Investigating the Potential of Renewable-Hydrogen Energy Storage Systems (RHES) in Enabling Scotland’s Farming Communities Net-Zero Transition and Sizing the Proposed RHES System

Introduction

The implementation of clean energy production from renewables is a key component towards achieving the net-zero target. However, given the intermittent nature of renewable energy systems (RES), energy storage is critical to mitigate this intermittency problem and realise the full potential of renewable energy. Energy storage devices can be classified according to a range of characteristics, including their storage capacity and duration, life expectancy, size, cost and safety, and environmental effect, including their recyclability [1]. There are numerous storage options; these include flow batteries which store energy directly in the electrolyte but are still in their infancy in terms of deployment, sodium-sulfur batteries which have a higher energy density than Li-ion batteries but have an inconvenient hot liquid metal electrolyte partially reducing the battery performance [2], supercapacitors which cannot provide electricity for an extended period of time, and compressed air and flywheels energy storage installations which are restricted by location requirements [3].

Hydrogen energy storage (HES) systems are distinguished from other types of renewable energy storage systems by their adaptability and capacity to deliver multiple services [4]. This quality is essential for grid operators to maintain system dependability and the integration of RES into the electricity, heating, and transportation infrastructures [4, 5]. Energy can be stored at large-scale using HES systems, ranging from 1 GWh to 1 TWh, whereas batteries generally range from 10 kWh to 10 MWh [5]. Fuel cell electric vehicles have additional hydrogen utilisation potential of interest [6]. The degree to which HES systems may enter energy storage markets will rely on a number of variables, including non-technological hurdles such as regulatory, safety, and economic concerns [7]. H₂ energy-storage is an emerging key enabler in Scotland’s decarbonisation plans, and there is a need to demonstrate how its implementation can contribute to meeting the net-zero greenhouse gas (GHG) emission goals. Sizing the capacity of the HES system components needed with renewables and their economic viability represent key research components in assessing the potential of HES in enabling the clean energy transition. To this extent, many researchers have carried out valuable insights into sizing and assessing the feasibility of HES technologies for stand-alone and grid-connected hybrid renewable energy systems. Castaneda et al. [8] have investigated the sizing of HES system components within a stand-alone hybrid renewable-H₂ system using four different methods including: deterministic technique, MATLAB-based technical optimization method using Simulink Design Optimization (SDO) toolbox, and a techno-economic optimization using HOMER and iHOGA software, the results showed satisfaction of the load demand criteria while minimizing the total cost. In [9], a stand-alone PV-Fuel Cell (PV-FC) system has been sized using HOMER for supplying the electrical load demand of a remote village. The objective function was set to minimize the net present cost while maintaining the system reliability. The obtained results have shown that the optimal sizes of system components have achieved a minimal net present cost while satisfying the required load demand. It has been also observed that most of the annual energy produced was utilized for hydrogen generation, reflecting the potential of H₂ storage systems in absorbing the excess of renewable energy, thus decarbonizing the electrification of remote areas. The authors have also compared the results of the proposed stand-alone PV-FC system to a stand-alone PV-batteries system, and it was found that the former is better in terms of net present cost. Bernoosi and Nazari [10] have studied the feasibility of stand-alone PV-Fuel Cell Combined-Heat-and-Power (CHP) system for feeding both electrical and heating load demands. In this work, the contribution of HES has been assessed by studying the system with and without the use of H₂ storage technology, considering the integration of the proposed PV-FC system with batteries and domestic water heater to evaluate the total cost of the system. In each case, the sizes of PV-panels, batteries and water heater were determined, and the obtained results have shown corresponding reduction in the size of water heater and batteries with the use of H₂ storage system. Darei et al. [11] have studied the effect of increasing the size of H₂ electrolyser and fuel-cell on the production and planning of an energy system in Sweden involving rooftop PVs and CHP plants for serving district heating network and electricity demands. The study involved sizing and modelling the proposed H₂ system components and results have shown that increased capacities of electrolyser and fuel-cell have contributed to great reduction in the grid energy purchase and production cost while increasing the heat recovery from H₂ system components and minimizing the load of CHP plants. Further research has been conducted for sizing Power-to-Gas (P2G) systems incorporating H₂ storage facilities for grid balancing services. Simonis and Newborough [12] have investigated the effect of multiple P2G configurations using different electrolyser capacities in conjunction with hydrogen storage buffers for capturing the excess of onshore wind energy in a German city and injecting it into the gas-grid. The study has been carried out based on predicting the excess of renewable electricity in the city during 2015–2020, with respect to the electricity and gas demand profiles. Results have indicated that upscaling the size of P2G system significantly increases the availability of hydrogen generation by electrolyser thus reducing wind curtailment scenarios in the considered region while increasing the contribution of green gas for balancing the gas-grid consumption. McDonagh et al. [13] have investigated multiple scenarios of integrating hydrogen P2G solutions with offshore wind farms from an investor perspective, considering different curtailment levels. A model has been developed to estimate the hourly generation and investor revenue of an offshore wind farm located in the Irish Sea in Ireland using the wind speed data and the electricity market price. Results have demonstrated that combining P2G installation with the considered wind farm allowed the most profitable configuration where the wind power is converted into hydrogen at lower energy rates and sold as electricity at higher energy rates. It has been concluded from the conducted literature that HES systems hold a great potential in decarbonizing multiple sectors, however sizing the capacity of HES systems, planning their operation and assessing their feasibility when integrated with renewables are key enablers for unlocking their full potential and promoting their large-scale development.

This paper investigates the potential of hybrid renewable-hydrogen energy storage systems in enabling the energy transition of Scotland’s farms using Glensaugh farm as a case study for the investigation. At Glensaugh farm, just outside Fettercairn in Aberdeenshire, the HydroGlen Project aims turning the farm into a net-zero carbon emissions farm through feeding the farm’s (electricity, heating, and transport) energy needs from a combination of renewable energy sources and on-site hydrogen production, compression, and storage [14]. Currently, Glensaugh farm has a renewable energy mix composed of a 50-kW wind turbine and 50 kW solar PV and is looking into installing more renewable capacity to meet all their residential, commercial, and transportation demands [14]. The farm’s energy demands data and RINA feasibility report [14] have been provided by the James Hutton Institute (JHI).

This study started by evaluating the farm currently installed renewable system in meeting the farm’s residential, commercial and transport demands. Based on this evaluation, two hybrid renewable-H₂ systems were proposed. Option (1), a Hybrid Solar-H₂ system that utilizes the existing PV together with an extra grid-connected PV capacity that was sized to allow meeting more residential and commercial demands and a H₂ generator that was sized to allow storing the solar surplus in the form of green H₂ to be used for meeting the transport demands. The effectiveness of this proposed and sized hybrid PV-H₂ system was then assessed. For option (2), a hybrid Wind-H₂ system is proposed, sized and assessed. The proposed hybrid Wind-H₂ system utilizes the currently installed wind turbine as it was found capable of meeting most of the residential and commercial demands, together with a H₂ generator that was sized to allow storing the wind surplus in the form of green H₂ to be used for meeting the transport demands. The green H₂ produced in each of the proposed hybrid renewable-H₂ systems is utilized as fuel for meeting the farm transport demands rather than converting the H₂ back to electrical energy for ‘on-grid’ consumption because H₂ fuel is more competitively priced if it is marketed as a high-value gas [15]. The LCOE for each of the proposed hybrid configurations was estimated and they were compared to select the more cost-competitive option. Based on this comparison, a Simulink model was developed for the selected hybrid PV-H₂ system to assess in more detail its operation and potential.

Evaluating Glensaugh Farm current operating renewable capacity

Table 1 shows the data provided for the farm demands [14], where the transport demand is comprised of 12 vehicles accounting for 5110 kWh with no variation throughout the year. The currently installed renewable capacity comprises a mixture of a 50-kW solar PV and 50-kW wind turbine [14]. The on-site 50 kW wind turbine is not included in the calculations as it feeds directly into the national grid, not the farm demands. The RINA report proposes a new, 800 kW turbine to be installed for use in combination with the solar PV.

The energy output of the currently installed 50-kW PV system was calculated using the PVsyst software and results are shown in Table 2. The PVsyst includes extensive meteorological and PV system components databases that allows the user to select the PV system components (module and inverter) [16]. Based on the information provided by the farm owner, the PV and inverter modules were set in PVsyst to JC250M-24/Bb-v and SolarLake 15000-TL respectively, the tilt was set to 20°, and the azimuth, which is the direction the PV is facing, was set at 50°. From the PVsyst simulation results shown in Table 2, it was found that the annual energy output of the currently installed 50-kW solar system can only meet 11% of the residential and commercial total demand (40,316 kWh/364,330 kWh).

Sizing the proposed PV-H₂ system

Given that the currently installed PV capacity is insufficient to meet the farm’s energy demands alone without adding an 800 kW wind turbine as proposed by RINA’s study [14], we here propose an alternative solution of extra PV capacity that allows meeting the farm’s demands by 100% solar PV. The capacity of the proposed PV system was accordingly sized, and its potential was assessed. A hydrogen energy storage system was also proposed and sized to store the excess in the PV production in the form of green hydrogen fuel to be used in clean fuelling of the farm’s 12 vehicles. Figure 1 shows the proposed PV-H₂ Energy System.

3.2.

Sizing the electrolyser (H₂ generator) needed to store the new PV system energy surplus as green H₂

Given the high cost of electrolysers, it is desirable to maximise their utilisation [15]. Based on industrial standards, the electrolyser size is usually chosen to be between 20% and 40% of the PV capacity to increase the electrolyser’s utilisation level. The possible downside is that there will be moments when total renewable generation surpasses the total electricity that the combined load and storage can absorb [15]. Selecting a 40% of the proposed 598 kW PV capacity suggests an electrolyser size of 239.2 kW. To allow better utilization, three units of 80 kW HySTAT 15–10 electrolysers from Hydrogenic manufacturer (highlighted in Figure 2) are accordingly selected.

3.3.

Sizing the hydrogen storage tank

To identify the storage tank size, it is first necessary to identify the amount of H₂ produced by the electrolyser based on the surplus in the solar output from the proposed PV system. Using Pvsyst, the surplus in solar production was estimated by comparing the monthly energy output of the newly sized PV capacity to the farm’s total residential and commercial load demand. The monthly energy output of the PV system was estimated using the PVsyst software after setting the following inputs: the location was set to a Latitude of 57.20° N and a longitude of −2.20° E. PV Tilt angle was set as 37° after examining different tilt values to find the optimal, additionally Fordham [20] proved that the optimal tilt angle of a PV is equal to the site’s latitude minus 20° thus in Scotland it is (57° − 20° = 37°). The azimuth was set at 0° as Scotland is in the northern hemisphere facing south. The PV module and inverter types were selected the same as the ones already installed on the farm, Figure 3 shows the system details. An optimized selection of inverter size is done by the PVsyst.

Table 4 demonstrates the PVsyst simulation results showing the proposed PV system monthly DC energy output and the monthly AC solar energy excess which takes into account the inverter and wiring losses.

Month	Monthly DC production from the proposed 598 kW PV system (kWh)	Monthly AC residential and commercial demand (kWh)	Monthly AC solar energy supplying the demand during the daytime (kWh)	Monthly AC solar energy excess (kWh)
January	19,740	43,810	7,850	11,450
February	29,900	35,970	9,630	19,660
March	54,040	37,980	14,220	38,790
April	65,220	22,330	10,960	53,020
May	82,320	23,480	13,240	67,540
June	74,280	20,860	12,600	60,250
July	73,070	18,530	11,090	60,560
August	63,230	19,280	10,070	51,950
September	54,020	19,320	8,680	44,300
October	32,920	31,580	10,030	22,200
November	19,340	45,110	9,010	9,880
December	13,740	46,100	6,610	6,770
Annual	581,820	364,350	123,990	446,370

Table 4

PVsyst energy output simulation results for the 598 kW PV system.

Based on the PV’s monthly energy surplus, the amount of hydrogen produced monthly by the previously selected and sized electrolyser was calculated by dividing the PV energy excess by the electrolyzer energy consumption of 5.4 kWh/N m³. The monthly hydrogen required for fuelling each vehicle in the farm was also calculated by using the 12 vehicles given total monthly consumption (5110 kWh) and the onboard H₂ fuel cell and sub-systems Round-trip efficiency (RTE) of 30% [21]. Given the RTE, the energy demand of the 12 vehicles becomes 17,033 kWh (5110/0.3), this equates to a monthly demand of 1419.4 kWh for each vehicle [14]. This amount is then converted from 1419.4 kWh to normal cubic meters (N m³ of hydrogen) by multiplying it by the conversation factor 0.333 giving 473 N m³ of H₂ required per month for each vehicle. This conversion factor was calculated based on the fact that at low heat value (LHV), 11.1 N m³ of hydrogen is equivalent to 33.3 kWh [19]. Using the calculated monthly H₂ produced by electrolyser and the H₂ required by each vehicle, the number of vehicles that can be fed by clean green H₂ fuel every month was then calculated. The monthly accumulation of H₂ excess was then calculated and used in fuelling more vehicles. Finally, the residual monthly accumulated H₂ after feeding the farm 12 vehicles was calculated to be sold as a commodity or used in generating clean electricity. Results of all these calculations are shown in Table 5.

Month	Monthly excess in PV energy (kWh)	Electrolyser energy consumption (kWh/N m3)	Monthly H2 Production (N m3)	One vehicle monthly H2 consumption needs (N m3)	No of vehicles that can be supplied each month	Monthly H2 excess after feeding the possible number of vehicles (N m3)	No of extra vehicles that can be supplied from the H2 monthly accumulation	Monthly accumulated H2 after feeding the farm’s 12 vehicles (N m3)
Jan	11,450	5.4	2120.4	473	4	228.4	0	228.4
Feb	19,660	5.4	3640.7	473	7	329.7	1	85.1
Mar	38,790	5.4	7183.3	473	12	1507.3	0	1592.4
Apr	53,020	5.4	9818.5	473	12	4142.5	0	5735.0
May	67,540	5.4	12,507.4	473	12	6831.4	0	12,566.4
Jun	60,250	5.4	11,157.4	473	12	5481.4	0	18,047.8
Jul	60,560	5.4	11,214.8	473	12	5538.8	0	23,586.6
Aug	51,950	5.4	9620.4	473	12	3944.4	0	27,531.0
Sep	44,300	5.4	8203.7	473	12	2527.7	0	30,058.7
Oct	22,200	5.4	4111.1	473	8	327.1	4	28,493.8
Nov	9,880	5.4	1829.6	473	3	410.6	9	24,647.4
Dec	6,770	5.4	1253.7	473	2	307.7	10	20,225.1

Table 5

H₂ accumulation calculation results.

The volume of the H₂ storage tank was then determined based on the maximum amount of accumulated hydrogen, which is 30,058.7 N m³. The volume of hydrogen can be lowered by employing a compressor. Using Boyles’ law (Equation (2)), the new volume of hydrogen following compression with the temperature remaining constant is calculated.

(2)

where, P₁ is the pressure of H₂ output from the selected electrolyser; V₁ is the volume of the maximum accumulated H₂ (30,058.7 N m³); P₂ is the pressure after compression; and V₂ is the volume after compression.

As specified in the HySTAT 15-10 electrolyser specifications, the hydrogen is supplied at a pressure P₁ = 10 bars. The pressure of the selected PURE ENERGY CENTER compressor P₂ is set to 200 bars [22]. Thus, using Equation (2), the volume of H₂ after compression will be:

Figure 4 illustrates the selected type of storage tank for this study: N5.0 (CP Grade H₂) Manifold cylinder pallet of maximum storage capacity 132 m³ and can store hydrogen at 200 bar [23]. To store the amount of H₂ accumulated, 12 storage tanks (12 ∗ 132 = 1584) are required, which accommodate any future expansion.

The Wind-H₂ system

A Wind-H₂ system, as illustrated in Figure 5, is proposed, sized and assessed as option (2). The RINA feasibility study (RINA 2021) identified that an 800 kW Enercon E-53 with a 50 m hub height would be suitable for meeting the farm’s demand. The sufficiency of this proposed wind turbine capacity in meeting the residential and commercial demands is therefore assessed, and based on the excess in the wind production, the best-suited hydrogen energy storage system was accordingly sized.

4.1.

Calculating the kWh output from the proposed wind turbine

Based on the assessment of several potential development areas (PDA) against several criteria (like terrain, wind speed, noise risk, etc.) as seen in Figure 8, the RINA feasibility study concluded that PDA 3 is the most suitable location for installing the new wind turbine capacity [14]. By matching the HydroGlen feasibility study location picture with Google maps, as illustrated in Figures 6 and 7, PDA 3 was found to be at a latitude of 56.913 and a longitude of −2.551.

Using Global wind Atlas, which is a free, web-based application, the wind speed was found by drawing a 3 km by 3 km rectangular on the PDA 3 location using the webpage map [24]. However, the wind speed was normalized as the result display the wind speed index as shown in Figure 8.

The denormalized wind speed was then calculated by multiplying the monthly wind speed index by the location wind speed which was found by the Global wind Atlas to be 9.81 m/s. Equation (3) [25] was then used to transform the obtained monthly wind speed at reference height of 100 m to the equivalent speed at the new wind turbine target height of 50 m as shown in Table 6.

(3)

where, z is the target height (m), z_r is reference height (m), z₀ is characteristic terrain length or roughness (m), U (z) is wind velocity at target height z, and U (z_r) is wind velocity at reference height z_r. The characteristic terrain length of the site (z) was found to be 0.03 by using the global wind atlas software, which matches the fallow field’s characteristic terrain length in [25].

Month	Monthly wind speed index at height 100 m (normalized wind speed)	Location wind speed (m/s)	Monthly denormalized wind speed at height 100 m (U (z)) (m/s)	Monthly wind speed at height 50 m (U (zr )) (m/s)
January	1.21	9.81	11.8701	10.86
February	1.23	9.81	12.0663	11.04
March	1.09	9.81	10.6929	9.78
April	0.87	9.81	8.5347	7.81
May	0.85	9.81	8.3385	7.63
June	0.75	9.81	7.3575	6.73
July	0.78	9.81	7.6518	7.00
August	0.84	9.81	8.2404	7.54
September	0.91	9.81	8.9271	8.16
October	1.03	9.81	10.1043	9.24
November	1.15	9.81	11.2815	10.32
December	1.32	9.81	12.9492	11.84

Table 6

Wind speed calculations at the proposed new turbine height of 50 m.

The wind speed time series (from Table 6) was then merged with the wind turbine power curve (yellow curve) shown in Figure 9 to find the available/theoretical wind power at each speed (P₀), results are shown in Table 7. According to Neill, only part of the available/theoretical power can be harvested by a wind turbine [26]. Therefore, the actual/extracted wind power (P_e) was calculated using Equation (4) [26]:

(4)

where C_p is the power coefficient which is the overall efficiency of a turbine extracted from Figure 9.

Month	Monthly wind speed at 50 m height (m/s)	Power coefficient Cp (using the Cp curve in Figure 9)	Wind power P0 (using the power curve in Figure 9) (kW)	Capacity factor CF	t (hours per month)	Wind production using Equation (5) (kWh)	Monthly residential and commercial demand (kWh)	Monthly wind energy excess after supplying residential and commercial total demand (kWh)
Jan	10.86	0.41	722	0.35	744	77,083.60	43,810	33,273.60
Feb	11.04	0.42	744	0.35	672	73,495.30	35,973	37,522.29
Mar	9.78	0.47	600	0.35	744	73,432.80	37,978	35,454.80
Apr	7.81	0.49	300	0.35	720	37,044	22,330	14,714
May	7.63	0.49	290	0.35	744	37,002.80	23,484	13,518.84
Jun	6.73	0.49	200	0.35	720	24,494.40	20,855	3,639.40
Jul	7.00	0.49	228	0.35	744	29,091.90	18,533	10,558.88
Aug	7.54	0.49	290	0.35	744	37,002.80	19,284	17,718.84
Sep	8.16	0.49	366	0.35	720	45,193.70	19,324	25,869.68
Oct	9.24	0.49	500	0.35	744	63,798	31,548	32,250
Nov	10.32	0.46	672	0.35	720	77,898.20	45,111	32,787.24
Dec	11.84	0.36	790	0.35	744	74,057.80	46,100	27,957.76
Annual						649,595.35	364,330	285,265.35

Table 7

Monthly wind energy production and monthly wind energy excess after feeding the residential and commercial total demand.

The wind output energy was then calculated using Equation (5) [26]:

(5)

where C_F is the capacity factor that describes the relation between the turbine’s power output and the maximum power, P_e is actual wind power in kW, and t is number of hours per month. In optimal conditions, a well-designed wind turbine may attain a maximum performance level (C_F) of 35% [28, 29]. Finally, the monthly excess in wind production after feeding the residential and commercial total demand was calculated and results shown in Table 7.

LCOE of PV-H₂ and Wind-H₂ systems

The levelized cost of energy (LCOE) was calculated for each of the proposed systems. The LCOE for a renewable source with hydrogen storage is given by Equation (6) [15]:

The CAPEX and OPEX of the PV-H₂ system, calculated by adding the costs of all units in the system as shown in Table 10, was found to be £1,332,328 and £5279.59 respectively.

The CAPEX and OPEX of the Wind-H₂ system, calculated by adding the costs of all units in the system as shown in Table 11, was found to be £2,055,894 and £43,141 respectively.

To calculate the LOCE of PV-H₂ and wind H₂ system, the E_{Res, t} (kWh) and E_{H 2, t} (kWh) (the renewable energy output utilised in meeting demands and the hydrogen production) were first calculated for each of the two systems based on the results in Tables 4 and 7.

The above values were then substituted in Equation (4) along with the CAPEX and OPEX to find the LCOE for each of the proposed systems which were found to be 0.3 £/kWh for the PV-H₂ system and 0.4 £/kWh for the Wind-H₂ system.

Analysis of the proposed hybrid PV-H₂ and Wind-H₂ systems

From Table 4 results, it can be found that the proposed PV system can meet just 35% of the load demand (123,990 kWh/364,350 kWh) because most of the demand is not during the sun availability. Only 21% (123,990 kWh/581,820 kWh) of the total solar energy is being utilized, resulting into around 77% (446,370 kWh/581,820 kWh) of excess in solar energy production (see Table 12). This solar energy excess is stored in the form of green H₂ to be used as clean fuel for the farm’s vehicles. On analyzing Table 5 results, it can be found that the proposed HES system was able to meet all the 12 vehicles’ transportation demand from March until September. From September to December, all the transportation demand was met after using the stored accumulation of hydrogen from previous months. Although not all the 12 vehicles were supplied in January and February, the accumulated hydrogen after one year will be sufficient to cover these months in the following year. Therefore, it can be concluded that PV-H₂ system is almost capable of meeting 100% of the transport fuel needs by green H₂. Furthermore, after supplying all vehicles, there are still extra H₂ that can be either sold as a commodity or can be converted back to electricity using a fuel cell to feed more residential and commercial demands during the lack of solar energy thus minimising grid imports.

In comparison, it can be concluded from Tables 7 that the proposed Wind-H₂ system can meet almost 100% of the farm’s residential and commercial demands by wind energy, and 44% (285,265.35 kWh/649,595.35 kWh) of the wind production is converted to H₂ for vehicles fuelling.

Figure 10 was then constructed using columns 3 and 4 from Tables 4 and column 7 from Table 7 that represent the monthly solar production, the monthly residential and commercial demands, and the monthly wind production respectively to investigate the ability of the PV-H₂ and the Wind-H₂ in meeting residential and commercial demands while producing green H₂ fuel from the excess to powering the transport demands. It can be seen that the wind generation profile closely matches the monthly residential and commercial demand, given the higher demand in the winter and lower demand in the summer. On the other hand, the Solar profile can be seen out of synchronism with the residential and commercial demand but allows plenty of green H₂ production to be used as fuel for meeting the transport needs as well as being sold as commodity.

From Table 10, it was found that LCOE for the PV-H₂ system is 0.3 £/kWh lower than that of the Wind-H₂ system of 0.4 £/kWh. This implies that the PV-H₂ system is more cost competitive than the wind-H₂ system at current prices, which might be attributed to the wind turbine’s high capacity and O&M costs. Reducing the CAPEX and/or increasing the round-trip efficiency of the hydrogen system in the future will allow reducing the LCOE of renewable-H₂ systems to become more financially competitive with other technologies such as natural gas, which is 213 US/MWh [39], this equates to 0.16 £/kWh.

Developing a MATLAB/Simulink model for the proposed PV-H₂ system

In order to assess in more detail, the potential of the PV-H₂ system (given that a PV-H₂ system was not considered by RINA), a Simulink model is developed in this section to simulate the generation of the proposed PV system and its overall utilization in the electrolyser to produce green H₂. An electrochemical model has been used to model the electrolyser’s green hydrogen production. The developed electrochemical model gives more accurate results as it considers the hydrogen production as function of the current output from the proposed PV system modelled using MATLAB/Simulink, as seen in Figure 11.

The amount of hydrogen produced by the electrolyser is also calculated using the electrolyser’s energy consumption calculation method in order to compare the result to the Simulink model results.

7.1.

Modelling the proposed PV array

The proposed 598-kW photovoltaic capacity was modelled using the MATLAB/Simulink PV Array block. The PV Array block is a five-parameter model that employs a light-generated current source (IL), a diode, series resistance (Rs), and shunt resistance (Rsh) to simulate the modules’ irradiance and temperature-dependent I–V characteristics, as seen in Figure 12. The diode I–V characteristics for a single module are defined by Equations (7) and (8) [40]:

(7)

(8)

where, all the symbols in Equations (7) and (8) are defined in Table 13.

Id	Diode current (A)
VT	Thermal voltage (V)
Vd	Diode voltage (V)
I0	Diode saturation current (A)
nI	Diode ideality factor, a number close to 1.0
k	Boltzmann constant = 1.3806 × 10−23 J⋅K−1
q	Electron charge = 1.6022 × 10−19 C
T	Cell temperature (K)
Ncell	Number of cells connected in series in a module

Table 13

Definitions of PV array equation.

The PV array block was then modified to model the proposed 598 kW solar capacity. The module was set as RenSola America J250M to be similar to the farm currently installed PV. The PV array block’s module parameters were then set to correspond to the PV module parameters listed in the PV datasheet [41]. The parallel strings and series modules were set to 104 and 23, respectively, as determined from the PVsyst simulation results shown in Figure 3. Based on this, the following values resulted for the output voltage and current:

(9)

(10)

where I_mp in (A) and V_mp in (V) denote respectively the current and voltage at the maximum power points obtained from the module cell parameters shown in Figure 13. Hence, the maximum power of the PV array is:

(11)

Figure 14 shows the output I–V characteristics of the modelled PV array block. It can be seen that, at Standard Test conditions (1000 w/m², Cell Temperature 25° C), the array output simulates the 598 kW capacity.

To ensure the PV array provides maximum power at all times, an MPPT (Maximum power point tracking) controller with incremental conductance technique block was integrated with the PV array to account for variables such as fluctuating irradiance (sunlight) and temperature [42]. The file exchange of the MPPT block can be found here [43].

7.2.

Electrolyser modelling

An Alkaline electrolyzer electrochemical model, as shown in Figure 15, is developed in this section.

7.2.1.

Hydrogen production block

The production rate of hydrogen in an alkaline electrolyser is related to the input current as given by Equation (12) [44]:

(12)

where, nH₂ is the molar flow rate (mol s⁻¹); 𝜂F is the Faraday efficiency; z is 2 (number of electrons transferred per reaction); I is the current (A); F is the Faraday constant 96,485 C mol⁻¹, and nc is the number of series cells in electrolyser cell stack. Since three electrolysers are used (refer to Section 3.2), then nc is 3. Faraday efficiency is given by the Equation (13) [44]:

(13)

where, A: is the electrode area (m²) = 0.25; I: is the current (A); f₁ is Faraday efficiency parameter (mA² cm⁻⁴) = 150; f₂ is the Faraday efficiency parameter number between 0 and 1 = 0.99, and f₁ and f₂ are selected empirically [44].

The flow rate obtained from Equation (12) is then converted from moles to N m³/h to facilitate comparison in the results section. The H₂ production in N m³/h (Q) is given by Equation (14) using the ideal gas volume V_std of 0.022414 m³∕mol [44].

(14)

A MATLAB/Simulink Hydrogen Production block was then developed as seen in Figure 16 using Equations (12) to (14).

7.2.2.

Electrolyser V-I model

However, the H₂ production rate obtained by using Equation (12) is independent of the input voltage which is incorrect since water can only split into hydrogen and oxygen when the cell voltage is sufficient to activate this [45]. Thus, the cell voltage is expressed as shown in Equation (15) [44].

(15)

where, V_cell is the cell voltage (V); V_rev is the reversible cell voltage (V); r_1,2 are the empirical ohmic resistance parameters of electrolyte (𝛺m²); T is the temperature (K); t_1,2,3 are the empirical overvoltage parameters of the electrode (mA⁻¹ m²); s is the overvoltage parameter of the electrode (V); A is the electrode area (m²), and I is the current (A). All parameters used in Equation (15) are listed in Table 14 and were obtained from Oystein Ulleberg [44].

Symbol	Description	Unit	Value
A	Area of electrode	m2	0.25
S	Overvoltage parameter of electrode	V	0.185
t1	Empirical overvoltage parameter of electrode	A−1 m2	1.002
t2	Empirical overvoltage parameter of electrode	A−1 m2 °C	8.424
t3	Empirical overvoltage parameter of electrode	A−1 m2 °C2	247.3
r1	Electrolyte ohmic resistive parameter	𝛺m2	8.05 × 10−5
r2	Electrolyte ohmic resistive parameter	𝛺m2 °C−1	−2.5 × 10−7

Table 14

Alkaline electrolyser constant parameters for Equations (4)–(10) [44].

V_rev was then calculated as 1.207 V at an ambient Temperature of 40° C (from datasheet) using the Equation (16) [44]:

(16)

Equation (15) was then used to develop the Electrolyser V-I block components, as shown in Figure 17.

Finally, the boundary condition was added to the developed Hydrogen Production modelling block to allow hydrogen production only when cell voltage is greater than V_rev. The complete electrolyser model is shown in Figure 15.

7.3.

The MATLAB/Simulink PV-H₂ model results and discussion

To simulate the electrolyser’s hourly H₂ output, the hourly solar irradiance and hourly ambient temperature data are required as input to the developed PV array block. To obtain this data, the PVsyst software was used to generate this hourly data for the 598 kW PV system. PVsyst generates various types of irradiance data, including global irradiation in the horizontal plane (GlobHor), global irradiation in the collector plane (GlobInc), and “Effective” global irradiation on collectors. GlobInc was the one utilised as input for the Simulink PV array block because refers to the total irradiance received (“viewed”) by the tilted plane [16].

Table 15 demonstrates some of the findings of the estimated H₂ output derived based on the electrolyser’s energy consumption (5.4 kWh/N m³), versus the estimated H₂ output from the developed Simulink model which is based on the electrolyser electrochemical model.

Month/day/hour	Hourly solar irradiance (W/m2) (GlolbInc using PVsyst)	Hourly T Amb (°C) (using PVsyst)	Hourly DC power from PV (kW) (using PVsyst)	Hourly DC PV power (kW) (using Simulink)	Estimated H2 output based on the electrolyser’s energy consumption and the PVsyst hourly excess power (N m3 /h)	Estimated H2 output from the developed Simulink model (N m3 /h)
Jan/31/13:00	74.38	10.5	41.07	44.75	7.60	8.05
Feb/28/13:00	564.91	12.86	311.54	358	57.69	60
Mar/15/13:00	522.50	8.54	296.02	337	54.81	56.17
Apr/15/13:00	298.69	8.47	171.25	186.32	31.71	34.25
May/15/13:00	719.95	14.01	389.71	435	72.17	79
Jun/15/13:00	683.62	17.35	366.75	421	67.91	75
Jul/15/13:00	384.51	24.90	215.43	227	39.89	42
Aug/15/13:00	790.33	18.79	414.15	417	76.69	87
Sep/15/13:00	326.98	15.7	182.96	201.75	33.88	35.85
Oct/31/13:00	159.85	8.25	91.13	97.5	16.87	17.66
Nov/30/13:00	303.96	7.28	171.72	191	31.8	34.05
Dec/15/13:00	42.76	7.5	21.81	22	4.04	4.5
Dec/31/13:00	192.39	9.1	107.701	119.6	19.94	21.9

Table 15

The estimated H₂ output from electrolyser using the electrolyser’s energy consumption versus that using the developed Simulink model.

The effect of the irradiance and temperature variations on the solar power generation and H₂ production were investigated over different hours of the years. Figure 18 shows the solar power produced from the proposed PV system when using PVsyst simulation versus the Solar power production on using the Simulink model; it can be observed that solar production in summer days is higher than that of winter days as irradiance decreases. It can also be observed that the estimated power generation on using Simulink is higher than that using PVsyst, and this is due to the fact that Simulink model does not account for losses such as soiling loss and wiring loss.

To minimize the number of parameters involved in the electrolyser simulation analysis, a simpler Faraday efficiency equation with non-temperature-dependent coefficients (Equation (13)) was utilized. A deeper examination of the hydrogen generation results on accounting for the varying electrical current (the orange bar in Figure 19) reveals that the previously predicted H₂ based on based on the electrolyser’s energy consumption (5.4 kWh/N m³) correspond well with the modelled one and thus validate the calculations undertaken in the design process. However, the H₂ generated by the electrolyser electrochemical model was found slightly more than that produced based on the electrolyser’s energy consumption (calculation method), this is because the Simulink model does not account for losses such as soiling loss and wiring loss.

Conclusion

Two combinations of Renewable-H₂ energy systems were proposed, sized and assessed in this paper to identify the scenario that meets most of Glensaugh farm residential and commercial demands with green energy as well as providing green H₂ fuel for the farm transport demand. It was found that the proposed grid-connected PV-H₂ system is capable of feeding almost 100% of Glensaugh transportation fuel requirements with green hydrogen and 35% of Glensaugh residential and commercial demands with clean solar energy, with the gird meeting the remaining demands. The proposed wind-H₂ system was found capable of meeting most of the residential and commercial demands by clean wind energy in addition to around 44% of the transport demand by green H₂.

The results obtained for the PV-H₂ system is due to the fact that most of the residential demands are during evenings resulting into a lot of solar daytime energy converted to green H₂ for powering the vehicles. The H₂ accumulated after feeding all vehicles could be either sold as a commodity or converted back to electricity by using a fuel cell to feed the residential and commercial demands during the shortage of solar energy, thus improving the overall system efficiency. The results obtained for the wind-H₂ system, on the other hand, is because the wind energy profile is very closely matched with the residential and commercial demand, and thus most wind energy is consumed by this demand, leaving only a small amount of wind energy excess to meet the green H₂ fuel transport demands.

It was also found that the levelized cost of energy of the proposed PV-H₂ system is 0.3 £/kWh, more cost competitive than that of the wind-H₂ of 0.4 £/kWh. On the other hand, the reduction in carbon footprint achieved on using Wind-H₂ system was found higher that of PV-H₂ system. Given that this paper is focusing on assessing the PV-H₂ system, a Simulink model was developed for the PV-H₂ system, and it utilized an electrolyser electrochemical model to model the system green hydrogen production.

Implementing the proposed Renewable-H₂ systems in Scottish farms will provide an excellent opportunity in maximizing the implementation of green energy and green H₂ to meet the current Scottish Government’s goal of reaching at least 5 GW of renewable and hydrogen generation by 2030 and at least 25 GW of hydrogen production by 2045 [38]. Developing and promoting such ecologically sustainable green concepts will be a crucial step in transforming Scottish Farms into an energy-efficient and environmentally sustainable communities.

For future work, it is recommended to investigate a renewable energy mix scenario, with wind and solar employed together to reduce the grid import and meet the transport fuel demand by green H₂. Wind energy may be sized to meet the commercial and residential demand since its generation profile closely matches the monthly consumption. At the same time, a solar-hydrogen system could be employed to meet the transportation demand fuel needs and the excess in green H₂ to be sold as a commodity. Furthermore, the system can be investigated with a grid connection import/export capacity to facilitate additional revenue through grid export.

It is also recommended to investigate the sale of the O₂ produced by the electrolyser as a commodity to increase the system economic efficiency (overall competitive value) of the system.

To allow investigating a large number of data such as examining the variation of irradiance every hour of the year, it is suggested to use a strain generator rather than a constant block to allow input a series of irradiance data. This could be more efficient. Further research into the impact of thermal transients in electrolysers could also be investigated.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgements

The data and feasibility reports used for this case study have been provided by the James Hutton Institute (JHI).