and accurate output across all irradiance conditi ons, including at lo w le v els, demonstrating its rob ustness and adaptability to en vironmental disturbances. In contrast, the con v entional ANN e xhibits slight de viations at 300 W/m², at trib uted to limited e xposure to lo w-irradiance cases during the training phase. The PID controller once ag ain underperforms, sho wing ripple, v oltage sag, and delayed response—issues that stem from its static control structure and inadequate handling of input v ariability . Int J Po w Elec & Dri Syst, V ol. 16, No. 3, September 2025: 1711–1720 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Po w Elec & Dri Syst ISSN: 2088-8694 ❒ 1717 In the nal test scenario, S4, illustrated in Figure 10, both irradiance and reference v oltage are v aried simultaneously to assess controller performance under compounded disturbances. The R O A-ANN maintains f ast, accurate, and stable tracking throughout, ef fecti v ely handling the multi v ariable dynamics without introducing steady-state error . This highlights it s capability to generalize and adapt in highly dynamic en vironments. The con v entional ANN displays slo wer adaptation and increased output noise during concurrent transitions, particularly under lo w irradiance conditions. The PID controller f ails to re gulate the output reliably , e xhibiting pronounced instability and lar ge tracking errors, reaf rming its limitations in nonlinear systems subjected to simultaneous uctuations. Across all test scenarios, the R O A-ANN controller consistently outperforms con v entional ANN and PID approaches in terms of rise time, set tling response, noise attenuation, and tracking precision. Its resilience under both reference v oltage shifts and irradiance uctuations highlights the ef fecti v eness of R O A-dri v en optimization in enhancing real-time control adaptability . These ndings reinforce the suitability of R O A-ANN architectures for practical PV applications that demand f ast and reliable v oltage re gulation under dynamic operating conditions. Figure 7. Constant irradiance and x ed reference v oltage (S1) Figure 8. Constant irradiance and v ariable reference v oltage (S2) Optimization of ANN-based DC volta g e contr ol using hybrid r ain optimization algorithm ... (Mohcine Byar) Evaluation Warning : The document was created with Spire.PDF for Python.
1718 ❒ ISSN: 2088-8694 Figure 9. V ariable irradiance and constant reference v oltage (S3) Figure 10. V ariable irradiance and reference v oltage (S4) 4. CONCLUSION This study introduced a v oltage re gulation strate gy based on an R O A-optimized ANN controller for transformerless high-g ain boost con v erters in standalone PV systems. The controller’ s ef fecti v eness w as v alidated through MA TLAB–OrCAD co-simulation under multiple test conditi o ns in v olving v arying irradiance and reference v ol tages. The results demonstrated that the R O A-ANN consistently outperformed both con v entional ANN and PID controllers in terms of dynamic response, tracking precision, and o v erall rob ustness. These ndings underscore the v alue of inte grating metaheuristic optimization into neural netw ork training for adapti v e control in nonlinear po wer electronic applications. Future w ork will e xplore e xperimental v alidation and the e xtension of the proposed scheme to in v erter -le v el control architectures. FUNDING INFORMA TION Authors state no funding in v olv ed. Int J Po w Elec & Dri Syst, V ol. 16, No. 3, September 2025: 1711–1720 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Po w Elec & Dri Syst ISSN: 2088-8694 ❒ 1719 A UTHOR CONTRIB UTIONS ST A TEMENT This journal uses the Contrib utor Roles T axonomy (CRediT) to recognize indi vidual author contrib utions, reduce authorship disputes, and f acilitate collaboration. Name of A uthor C M So V a F o I R D O E V i Su P Fu Mohcine Byar ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Abdelouahed Abounada ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ C : C onceptualization I : I n v estig ation V i : V i sualization M : M ethodology R : R esources Su : Su pervision So : So ftw are D : D ata Curation P : P roject Administration V a : V a lidation O : Writing - O riginal Draft Fu : Fu nding Acquisition F o : F o rmal Analysis E : Writing - Re vie w & E diting CONFLICT OF INTEREST ST A TEMENT The authors declare that there are no nancial, personal, or professional conicts of interest that could ha v e inuenced the ndings, interpretations, or conclusions presented in this w ork. All contrib utions were made independently and according to ethical research standards. D A T A A V AILABILITY The data supporting the ndings of this study are a v ailable from the corres po ndi ng author , [MB], upon reasonable request. All simulation models and co-simulation setups were de v eloped by the authors and are not publicly archi v ed due to ongoing related research. REFERENCES [1] H. M. Saleh and A. I. Hassan, “The challenges of sustainable ener gy transition: A focus on rene w able ener gy , ” Applied Chemical Engineering , v ol. 7, no. 2, p. 2084, Apr . 2024, doi: 10.59429/ace.v7i2.2084. [2] A. I. Osman et al. , “Cost, en vironmental im pact, and resilience of rene w able ener gy under a changing climate: a re vie w , ” En vir onmental Chemistry Letter s , v ol. 21, no. 2, pp. 741–764, Apr . 2023, doi: 10.1007/s10311-022-01532-8. [3] A. Zisos , D. Chatzopoulos, and A. Efstratiadis, “The concept of spatial reliabi lity across rene w able ener gy systems—an application to decentralized solar PV ener gy , ” Ener gies , v ol. 17, no. 23, p. 5900, No v . 2024, doi: 10.3390/en17235900. [4] N. Mlilo, J. Bro wn, and T . Ahfock, “Impact of intermit tent rene w able ener gy generation penetration on the po wer system netw orks – A re vie w , ” T ec hnolo gy and Economics of Smart Grids and Sustainable Ener gy , v ol. 6, no. 1, p. 25, Dec. 2021, doi: 10.1007/s40866-021-00123-w . [5] M. Byar , G. Chbirik, A. Brahmi, and A. Abounada, “Hybrid ANN-PSO MPPT with high-g ain boost con v erter for standalone photo v oltaic systems, ” Bulletin of Electrical Engineering and Informatics , v ol. 14, no. 4, pp. 3189–3201, Aug. 2025, doi: 10.11591/eei.v14i4.9451. [6] A. F . de Souza, F . L. T ofoli, and E. R. Ribeiro, “Switched capacitor DC-DC con v erters: A surv e y on the main topologies, design characteristics, and applications, ” Ener gies , v ol. 14, no. 8, p. 2231, Apr . 2021, doi: 10.3390/en14082231. [7] P . Sumath y , N. Di vya, J. Sathik, A. La v an ya, K. V ijayakumar , and D. Almakhles, “ A comprehensi v e study on v arious DC–DC con v erter v oltage-boosting topologies and their applications, ” Cir cuit W orld , v ol. 48, no. 4, pp. 529–549, 2022, doi: 10.1108/CW -12-2020-0338. [8] F . Y i and F . W ang, “Re vie w of v oltage-b ucking/boosting techniques, topologies, and applications, ” Ener gies , v ol. 16, no. 2, p. 842, Jan. 2023, doi: 10.3390/en16020842. [9] B . T . Rao and D. De, “ A coupled inductor -based high-g ain ZVS DC–DC con v erter with reduced v oltage stresses, ” IEEE T r ansactions on P ower Electr onics , v ol. 38, no. 12, pp. 15956–15967, Dec. 2023, doi: 10.1109/TPEL.2023.3310577. [10] M . K umar , V . K. Y ada v , and A. K. V erma, “Switched capacitor based high g ain boost con v erter for rene w able ener gy application, ” IEEE J ournal of Emer ging and Selected T opics in Industrial Electr onics , v ol. 4, no. 3, pp. 818–826, Jul. 2023, doi: 10.1109/JESTIE.2023.3262451. [11] M. Rezaie and V . Abbasi, “Ef fecti v e combination of quadratic boost con v erter with v oltage multiplier cell to increase v oltage g ain, ” IET P ower Electr onics , v ol. 13, no. 11, pp. 2322–2333, Aug. 2020, doi: 10.1049/iet-pel.2019.1070. [12] D. Prasad, N. K umar , R. Sharma, H. Malik, F . P . Garcia M ´ arquez, and J. M. Pinar -P ´ erez, “ A no v el ANR O A based control approach for grid-tied multi-functional solar ener gy con v ersion system, ” Ener gy Reports , v ol. 9, pp. 2044–2057, Dec. 2023, doi: 10.1016/j.e gyr .2023.01.039. [13] R. Aghataher , H. Rabieif ar , N. N. Saman y , and H. Rezayan, “The suitability mapping of an urban spatial structure for earthquak e disaster response using a gradient rain optimization algorithm (GR O A), ” Heliyon , v ol. 9, no. 10, p. e20525, Oct. 2023, doi: 10.1016/j.heliyon.2023.e20525. Optimization of ANN-based DC volta g e contr ol using hybrid r ain optimization algorithm ... (Mohcine Byar) Evaluation Warning : The document was created with Spire.PDF for Python.
1720 ❒ ISSN: 2088-8694 [14] Y . V . R. S. V isw anadham and K. Jaya v el, “ A frame w ork for data pri v ac y preserving in supply chain management using h ybrid meta-heuristic algorithm with Ethereum blockchain technology , ” Electr onics (Switzerland) , v ol. 12, no. 6, p. 1404, 2023, doi: 10.3390/electronics12061404. [15] C . B. Ali, A. H. Khan, K. Perv ez, T . M. A w an, A. Noorw ali, and S. A. Shah, “High ef cienc y high g ain DC-DC boost con v erter using PID controller for photo v oltaic applic ations, ” in 2021 International Congr ess of Advanced T ec hnolo gy and Engineering (ICO TEN) , IEEE, Jul. 2021, pp. 1–7. doi: 10.1109/ICO TEN52080.2021.9493476. [16] R . W . Erickson and D. Maksimo vi ´ c, “The discontinuous conduction mode, ” in Fundamentals of P ower Electr onics , Cham: Springer International Publishing, 2020. doi: 10.1007/978-3-030-43881-4. [17] A. W . Kha w aja et al. , “ Angle stability impro v ement using optimised proportional inte gral deri v ati v e with lter controller , ” Heliyon , v ol. 10, no. 19, p. e38944, Oct. 2024, doi: 10.1016/j.heliyon.2024.e38944. [18] A. Daraz, S. A. Malik, A. T . Azar , S. Aslam, T . Alkhalif ah, and F . Alturise, “Optimized fractional order inte gral-tilt deri v ati v e controller for frequenc y re gulation of interconnected di v erse rene w able ener gy resources, ” IEEE Access , v ol. 10, pp. 43514–43527, 2022, doi: 10.1109/A CCESS.2022.3167811. [19] I. Alhamrouni et al. , “ A comprehensi v e re vie w on the role of articial intelligence in po wer system stability , c ontrol, and protection: insights and future directions, ” Applied Sciences , v ol. 14, no. 14, p. 6214, Jul. 2024, doi: 10.3390/app14146214. [20] M. M. Hammad, “Deep learning acti v ation functions: Fix ed-shape, parametric, adapti v e, stochastic, mis cellaneous, non-standard, ensemble, ” arXiv pr eprint arXiv:2407.11090 , 2024, doi: 10.48550/arXi v .2407.11090. [21] A. R. Moazzeni and E. Khamehchi, “Rain optimization algorithm (R O A): A ne w metaheuristic method for drilling optimization solutions, ” J ournal of P etr oleum Science and Engineering , v ol. 195, p. 107512, Dec. 2020, doi: 10.1016/j.petrol.2020.107512. [22] F . M. Alamgir and M. S. Alam, “ An articial intelligence dri v en f acial emotion recognition system using h ybrid deep belief rain optimization, ” Multimedia T ools and Applications , v ol. 82, pp. 2437–2464, 2023, doi: 10.1007/s11042-022-13378-x. [23] M. Ghasemi, M. Zare, P . T rojo vsk ´ y, R. V . Rao, E. T rojo vsk ´ a, and V . Kandasamy , “Optimization based on the smart beha vior of pl ants with its engineering applications: Ivy algorithm, ” Knowledg e-Based Systems , v ol. 295, p. 111850, Jul. 2024, doi: 10.1016/j.knosys.2024.111850. [24] M. Abd Elaziz et al. , “ Adv anced metaheuristic optimization techniques in applications of deep neural netw orks: a re vie w , ” Neur al Computing and Applications , v ol. 33, pp. 14079–14099, No v . 2021, doi: 10.1007/s00521-021-05960-5. [25] F . M. T alaat and S. A. Gamel, “RL based h yper -parameters optimization algorithm (R O A) for con v olutional neural netw ork, ” J ournal of Ambient Intellig ence and Humanized Computing , v ol. 14, pp. 13349–13359, Oct. 2023, doi: 10.1007/s12652-022-03788-y . BIOGRAPHIES OF A UTHORS Mohcine Byar recei v ed his master’ s de gree in Rene w able Ener gy and Ener gy Ef cienc y Engineering from Sult an Moulay Slimane Uni v ersity , Polydisciplinary F aculty , Beni Mellal, Morocco, in 2020. He is currently pursuing his Ph.D. at the F aculty of Science and T echnology , Beni Mellal. His research focuses on inno v ati v e control methods for solar ener gy systems, with a particular emphasis on optimizing solar technology to enhance ener gy ef cie nc y and sustainability . His w ork in v olv es de v eloping adv anced control strate gies to impro v e the performance and reliability of solar systems in di v erse applications. He can be contacted at email: mohcinebyar2013@gmail.com. Abdelouahed Abounada recei v ed his B.Sc. de gree from C adi A yyad Uni v ersity , Marrak ech, Morocco, in 1989, and his Ph.D. de gree in Automatic Control from the same uni v ersity in 1992. He is currently a professor at Sultan Moulay Sli mane Uni v ersity , Beni-Mellal, Morocco. His research interests include adv anced methods for numerical automation, po wer electronics, and the inte gration of rene w able ener gies into electrical systems. Abdelouahed is committed to adv ancing kno wledge and technology in these elds to contrib ute to the de v elopment of sustainable ener gy solutions. He can be contacted at email: a.abounada@gmail.com. Int J Po w Elec & Dri Syst, V ol. 16, No. 3, September 2025: 1711–1720 Evaluation Warning : The document was created with Spire.PDF for Python.