Design and Simulations of Solar-Based Hydrogen Production System via Methane Decomposition

System and collector types

For solar-based hydrogen production systems, Dincer and Joshi [5] classified them into four types: photovoltaic, thermal energy, photo-electrolysis, and bio-photolysis. The temperature requirement of methane pyrolysis processes may be divided into three categories. The most common one a lower temperature (500–700 °C) thermochemical path in presence of a catalyst (catalytic decomposition) [5–7]. However other pathways such as thermal decomposition at higher temperatures (1200 °C) and thermal plasma decomposition at much higher temperatures (e.g., >2000 °C for plasma torch in plasma-assisted reforming of methane (PARM)) are also being considered [8].

In the thermal energy-based hydrogen production systems, different solar collector types can be used with varying operating temperatures, concentration factors, and respective power capacities. Solar collector types like flat plate, vacuum-tube, concentrating trough, field mirror, and parabolic were elucidated and then modified by Dincer and Joshi [5]. It was demonstrated that the concentrating solar trough type has a concentration factor ranging from 40–80 with an operating temperature <350 °C compared to a concentration factor of only 1 for flat plate collector with an operating temperature <200 °C. Methane cracking requires a temperature exceeding 1200 °C to break the strong C–H bonds of methane molecules [9].

Yield and economics

Laboratory-scale reactions with solar concentrators were carried out by Abanades and Flamant [10] to examine the decomposition of methane. They found that a nearly complete conversion of CH₄ with a hydrogen yield of about 90% can be achieved. Meanwhile, Abbas and Daud [2] posited that the utilization of ethane versus methane results in an increase of hydrogen production at low temperatures and pressures. Additionally, they found that the addition of a noble gas such as Argon increases the yield of hydrogen at high pressures. In methane pyrolysis experiments conducted by Geißler et al. [11], increasing liquid metal temperatures resulted in increased hydrogen yields, leading to a maximum hydrogen yield of 78% at 1175 °C.

To make solar-based hydrogen production systems commercial and cost effective, Rodat et al. [12] suggested the integration of a carbon sequestration unit. Their economic evaluation showed that the hydrogen production cost can be competitive versus conventional SMR if the carbon black generated is sold. Dincer and Joshi [5] posited that photo-catalytic solar-based water splitting is the most economical and sustainable technology because hydrogen can be obtained directly from abundant and renewable water and sunlight. However, the water used for splitting hydrogen and oxygen often requires treatment and/or desalination, the cost of which was not considered.

Catalytic methane decomposition (CMD)

Methane gas is decomposed as per Equation (1) below:

Using catalysts in methane pyrolysis results in significantly lower temperature (500–700 °C) requirements, making them advantageous from the operational, integrity, and cost perspectives. Recently, Gamal et al. [13] thoroughly discussed catalyst types which can be used for methane cracking including Co-, Fe-, Ni-, Cu-, and C-based catalysts, in addition to self-standing catalysts. There is no consensus in the literature on the best catalyst for production of H₂ using CMD. The use of commercial Ni catalysts is quite common, though it suffers from the possibility of causing coking issues due to the formation, diffusion, and dissolution of C in Ni-alloys. Other alloys containing elements such as Ru, Rh, Pd, Ir, and Pt yield much less coking but come at a much higher cost. Fe-based catalysts may quickly oxidize under reaction conditions, whereas Co cannot withstand the higher process pressures. Therefore, for cost, operational, and process factors, Ni-based catalysts were considered for this work. Further discussions of process-induced catalyst breakdown or catalyst-induced degradation of system materials or surfaces is considered beyond the scope of this work.

Process description and flow

The hydrogen production system designed in this work is through a solar-based CMD process. HITEC molten salt was previous characterized for thermal behaviour in solar linear concentrated technologies by Fenandez et al. [14], and is a eutectic mixture of water-soluble, inorganic salts of potassium nitrate (53 wt% KNO₃), sodium nitrite (40 wt% NaNO₂) and sodium nitrate (7 wt% NaNO₃) [15]. It is initially stored in the “cold” tanks, then is transferred to the parabolic trough solar collectors where the solar radiation is captured, increasing its temperature from 270 °C to around 590 °C. There are four storage tanks for the “hot” molten salt, one of which is kept for night-time supply. Molten salt from any of these tanks enters the decomposition reactor at 590 °C. The other feed into the decomposition reactor is methane gas. The Ni-based catalyst filling the reactor allows the decomposition of methane into H₂ and carbon to occur at around 600 °C. The hydrogen gas leaves the reactor from the top where it is then cooled through a series of heat exchangers and sent to a tank farm for temporary storage and subsequent use in refining units on-demand. Figure 1 shows the process flow diagram for solar-based hydrogen production system. It is noted that the system produces carbon as a by-product, which although could be sellable to external users, will not be discussed further in the present work.

Solar energy collection system selection

A set of 9 selection criteria were considered in a choice matrix. These were: cost of the solar field, peak solar efficiency, annual solar efficiency, concentration ratio, construction requirement, operating temperature, system reliability, land requirement, and thermal storage. Table 1 lists the score of each criterion against the four different solar collection system configurations. A weighting fraction is assigned to each selection criteria. A 0.1 fraction is applied to the scores for the cost of the solar field, peak efficiency, concentration ratio, and operating temperature, since these reasonably contribute to the CMD process. Conversely, land requirement is less significant for refinery sites since they often situated in remote areas such as the desert for UAE. System reliability is a key criterion with a fraction of 0.2, since it is essential to ensure the continuous and reliable production of hydrogen for the refinery. Based on these selection criteria and weights, the parabolic trough solar system was found to be the optimal solar energy collection system for use in the UAE for this work. This choice is supported by related studies by Palenzuela et al. [16] on the engineering and economics of concentrating solar power types for desalination plants, which show that the parabolic trough system provides long-term reliability, adequate concentration ratios and operating temperature projections, and appropriate efficiencies, all whilst maintaining reasonable cost.

Process modelling and simulation

The solar-based hydrogen production system in this work is simulated using the System Advisor Model (SAM) developed by the National Renewable Energy Laboratory (NREL) of the U.S. Department of Energy (DOE) [17]. The modelling process steps in the SAM software began with configuring the location of where the system will be built. In this work, the city of Abu Dhabi in the UAE is chosen, and the solar data was summoned for this location with latitude and longitude of 24.4539 °N and 54.3773 °E, respectively.

Then, the receiver and collector (SCA, solar collector assembly) components were configured. The heat thermal fluid (HTF) was chosen as HITEC molten salt with an operating temperature of 550 °C, required for the CMD process. The factors and parameters considered for selecting HITEC molten salt as the HTF included viscosity, density, heat capacity, and associated pressure loss behaviour. Afterwards, the transport operation limits and loop configuration of the collectors were set up. The number of assemblies per loop is optimizable along with field subsections. The loop thermal efficiency calculation uses the receiver estimated average heat loss, a calculation embedded in the SAM software.

Three cases for SCA makes were studied in this work, as shown in Table 2. Considerations for selecting the collectors included their design weight (low), fasteners requirements, welding or specialized manufacturing requirements, time required for field assembly, extruded parts costs, and mirrors alignment requirements. The values presented by Wagner [18] for HTF factors and parameters were used, and shown in Tables 3,  4, and 5 below.

Table 3

Case 1 major simulation data inputs.

Table 4

Case 2 major simulation data inputs.

Table 5

Case 3 major simulation data inputs.

Cases and calculations

For the three cases discussed below, the design parameters are driven by the weather library values for the city of Abu Dhabi in the SAM software. The direct normal irradiation (DNI) parameter, which is the quantity of solar radiation received per unit area of the collector, is considered as 950 W/m². The density of the HTF (i.e., HITEC molten salt) simulated is 1819.7 kg/m³ and operating temperature is 550 °C, adequate for the CMD process. To avoid the freezing of the HTF, the loop inlet temperature parameter is maintained at temperature higher than the HTF freezing point of 290 °C. Initially, the HTF storage time is set as 14 h. Parallel tank pairs were considered for all cases, each tank with a height of 15 m and a diameter of 9.1 m. Figure 2 illustrates the system design with values for the Case 1 as an example.

Simulation outputs

In this section, the results from the SAM software for the three cases are presented and discussed. Based on the inputs described in the previous sections, Figure 3 shows the annual net energy, annual gross energy, annual thermal freeze protection requirement, and annual electricity load for the three cases considered in this work. As seen in this figure, the annual net energy (kWh) for Case 1 is the highest, followed by Case 2 and then Case 3. Notably, in Case 2 although the annual gross energy (kWh) is higher than in Case 1, the thermal freeze protection requirements are also higher, making the annual net energy lower than Case 1. The levelized cost of heat (LCOH) for the three cases considered in this work is shown in Figure 4. As seen in Figure 4, the LCOH is lowest for Case 1 at around 9.32 ¢/kWh. When evaluated on an annual profile, field protection requirements for the three cases were expectedly higher at the beginning of the year due to the winter condition in the UAE. Then, as the summer season commences, the field freeze protection requirement starts reducing for all three cases. Case 1 exhibits the minimum freeze protection requirement among the three cases, which makes it the optimum choice in terms of LCOH.

Simulation outputs for three of the most important parameters, namely field pressure drop, receiver thermal losses, and field protection (from molten salt freezing) requirements are discussed here. Figure 5 illustrates the field pressure drop in bar for the three cases studied. The profile increases in the summer to around 10 bar. The profile for Case 2 exhibits a nearly identical fashion and very similar values to that of Case 1. However, for Case 3 shown in Figure 5, the pressure drop soared to about 50 bar around halfway through the year. The same significant difference between the first two cases and Case 3 is observed for the receiver thermal losses parameter (MW), shown in Figure 6. As seen in Figure 6, the receiver thermal losses for Case 1 records a maximum of nearly 7 MW in the annual profile obtained from the SAM software for the city of Abu Dhabi. A similar behaviour and nearly identical values are exhibited in Case 2. Contrastingly, the receiver thermal losses in Case 3 reached double that of Cases 1 and 2, at around 14 MW.

As shown in Figure 7, the field protection (from molten salt freezing) requirement for the three cases are expectedly higher at the beginning of the year due to winter condition in the UAE. They peak at values between 2 and 3 MW for all three cases. Then, for all three cases, as summer season commences, the field freeze protection requirements start reducing. Case 1 exhibits the minimum freeze protection requirement among the three cases, which translates to the lowest cost requirement for continued operation throughout the year.

The influence of molten salt type (HITEC models) on the LCOH was also studied. By maintaining the configuration of the Case 1 in terms of collector and receiver types, the HTF type was the adjustment variable. The control HTF studied was HITEC solar salt (specific to CSP applications) with maximum operating temperature capabilities of about 593 °C. This was compared to HITEC XL with a maximum operating temperature of around 500 °C and HITEC standard with a maximum operating temperature up to 538 °C. Based on the LCOH obtained for the three HTF cases, shown in Figure 8, the control case (i.e., HITEC solar salt used throughout this paper) maintained its superiority over the HITEC XL and HITEC standard counterparts based on the highest operating temperature to LCOH (T:LCOH) ratio. Although the LCOH for the HITEC XL is lower than HITEC solar salt, its maximum operating temperature of 500 °C is too low to sustain the CMD process at profitable yields. Furthermore, the T:LCOH ratio of HITEC XL is only 54, compared to 64 for HITEC solar salt.

Regression analysis

Since the pressure drop is one of the main variables affecting the integrity of the solar-based hydrogen production system, it is essential to understand which operating or design variables affect it. Also, the behaviour of pressure drop versus these variables needs to be characterized. To address this concern, a regression analysis was carried out using Minitab to examine the pressure drop response. Initially, ten variables were identified for the regression analysis, namely: field thermal power incident, field averaged outlet temperature, field total mass flow delivered, receiver mass flow rate, receiver thermal losses, receiver thermal power incident, TES (thermal energy storage) freeze protection power, TES thermal losses, resource beam normal irradiance, and resource dry bulb temperature. These variables are denoted x₁ through x₁₀ in the regression Equation (27). These independent variables influence the output variable (also known as the response variable), namely the pressure drop in this analysis which is denoted y in Equation (27). These ten variables are accepted or rejected based on the p-value generated from the regression analysis, with a 5% significance screening indicator, or 𝛼 = 0.05.

The regression equation was obtained with the ten variables for the response of the pressure drop. A total of 8760 data points were considered in this analysis based on the hourly data for each variable. As shown in Figure 9, the majority of data points were within ±2 residuals for the output variable (i.e., pressure drop). Three input variables were rejected as a result of the analysis, namely field thermal power incident, TES freeze protection power, and resource beam normal irradiance, with p-values 0.877, 0.867, and 0.862, respectively. Eventually, only seven variables are considered acceptable as affecting the pressure drop response. However, of these, only the receiver mass flow rate could be considered as a main contributing variable towards the pressure drop behaviour. This is based on the R²-value for this variable being >95%, the rest were <75%. By knowing that the mass flow rate is the main parameter affecting pressure drop, the operator of the system can better control adjustable system inputs to delay or avoid failures of this solar-based hydrogen production system.