Inter national J our nal of A pplied P o wer Engineering (IJ APE) V ol. 15, No. 1, March 2026, pp. 69 ∼ 79 ISSN: 2252-8792, DOI: 10.11591/ijape.v15.i1.pp69-79 ❒ 69 Enhancing electr olyzer perf ormance f or h ydr ogen pr oduction in a solar system using a b uck con v erter with sliding mode contr ol Abdellah El Idrissi 1 , Belkasem Imodane 1 , M’hand Oubella 1 , Hatim Ameziane 2 , Mohamed Benydir 1 , Kaoutar Dahmane 1 , Driss Belkhiri 1 , Mohamed Ajaamoum 1 1 Laboratory of Engineering Sciences and Ener gy Management (LASIME), National School of Applied Sciences, Ibn Zohr Uni v ersity , Ag adir , Morocco 2 Laboratory of Science and T echnology for the Engineer (LaSTI), Electrical Engineering Department, National School of Applied Sciences (ENSA), Sultan Moulay Slimane Uni v ersity , Khouribg a, Morocco Article Inf o Article history: Recei v ed Jan 12, 2025 Re vised Aug 26, 2025 Accepted Oct 16, 2025 K eyw ords: DC-DC con v erter Electrolyzer Hydrogen production Rene w able ener gy Sliding mode control ABSTRA CT As the w orld increasingly turns to rene w able ener gy , green h ydroge n produced through w ater electrolysis has emer ged as a cl ean and promising alternati v e to fossil fuels. In this w ork, we e xplore a solar -po wered h ydrogen production system that uses real data from an operational photo v oltaic (PV) installation, ensuring accurate and realistic modeling of en vironmental conditions. A DC-DC b uck con v erter is used to re gulate the uctuating PV output, supplying the precise v oltage needed by a PEM electrolyzer . Sli ding mode control (SMC) strate gy is applied to maintain v oltage stability , and its performance is compared with a tradit ional proportional-inte gral (PI) controller . Simulations in MA TLAB/Simulink demonstrate that SMC of fers better dynamic performance, including minimal o v ershoot, f aster response, and an impressi v e h ydrogen production rate of 0.98 L/min (98% ef cienc y). By pro viding more consistent v oltage to the electrolyzer , SMC signicantly boosts o v erall system performa nce. These ndings underline the potential of adv anced control strate gies, support ed by real-w orld data, to mak e rene w able h ydrogen production more reliable and ef cient. This is an open access article under the CC BY -SA license . Corresponding A uthor: Abdellah El Idrissi Laboratory of Engineering Sciences and Ener gy Management (LASIME) National School of Applied Sciences, Ibn Zohr Uni v ersity Ag adir 80000, Morocco Email: abdellah.elidrissi@edu.uiz.ac.ma 1. INTR ODUCTION The accelerat ing depletion of fossil fuels, coupled with the ur gent need to mitig ate climate change, has intensied the global transition to w ard cleaner and more sustainable ener gy systems. Among the emer ging solutions, green h ydrogen that produced through electrolysis po wered by rene w able ener gy sources such as solar or wind, has g ained prominence for its potential to decarbonize se v eral sectors including industry , transportation, and po wer generation [1]. Morocco is well positioned to become a re gional and global hub for green h ydrogen production thanks t o its ab undant sunshine, rene w able ener gy resources, and strate gic proximity to major e xport mark ets. This potential is reinforced by the country’ s ambitious strate gy , which includes lar ge-scale in v estments in h ydrogen and its deri v ati v es, such as ammonia and methanol, thereby J ournal homepage: http://ijape .iaescor e .com Evaluation Warning : The document was created with Spire.PDF for Python.
70 ❒ ISSN: 2252-8792 consolidating its role in the global ener gy transition. In the broader conte xt of carbon neutrality goals, green h ydrogen emer ges as an essential ener gy carrier , of fering a path to decarbonization in hard-to-abate sectors while supporting the global transition to sustainable ener gy systems [2]. The electrolysis process in v olv es spl itting w ater molecules into h ydrogen and oxygen using electr icity from distrib uted ener gy sources (DES) lik e photo v oltaic (PV) or wind system [3]. When po wered by rene w ables, this process emits no carbon dioxide, making green h ydrogen a truly clean ener gy carrier [4]. Current research focuses on impro ving electrolysis ef cienc y while optimizing the inte gration of DES with electrolyzer systems to ensure economic viability and operational stability [5]. Among the dif ferent types of electrolyzers, proton e xchange membrane (PEM) electrolyzers stand out due to their f ast response, compactness, and compatibility with v ariable inputs. Ho we v er , coupling PEM systems with intermittent ener gy sources such as solar and wind remains technically challenging, primarily due to uctuating v oltage and current le v els [6], [7]. Photo v oltaic-po wered PEM systems represent a promising conguration, b ut the v ariable nature of solar irradiance—af fected by f actors lik e cloud co v er , tim e of day , and temperature—introduces issues that impact h ydrogen production rates and electrolyzer durability [8], [9]. T o address this, adv anced control strate gies are necessary to maintain v oltage stability , reduce ener gy losses, and ensure continuous h ydrogen generation under changing en vironmental conditions [10]. Gi v en the PEM electrolyzer’ s requirement for lo w v oltage and high current, inte grating a DC-DC b uck con v erter becomes essential to adapt the PV output to the required input l e v el s [11], [12]. The performance of this po wer conditioning stage strongly depends on the ef fecti v eness of its control method. T raditionally , proportional inte gral (PI) controllers ha v e been emplo yed in PV -electrolyzer systems due to their simplicity and satisf actory steady-state performance [13], [14]. Ho we v er , the y often f all short under dynamic and nonlinear operating conditions, which are common in solar -po wered systems. to o v ercome this chall enges, more adv anced control strate gies ha v e been introduced such as fuzzy logic control (FLC) [15], [16], model predicti v e control (MPC) [17], and neural netw orks (NN) [18] ha v e demonstrated better adaptability to system uncertaint ies and disturbances. Ne v ertheless, their implementation can be comple x and computationally demanding, limiting their widespread deplo yment. Out of the v arious rob ust control strate gies, sliding mode contr o l (SMC) has emer ged as a particularly attracti v e option for PV–PEM h ydrogen production systems due to its insensiti vity to parameter v ariations, f ast dynamic response, and strong disturbance rejection capabilities [19]. In dynamic solar en vironments, con v entional controllers such as PI often suf fer from slo wer tra n s ient response, sensiti vity to parameter changes, and performance de gradation under rapid irradiance and temperature uctuations. In contrast, SMC of fers superior rob ustness, f aster con v er gence, and better tracking accurac y , making it highly ef fecti v e for maintaining v oltage stability and optimizing h ydrogen production ef cienc y . Its ability to handle system nonlinearities, parameter v ariations, and e xternal disturbances ensures stable operation and high ef cienc y , e v en under rapidly changing solar input. By ef fecti v ely managing these dynamic conditions, SMC enhances both system reliability and h ydrogen yield. Unlik e pre vious studies that rely on idealized or simulated solar proles, this w ork distinguishes itself by emplo ying real-w orld solar data collected from an operati on a l PV installation at the Higher School of T echnology of Ag adir . This enables a more accurate e v aluation of control strate gies under realistic en vironmental v ariations. The performance of SMC is compared with that of a con v entional PI controller , with the goal of enhancing system stabil ity , h ydrogen production ef cienc y , and o v erall operational rob ustness under uctuating solar conditions. 2. DESCRIPTION AND DESIGN OF THE SYSTEM The photo v oltaic h ydrogen production system st ud i ed in this w ork is sho wn in Figure 1, It comprises a PV array , a DC-DC b uck con v erter , and a PEM electrolyzer . The PV array con v erts solar irradiance into DC electr icity , while a control unit re gulates the v oltage to match system requirements. The b uck con v erter adjusts this v oltage to the le v el needed by the electrolyzer , which then uses the re gulated po wer to perform w ater electrolysis for h ydrogen production and storage. 2.1. Model of the PEM electr olyser Modeling a proton e xchange membrane (PEM) electrolyzer as an equi v alent electrical circuit enables safe testing of control strate gies without ph ysical equipment. The model includes the re v ersible v oltage E re v and three resistances: R ohm (internal losses), R act (acti v ation losses), and R con (concentration limitat ions). The total v oltage V elz is gi v en by (1) [20]. T o analyze v oltage and po wer v ariations with current, a commercial Int J Appl Po wer Eng, V ol. 15, No. 1, March 2026: 69–79 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Appl Po wer Eng ISSN: 2252-8792 ❒ 71 400 W PEM electrolyzer w as selected; its specications are in T able 1. The system can be approximated linearly by a v oltage source in series with resistance, described by (2). V elz = E re v + I elz ( R act + R ohm + R con ) (1) V elz = 0 . 0625 I elz + 4 . 375 (2) The linear model coef cients were obtained by interpolating e xperimental data o v er a current range of 3–50 A [21], sho wing a nearly linear v oltage–current relationship, as sho wn in Figure 2. This v alidates the simplied equi v alent circuit model with a re v ersible v oltage source and a resisti v e component. Hydrogen production ( ˙ N H 2 ) is calculated using F araday’ s la w [22]. ˙ N H 2 = nI 2 F (moles/s) , ˙ N H 2 = 0 . 00696 nI (L/min) (3) Where n is the number of cells, I the applied current, and F F araday’ s constant. F araday ef cienc y ( η F ) e xpresses the fraction of char ge con v erted into h ydrogen, as in (4). η F = Actual moles of H 2 Theoretical moles of H 2 × 100% (4) Higher ef cienc y increases h ydrogen yield and reduces ener gy consumption. 2.2. Solar PV array The solar PV unit is used to produce po wer output. The s olar panel in thi s article is a single unit with a maximum po wer input capacity of 120.7 W . The solar module specications are sho wn in T able 2 [23]. T able 1. PEM electrolyzer specications Specication V alues (unit) Rated po wer ( P el ) 400 W Operating v oltage ( V elz ) 2.2–8 V Electrolyzer current ( I elz ) 0–50 A Output pressure 0.1–10.5 bar H 2 o w rate 1 L/min ( T = 20 °C, P = 1 bar) Cell Numbers 3 T able 2. Solar panel parameters for W aaree Ener gies WU-120 P arameters V alue I mp 7.1 A V mp 17 V P max,e 120.7 W I sc 8 A V oc 21 V PV solar irradiation (G) 1000 W/m² PV operation temperature (T) 25 °C DC DC C on tr ol l e r   H yd r oge n   S tor age   D i s ti l l e d  w ate r H 2 O ½ O 2 H2 A n od e C at h od e   M e m b r an e H+ - + Figure 1. Schematic of the proposed system to produce h ydrogen using solar ener gy Enhancing electr olyzer performance for hydr o g en pr oduction in a solar system using ... (Abdellah El Idrissi) Evaluation Warning : The document was created with Spire.PDF for Python.
72 ❒ ISSN: 2252-8792 0 10 20 30 40 50 Current (A) 0 100 200 300 400 Power P e l z  (W) Theoretical Power Experimental Data (a) 0 5 10 15 20 25 30 35 40 45 50 Current (A) 0 2 4 6 8 VoltageV e l z   (V) Theoretical Voltage Experimental Data (b) Figure 2. Static characteristics of the selected PEM electrolyzer cells: (a) electrolyzer po wer vs. current and (b) v oltage vs. current, with e xperimental data (o) and model tting (solid line) 2.3. Modeling of b uck con v erter The b uck con v ert er is a type of DC-DC con v erter that is widely used in po wer electronics appli cations due to its simplicity , ef cienc y , and ability to step do wn the input v oltage. The con v erter consists of a switch (S), an inductor (L), a diode (D), and a capacitor (C). Electrolyzers typically operate at a lo w DC v oltage for w ater electrolysis, so the use of a DC-DC con v erter , as sho wn in Figure 3 is essential. In addition to reducing the v oltage, these con v erters manage v oltage adaptation to handle uctuations in the v oltage pro vided by the solar panels. By selecting the appropriate components and adjusting the con v erter parameters according to the specic application requirements, we minimize po wer loss es, and e xtend the life of the components [24]. T able 3 sho ws the sizing of the v arious components of the b uck con v erter . C V i n V C V e l z L i C R D S DC Figure 3. Electrical schematic of the DC-DC b uck con v erter T able 3. Buck con v erter parameters P arameters V alues Inductor L 497 µH Capacitor C 7.94 µF Switching frequenc y f 40 kHz Duty c ycle α 22.8% Input v oltage V in 30 V Output v oltage v elz 7.5 V The b uck con v erter operates in tw o modes depending on the switch state. When ON, the inductor and capacitor dynamics are go v erned by (5). di L dt = V in − v elz L , dv elz dt = i L C − v elz R C (5) When OFF as in (6). di L dt = − v elz L , dv elz dt = i L C − v elz R C (6) These can be written in the state-space form as (7). ˙ x = Ax + B V in , Y = N x (7) Int J Appl Po wer Eng, V ol. 15, No. 1, March 2026: 69–79 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Appl Po wer Eng ISSN: 2252-8792 ❒ 73 Where: x = i L v elz , N = 0 1 State-space a v eraging o v er one switching period, as (8). A = α A 1 + (1 − α ) A 2 , B = α B 1 (8) W ith: A 1 = A 2 = − 1 L 0 1 C − 1 R C , B 1 = 1 L 0 The a v eraged model becomes (9). ˙ x = " − (1 − α ) L 0 (1 − α ) C − 1 R C # x + 1 L 0 V in (9) 3. CONTR OL STRA TEGIES Sliding mode controller (SMC) i s a po werful control technique that can pro vide f ast and rob ust cont rol of systems. Its ef fecti v eness has been demonstrated in v arious applications, including DC-DC con v erters, where it can pro vide superior performance compared to other control techniques [25]. The control v ariable x Buck is presented in (10), where the v ariables x 1 , x 2 , and x 3 are respecti v ely the v oltage error , the deri v ati v e of the error , and the inte gral of the error . The instantaneous capacitor , inductor , and load currents, respecti v ely; V ref , v i , and β v elz denote the reference, instantaneous input, and instantaneous output v oltages, respecti v ely; β denotes the feedback netw ork ratio; and u = 0 or 1 is the switching state of po wer switch SW . The equation of state for the control system in the v ector space is written in (10). x Buck = x 1 = V ref − β v elz x 2 = ˙ x 1 = d dt ( V ref − β v elz ) x 3 = R x 1 dt = R ( V ref − β v elz ) dt and ˙ x Buck = ˙ x 1 = x 2 ˙ x 2 = − 1 R C x 2 + β v in LC u + β v elz LC ˙ x 3 = x 1 (10) The state-space model describing the system can be deri v ed as (11). ˙ x Buck = 0 1 0 0 − 1 R C 0 1 0 0 x + 0 − β v in LC 0 u eq + 0 β v elz LC 0 (11) The SMC la w uses a switching function to determine the control signal, as (12). u = ( 1 when S > 0 0 when S < 0 where S dened as S = a 1 x 1 + a 2 x 2 + a 3 x 3 = J T x (12) Where S is the instantaneous state trajectory and J T = [ a 1 , a 2 , a 3 ] and a 1 , a 2 , a 3 are the sliding coef cients. The sliding mode control ensures that the system meets the sliding conditions: hitting, e xistence, and stability . The ramp si gnal and control signal are compared to get the output switching signal, which has a frequenc y identical to the ramp signal. By xing the ramp signal frequenc y , the output switching signal frequenc y remains constant. Therefore, using the PWM technique in controller design ensures a x ed frequenc y for the proposed method. In the rst step, the equi v alent c o nt rol signal u eq is deri v ed using the in v ariance condition. In the second step, u eq is translated to the duty ratio α of the PWM during the deri v ation process. The equi v alent control signal u eq is obtained from the equation ˙ S = J T Ax + J T B u eq + D = 0 , which yields the equi v alent control function, as (13). Solving for u eq : u eq = − β L β v in a 1 a 2 − 1 R C i C + a 3 LC a 2 β v in ( V r ef − β v el z ) + v el z v in (13) Enhancing electr olyzer performance for hydr o g en pr oduction in a solar system using ... (Abdellah El Idrissi) Evaluation Warning : The document was created with Spire.PDF for Python.
74 ❒ ISSN: 2252-8792 T ranslating the equi v alent control as in (13) to the duty ratio α , where 0 < α = v c ˆ v ramp < 1 , gi v es the follo wing relationships for the control signal v c and ramp signal ˆ v ramp as in (14). F or the practical implementation of the PWM-based sliding mode controller , the electrical schematic of the SMC controller is sho wn in Figure 4. v c = u eq = k 1 i C + k 2 ( V ref − β v elz ) + β v elz and ˆ v ramp = β v in (14) Where: k 1 = − β L a 1 a 2 − 1 R C , k 2 = a 3 a 2 LC K2 + - V r e f β V elz S l i d i n g M od e  C on tr ol l e r Vc K1 ic + + + PWM - + β V i n V r am p C V i n V C V e l z L i C R D S DC Figure 4. Electrical schema of SMC controller 4. RESUL TS AND DISCUSSION 4.1. Steady-state perf ormance T o v erify the ef cac y of sliding mode control in re gulating v oltage to the desired v alue for po wering the electrolyzer used in the study , we compare the results obtained us ing this method with those obtained using a classical PI controller , which w as also emplo yed for the same purpose. W e used the MA TLAB/Simulink platform for this comparison. The steady-state operational conditions of the system, including the solar PV parameters, b uck con v erter , and PEM electrolyzer , are sho wn under rated conditions. The solar iradiation is x ed at 1000 W/m², and the temperature is held at 25 °C. The solar panel v oltage, current, and po wer are stable at 30 V , 30 A, and 897 W , respecti v ely , under full irradiance. These v alues indicate that the PV system is operating at its full rated capacity . The PEM electrolyzer , po wered by the solar PV , also operates at full rated po wer under these conditions. The steady-state electrolyzer v oltage (7.5 V), current (47.15 A), and po wer (354 W) are achie v ed by both PI and SMC control strate gies, Figure 5 illustrates the performance of both control methods. In Figure 5(a), we observ e the electrolyzer v oltage ( V el z ), and in Figure 5(b), the h ydrogen production rate ( H 2 ). Both controllers successfully ensure st able operation, with the v oltage settling around 7.5 V and the h ydrogen production rate stabilizing at approximately 0.985 L/min. These results conrm the system’ s ef fecti v e and reliable performance under steady-state conditions. Ho we v er , as demonstrated in Figure 5, the SMC pro v es to be slightly more accurate and precise than the PI controller in steady-state mode. The results sho w that the SMC performs signicantly better than the PI controller in terms of stability , accurac y , and response time. In summary , the SMC’ s lo wer v alues across both system response and error metrics mak e it the more ef fecti v e controller , with greater stability , f aster settling, and a much lo wer error prole compared to the PI controller . Int J Appl Po wer Eng, V ol. 15, No. 1, March 2026: 69–79 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Appl Po wer Eng ISSN: 2252-8792 ❒ 75 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (s) 0 2 4 6 8 10 V e l z  (V) Electrolyzer Voltage PI Electrolyzer Voltage SMC 0.02 0.04 7 7.5 8 8.5 2 2.05 2.1 7.48 7.5 7.52 (a) 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (s) 0 0.5 1 H 2  Production (L/min) Hydrogen Production Rate PI Hydrogen Production Rate SMC 0.02 0.04 0.06 0.08 0.95 1 1.05 2 2.05 2.1 0.98 0.985 0.99 (b) Figure 5. Performance of the system under constant solar irradiation: (a) the electrolyzer v oltage ( V el z ) and (b) the h ydrogen production rate ( H 2 ) 4.2. Dynamic perf ormance In order to e v aluate the dynamic performance of our controller , we used e xperimental data collec ted from a photo v oltaic panel installed on the rooftop of our laboratory at the Higher School of T echnology of Ag adir . The dataset includes solar irradiance, panel surf ace temperature, and output v oltage. illustrated in Figure 6, the data sho w noticeable uctuations: irradiance ranges from 400 W/m 2 to 1000 W/m 2 , while the surf ace temperature of the PV panel v aries between 42 °C and 62 °C, as sho wn in Figures 6(a) and 6(b). These v ariations directly af fect k e y PV system parameters, particularly the panel v oltage ( V pv ), depicted in Figure 6(c), and panel po wer ( P pv ) in Figure 6(d) respecti v ely . The irradiance prole initially sho ws a steady increase, follo wed by a sharp decline near the midpoint, indicati n g a temporary drop in solar input. This change is clearly reected in the corresponding v oltage and current responses of the system. The results in Figures 7(a) and 7(b) re v eal that the SMC consistently outperforms the PI controller in terms of stability , accurac y , and dynamic response. The SMC demonstrates minimal o v ershoot at 0.0299%, signicantly lo wer than the PI’ s 12.2873%, indicating a more controlled and steady response close to the tar get v alue. While both controllers e xhibit rapid response times, the SMC achie v es a much f aster settling time of 0.0032 seconds compared to the PI’ s 0.0267 seconds. Additionally , the SMC has a lo wer mean absolute percentage error (MAPE) of 0.21%, is signicantly lo wer for the SMC, indicating better accurac y in achie ving the desired output. Here, V ref represents the desired v alue (7.5 V) y actual ,i is the actual response at each time step i , and n is the number of time steps. contrasting sharply with the PI’ s 6.96%, further emphasizing its superior tracking accurac y , When e xamining error metrics, the SMC ag ain distinguishes i tself with superior performance indices. The error signal is dened as e ( t ) = V ref − V elz ( t ) . 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (s) 400 600 800 1000 Ir (W/m 2 ) Irradiation (a) 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (s) 40 45 50 55 60 65 T p v  (°C) PV Temperature (b) 0 0.5 1 1.5 2 2.5 3 3.5 Time (s) 10 15 20 25 30 V p v  (V) PV Voltage (c) 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (s) 200 400 600 800 1000 P p v  (W) PV Power (d) Figure 6. Dynamic performance of the system: (a) solar irradiance ( I r ), (b) temperature of the photo v oltaic panel ( T pv ), (c) the panel v oltage ( V pv ), and (d) the panel po wer ( P pv ) Enhancing electr olyzer performance for hydr o g en pr oduction in a solar system using ... (Abdellah El Idrissi) Evaluation Warning : The document was created with Spire.PDF for Python.
76 ❒ ISSN: 2252-8792 Where V ref (7.5 V) is the desired v oltage and V elz ( t ) is the actual electrolyzer v oltage at time t , the inte gral of squared error (ISE) for the SMC is lo wer at 0.0386 compared to 0.0476 for the PI controller , reecting a more ef cient reduction in o v erall error ener gy . Similarly , the inte gral of absolute error (IAE) is signicantly reduced to 0.0114 for the SMC, whereas the PI controller reaches 0.1362. The most notable impro v ement appears in the inte gral of time-weighted absolute error (IT AE), where the SMC achie v es a remarkably lo w v alue of 0.0039, in contrast to 0.2378 for the PI controller . These results conrm that the SMC not only minimizes the o v erall error b ut also reacts more promptly to disturbances, resulting in smoother and more accurate system performance, T able 4 summarizes the dynamic performance and error metrics of both controllers. 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (s) 0 2 4 6 8 V e l z  (V) Electrolyzer Voltage PI Electrolyzer Voltage SMC 0.02 0.03 0.04 7 7.5 8 1.5 2 2.5 7.48 7.5 7.52 (a) 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (s) 0 0.5 1 H 2  Production (L/min) Hydrogen Production Rate PI Hydrogen Production Rate SMC 0.02 0.04 0.06 0.9 1 1.1 2.05 2.1 0.984 0.9845 0.985 (b) Figure 7. Dynamic performance of the system with v arying solar irradiation and temperature: (a) the electrolyzer v oltage ( V el z ) and (b) the h ydrogen production rate ( H 2 ) T able 4. Dynamic performance and error metrics of controllers Controller MAPE (%) Settling time (s) Ov ershoot (%) ISE IAE IT AE Equation – – – Z e 2 ( t ) dt Z | e ( t ) | dt Z t | e ( t ) | dt PI 6.96 0.0267 12.2873 0.0476 0.1362 0.2378 SMC 0.21 0.0032 0.0299 0.0386 0.0114 0.0039 5. CONCLUSION This study in v estig ated the use of adv anced control techniques, particularly sliding mode control (SMC), to address the challenges of inte grating uctuating rene w able ener gy sources with PEM electrolyzers for green h ydrogen production. The system includes a photo v oltaic array connected to an electrol ysis unit via a DC-DC b uck con v erter , which adjusts the solar panel v oltage to the le v el needed by the electrolyzer . MA TLAB/Simulink simulations demonstrated that SMC outperforms the traditional P I controller by achie ving nearly zero o v ershoot, a v ery f ast settling time of 0.003 s econds, and a lo w error rate (MAPE of 0.21%). Additionally , SMC enabled a high h ydrogen production rate of 0.98 liters per minute with 98% ef cienc y , thanks to its ability to pro vide a stable and consistent v oltage to the electrolyzer . These results underscore the potential of adv anced control strate gies, supported by real-w orld data, to enhance the reliability and ef cienc y of rene w able h ydrogen production, laying a solid foundation for future w ork aimed at further optimizing control methods and impro ving system performance. FUNDING INFORMA TION Authors state no funding in v olv ed. 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. Int J Appl Po wer Eng, V ol. 15, No. 1, March 2026: 69–79 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Appl Po wer Eng ISSN: 2252-8792 ❒ 77 Name of A uthor C M So V a F o I R D O E V i Su P Fu Abdellah El Idrissi ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Belkasem Imodane ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ M’hand Oubella ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Hatim Ameziane ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Mohamed Ben ydir ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Kaoutar Dahmane ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Driss Belkhiri ✓ ✓ ✓ ✓ ✓ Mohamed Ajaamoum ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 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 the y ha v e no kno wn competing nancial interests or personal relations hips that could ha v e appeared to inuence the w ork reported in this paper . D A T A A V AILABILITY The MA TLAB simulation data that support the ndings of this study will be made a v ailable in an open-access repository upon acceptance of the manuscript. REFERENCES [1] T .-Z. Ang, M. Salem, M. Kamarol, H. S. Das, M. A. Nazari, and N. Prabaharan, “ A comprehensi v e study of rene w able ener gy sources: classications , challenges and suggestions, ” Ener gy Str ate gy Re vie ws , v ol. 43, p. 100939, Sep. 2022, doi: 10.1016/j.esr .2022.100939. [2] A. Caillard, R. Y e g an yan, C. Cannone, F . Plazas-Ni ˜ no, and M. Ho wells, “ A critical analysis of Morocco’ s green h ydrogen roadmap: a modelling approach to assess country re adiness from the ener gy trilemma perspecti v e, ” Climate , v ol. 12, no. 5, p. 61, Apr . 2024, doi: 10.3390/cli12050061. [3] P . Fern ´ andez-Arias, ´ A. Ant ´ on-Sancho, G. Lampropoulos, and D. V er g ara, “On green h ydrogen generation technologies: a bibliometric re vie w , ” Applied Sciences , v ol. 14, no. 6, p. 2524, Mar . 2024, doi: 10.3390/app14062524. [4] W . Mei, L. Sun, and Y . Zhao, “Ov ervie w of h ydrogen ener gy and general aspects of w ater electrolysis, ” in Gr een Hydr o g en Pr oduction by W ater Electr olysis , Boca Raton: CRC Press, 2024, pp. 1–26, doi: 10.1201/9781003368939-1. [5] M. K oundi et al. , “In v estig ation of h ydrogen production system -based PEM EL: PEM EL modeling, DC/DC po wer con v erter , and controller design approaches, ” Clean T ec hnolo gies , v ol. 5, no. 2, pp. 531–568, Apr . 2023, doi: 10.3390/cleantechnol5020028. [6] A. Baraean, M. Kassas, M. S. Alam, and M. A. Abido, “Ph ysics-informed NN-based adapti v e backstepping terminal sliding mode control of b uck con v erter for PEM el ectrolyzer , ” Heliyon , v ol. 10, no. 7, p. e29254, Apr . 2024, doi: 10.1016/j.heliyon.2024.e29254. [7] S. G. Nnab uife, A. K. Hamzat, J. Whidborne, B. K uang, and K. W . Jenkins, “Inte gration of rene w able ener gy sources in tandem with electrolysis: A technology re vie w for green h ydrogen production, ” International J ournal of Hydr o g en Ener gy , v ol. 107, pp. 218–240, Mar . 2025, doi: 10.1016/j.ijh ydene.2024.06.342. [8] I. Arias et al. , “ Assessing syst em-le v el syner gies between photo v oltaic and proton e xchange membrane electrolyzers for solar -po wered h ydrogen production, ” Applied Ener gy , v ol. 368, p. 123495, Aug. 2024, doi: 10.1016/j.apener gy .2024.123495. [9] H. A. Z. AL-bonsrulah et al. , “Design and simulation studies of h ybrid po wer systems bas ed on photo v oltaic, wind, electrolyzer , and PEM fuel cells, ” Ener gies , v ol. 14, no. 9, p. 2643, May 2021, doi: 10.3390/en14092643. [10] D. Y . Ga vrailo v , S. V . Bo yche v a, and X. Gao, “ A direct coupled photo v oltaic - electrolyser system for producing green h ydrogen, ” IOP Confer ence Series: Earth and En vir onmental Science , v ol. 1380, no. 1, 2024, doi: 10.1088/1755-1315/1380/1/012010. [11] A. Al obaid and R. A. Adomaitis, “Optimal design of a coupled photo v oltaic–el ectrolysis-battery system for h ydrogen generation, ” Sustainable Ener gy & Fuels , v ol. 7, no. 6, pp. 1395–1414, 2023, doi: 10.1039/D2SE01555B. [12] M. Chen, S.-F . Chou, F . Blaabjer g, and P . Da v ari, “Ov ervie w of po wer electronic con v erter topologies enabling lar ge-scale h ydrogen production via w ater electrolysis, ” Applied Sciences , v ol. 12, no. 4, p. 1906, Feb . 2022, doi: 10.3390/app12041906. [13] M . E. S ¸ ahin, “ A photo v oltaic po wered electrolysis con v erter system with maximum po wer point tracking control, ” International J ournal of Hydr o g en Ener gy , v ol. 45, no. 16, pp. 9293–9304, Mar . 2020, doi: 10.1016/j.ijh ydene.2020.01.162. [14] R . K. K umar and P . Samuel, “Designing a h ydrogen generation system through PEM w ater electrolysis with the capability to adjust f ast uctuations in photo v oltaic po wer , ” International J ournal of Hydr o g en Ener gy , v ol. 82, pp. 1–10, Sep. 2024, doi: 10.1016/j.ijh ydene.2024.07.376. Enhancing electr olyzer performance for hydr o g en pr oduction in a solar system using ... (Abdellah El Idrissi) Evaluation Warning : The document was created with Spire.PDF for Python.
78 ❒ ISSN: 2252-8792 [15] M . Ben ydir , M. Oubella, S. Mouslim, M . Ajaamoum, K. Dahmane, and B. Imoudane, “Implementation and analysis of a fuzzy logic and sliding mode controller on a boost DC/DC con v erter in a PV array , ” International J ournal of Rene wable Ener gy Resear c h , v ol. 12, no. 1, pp. 294–301, 2023, doi: 10.20508/ijrer .v13i1.13862.g8683. [16] B . Imodane, M. Ben ydir , B. Bouachrine, M. Ajaamoum, and K. Dahmane, “Experimental v alidati on of tw o v oltage re gulation strate gies for boost con v erters in wind systems, ” International J ournal of P ower Electr onics and Drive Systems (IJPEDS) , v ol. 16, no. 1, pp. 509–518, Mar . 2025, doi: 10.11591/ijpeds.v16.i1.pp509-518. [17] W . El Aouni and L. A. Dessaint, “Real-time implementation of input-s tate linearization and model predicti v e control for rob ust v oltage re gulation of a DC-DC boost con v erter , ” IEEE Access , v ol. 8, pp. 192101–192108, 2020, doi: 10.1109/A CCESS.2020.3032327. [18] S . D. Gangula, T . K. Nizami, R. R. Udumula, A. Chakra v arty , and P . Singh, “ Adapti v e neural netw ork control of DC–DC po wer con v erter , ” Expert Systems with Applications , v ol. 229, p. 120362, No v . 2023, doi: 10.1016/j.esw a.2023.120362. [19] Y . Mousa vi, G. Be v an, I. B. K ucukdemiral, and A. Fekih, “Sliding mode control of wind ener gy con v ersion systems: trends and applications, ” Rene wable and Sustainable Ener gy Re vie ws , v ol. 167, 2022, doi: 10.1016/j.rser .2022.112734. [20] D. S. F alc ˜ ao and A. M. F . R. Pinto, “ A re vi e w on PEM electrolyzer modelling: guidelines for be ginners, ” J ournal of Cleaner Pr oduction , v ol. 261, p. 121184, Jul. 2020, doi: 10.1016/j.jclepro.2020.121184. [21] F . Alonge, S. M. Collura, F . D’Ippolito, D. Guilbert, M. Luna, and G. V itale, “Design of a rob ust controller for DC/DC con v erter–electrolyzer systems supplied by µ WECSs subject to highly uctuating wind speed, ” Contr ol Engineering Pr actice , v ol. 98, p. 104383, May 2020, doi: 10.1016/j.conengprac.2020.104383. [22] B. Y odw ong, D. Guilbert, M. Phattanasak, W . Kae wmanee, M. Hinaje, and G. V itale, “F araday’ s ef cienc y modeling of a proton e xchange membrane electrolyzer based on e xperimental data, ” Ener gies , v ol. 13, no. 18, p. 4792, Sep. 2020, doi: 10.3390/en13184792. [23] L. Nadam, M. Chakra v arth y , and M. Manjula, “Optimized ener gy management for PV h ybrid po wer systems with DC b us v oltage control, ” International J ournal of Electrical and Electr onics Resear c h , v ol. 12, no. 4, pp. 1173–1180, 2024, doi: 10.37391/IJEER.120407. [24] N. Deshmukh, S. Prabhakar , and S. Anand, “Po wer loss reduction in b uck con v erter based acti v e po wer decoupling circuit, ” IEEE T r ansactions on P ower Electr onics , v ol. 36, no. 4, pp. 4316–4325, 2020, doi: 10.1109/TPEL.2020.3024721. [25] ¨ O. T ¨ urkso y and A. T ¨ urkso y , “ A f ast and rob ust sliding mode controller for automatic v oltage re gulators in electrical po wer systems, ” Engineering Science and T ec hnolo gy , an International Journal, v ol. 53, 2024, doi: 10.1016/j.jestch.2024.101697. BIOGRAPHIES OF A UTHORS Abdellah El Idr issi is an Ph.D. student in engineering sciences at Ibn Zohr Uni v ersity (UIZ), Ag adir , Morocco, and a member of Laboratory of Engineering Sciences and Ener gy Management (LASIME) at the High School of T echnologies of Ag adir (EST A). His doctoral research focuses on rene w able ener gy systems for h ydrogen production, emphasizing the optimization of solar -po wered electrolyzers through adv anced control strate gies and the inte gration of po wer electronics. He can be contacted at email: abdellah.elidrissi@edu.uiz.ac.ma. Belkasem Imodane is an Ph.D. student in electrical engineering at the Uni v ersity of Ibn Zohr , Ag adir . He gradua ted as an embedded systems engineer in 2021 from the National School of Applied Sciences, Ag adir , Moroc co. Subsequently , he joined the research group at the Engineering Sciences and Ener gy Management Laboratory , Uni v ersity of Ibn Zohr , Ag adir , Morocco. His research focuses on rene w able ener gies for his doctoral thesis. He can be contacted at email: b .imodane@uiz.ac.ma. M’hand Oubella holds the position of professor in higher education at the High School of T echnologies of Ag adir (E ST A), Ibn Zohr Uni v ersity , Ag adir , M orocco. He obtained his Ph.D. in ener getic and process engineering from the National School of Appl ied Sciences (ENSA) of Ag adir in 2014. Currently , M’hand Oubella is a member of the Laboratory of Engineering Sciences and Ener gy Management (LASIME) at the High School of T echnologies of Ag adir (EST A), and his research focuses on intelligent systems and ener gy management, with a particular e mphasis on rene w able ener gies. This research is conducted within the frame w ork of the research team kno wn as Intelligent Systems and Ener gy Management (ERSIME). He can be contacted at email: m.oubella@uiz.ac.ma. Int J Appl Po wer Eng, V ol. 15, No. 1, March 2026: 69–79 Evaluation Warning : The document was created with Spire.PDF for Python.