Enriched methane production using solar energy: an assessment of plant performance

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Enriched methane production using solar energy: anassessment of plant performance

Marcello De Falcoa,*, Alberto Giaconiab, Luigi Marrellia, Pietro Tarquinib,Roberto Grenab, Giampaolo Caputob

aDepartment of Chemical Engineering, Materials and Environment, University of Rome ‘‘La Sapienza’’, via Eudossiana,

18, 00184 Rome, ItalybENEA Research Center ‘‘Casaccia’’, via Anguillarese, 301, 00123 Rome, Italy

a r t i c l e i n f o

Article history:

Received 7 August 2008

Received in revised form

20 September 2008

Accepted 20 September 2008

Available online 18 November 2008

Keywords:

Enriched methane

Solar energy

Hydrogen production

CSP plant

Molten salt

a b s t r a c t

A novel hybrid plant for the production of a mixture of methane and hydrogen (17 vol%)

from a steam-reforming reactor whose heat duty is supplied by a concentrating solar

power (CSP) plant by means of a molten salt stream is here presented.

It has been demonstrated in the literature that mixtures containing up to 17 vol% of

hydrogen in a natural gas (NG) stream can be sent in the NG pipeline infrastructure and can

supply NG–ICE propulsion systems reducing the GHG emissions and improving the engine

efficiency.

A rigorous two-dimensional model has been developed to investigate the behaviour of the

steam-reforming reactor and the effects of some important operating conditions, such as

the reactant mixture residence time, the steam-to-carbon ratio in the feedstock, the

operating pressure and the inlet temperature.

The final outcome of this work is the dimensioning of a solar enriched methane plant for

the electricity and gas requirements of domestic users.

ª 2008 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights

reserved.

1. Introduction

The well-known world energetic matter, mainly due to the

worldwide growing energy consumption gone with a reduc-

tion of oil and gas availability, and to the environmental

effects of the indiscriminate use of fossil fuels in our economy,

is leading to an increasing interest on hydrogen as energy

carrier. Hydrogen could have the potential to reshape the

entire energy industry since it can be produced from renew-

able energy sources and it can be consumed in fuel cells,

producing electricity without local air pollutants or green-

house gas (GHG) emissions. On the other hand, the hydrogen

technology is not yet ready for a real commercial breakthrough

in a ‘‘hydrogen economy’’ seeing that production methods,

storage and end-use technologies suffer for high cost and, in

some case, low efficiency. Therefore, it seems evident that the

transition towards a hydrogen economy will pass through the

use of hybrid technologies, immediately applicable and

leading to important benefits in terms of reduction of GHG

emissions thanks to a partial use of renewable energy sources.

The aim of this work is to present a novel hybrid plant for

the production of natural gas enriched with hydrogen (HCNG),

very similar to Hythane�, a mixture of methane and hydrogen

(20 vol%), trademarked by Hydrogen Consultants Inc.

In this work the hydrogen concentration in the natural gas

(NG) has been limited to 17 vol%. Natural gas containing

* Corresponding author. Tel.: þ39 06 44585704.E-mail address: [email protected] (M. De Falco).

Avai lab le at www.sc iencedi rect .com

journa l homepage : www.e lsev ie r . com/ loca te /he

0360-3199/$ – see front matter ª 2008 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.ijhydene.2008.09.085

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9

Author's personal copy

17 vol% of hydrogen, here referred as HCNG-17, can be can be

transported, stored and used with the actual gas infrastruc-

ture. Accordingly, HCNG-17 can be sent into the medium or

low-pressure NG grid, immediately after the pressure-reduc-

tion stations [1]. In fact, no compressors are used in the

medium or low-pressure distribution grid, which facilitates

the use of the pipeline infrastructure for hydrogen transport.

Furthermore, using HCNG-17 the well-known hydrogen

storage problems are avoided since standard and currently

available storage systems for compressed NG are adaptable to

the ‘‘enriched’’ NG at low hydrogen content.

HCNG mixtures can also supply natural gas powered

internal combustion engines (NG-ICE): a number of papers,

appeared in the scientific literature, claim that increasing

hydrogen content in the NG engine allows BSFC, BSCO2, BSCO,

BSHC to be reduced, improving the engine efficiency and

reducing the pollutants’ emissions [2–5]. Since the higher

flame temperature can lead to an increase of NOx emissions,

an efficient catalytic converter may be required.

The proposed technology for HCNG production is based on

a consolidated production method such as steam methane

reforming (SMR), powered by solar heat by means of a prom-

ising, widely tested and pre-commercial technology that

makes use of molten salts as heat transfer fluid.

The SMR process is today the most important commercial

massive hydrogen production route; it is based on the

following two reactions:

CH4 þH2O4COþ 3H2 DH0298 K ¼ 206

kJmol

; (1)

CO ¼ H2O4CO2 þH2 DH0298 K ¼ �41

kJmol

; (2)

which, together, yield:

CH4 þ 2H2O4CO2 þ 4H2 DH0298 K ¼ 165

kJmol

; (3)

Steam reforming reactions are very fast over Ni-based

catalyst, so that equilibrium conditions are quickly reached;

Nomenclature

BSCO brake-specific production of carbon monoxide

BSCO2 brake-specific production of carbon dioxide

BSFC brake-specific fuel consumption

BSHC brake-specific production of unburned

hydrocarbons

cCH4 ;0 inlet methane concentration (kmol/m3)

ci i component concentration (kmol/m3)~ci the dimensionless i component concentration

cp,mix gas mixture specific heat (kJ/kgK)

cp,MS molten salt specific heat (kJ/kgK)

CSP concentrating solar power

ctot gas mixture concentration (kmol/m3)

Der effective radial mass diffusivity (m2/s)

dp equivalent particle diameter (m)

f friction factor

FCH4 ;0 inlet methane flow-rate (kmol/s)

FCH4 ;ex outlet methane flow-rate (kmol/s)

Ftot inlet total gas mixture flow-rate (kmol/h)

G superficial mass flow velocity (kg/sm2)

DH0298 K standard reaction enthalpy (kJ/mol)

(�DHj) heat of reaction j (kJ/mol)

HGNG natural gas containing hydrogen

HGNG-17 natural gas containing 17 vol% of hydrogen

hMS heat transport coefficient in the molten salt side

(kJ/sm2K)

hW wall-to-fluid heat transport coefficient (kJ/sm2K)

kmet metal tube conductivity (kJ/smK)

L reactor length

LHV low heat value (kJ/mol)

nreformers number of tubes in configuration shown in Fig. 3

NG natural gas~P dimensionless pressure in the reaction zone

P0 inlet pressure in the reaction zone (bar)

Pemr radial mass Peclet number (uzdp/Der)

QCH4 yearly pro-capite household consumption of

methane (m3/y)

QE.E. yearly pro-capite household consumption of

electricity (kWh/y)

Qen�CH4 yearly pro-capite household consumption of

enriched methane (m3/y)

qr heat flux from the external source to the reactor

packed bed (kW/m2)~r radial dimensionless coordinate

ri rate of reaction of the component i (kmol/skgcat)

Ri tube internal radius (m)

rj rate of reaction j (kmol/skgcat)

R0 tube external radius (m)

S/C steam-to-carbon ratio

tmet metal tube thickness (m)

TMS molten salt temperature (K)~TMS molten salt dimensionless temperature

TMS,in molten salt inlet temperature~TR dimensionless reaction temperature

TR,0 inlet reaction temperature

U overall heat transport coefficient between the

external energy source and the reaction bed (kJ/

sm2K)~uz dimensionless gas mixture velocity

uz,0 inlet gas mixture velocity (m/s)

XCH4 methane conversion

W/F gas mixture residence time (kgcats/mol)

wMS molten salt mass flow-rate (kg/s)~z axial dimensionless coordinate

Greek symbols

3 void fraction of packing

h effectiveness factor

ler effective radial thermal conductivity (kJ/smK)

mg gas mixture viscosity (kJ/sm)

rB catalytic bed density (kg/m3)

rcat catalyst density (kg/m3)

rg gas density (kg/m3)

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9 99

Author's personal copy

a significant hydrogen yield is achieved only at high temper-

atures (850–900 �C). This is the reason why SMR reactors are

usually placed inside furnaces that supply the high heat flux

required to get high methane conversion.

However, if just a 17 vol% content of hydrogen is requested,

a lower operating temperature is needed to limit methane

conversion to about 5%, that is sufficient to satisfy the speci-

fication on the enriched methane composition. This leads to

the possibility of efficiently matching the steam reforming

process with solar-derived heat available at temperatures

lower than 600 �C.

The concentrating solar power (CSP) plant basically consists

of a solar collector field, a receiver, a heat transfer fluid loop;

a suitable heat storage system is also required to maximize the

‘‘capacity factor’’ (i.e. productivity) of the solar plant, and to

provide solar heat at the desired rate regardless the instanta-

neous solar radiation availability and fluctuations [6,7]. The

mirrors of the solar field concentrate the direct solar radiation

on the solar receiver set at the focal point (if point concentra-

tors are adopted) or at the focal line (if linear concentrators are

used). The heat transfer fluid removes the high temperature

solar heat from the receiver and it is afterwards collected into

an insulated heat storage tank to be pumped, on demand, to

the heat users where it releases its sensible heat. Finally, the

heat carrier fluid is stored into a lower temperature tank ready

to restart the solar heat collection loop. A proper dimensioning

of the heat storage system allows to drive the process 24 h/24 h

in continuous at the designed working conditions.

Recently, some molten nitrate mixtures at temperatures

up to 550 �C have been positively tested as convenient, cost-

effective and environmental friendly heat transfer fluid and

storage medium for CSP plants [7–12].

Normally, the molten salt sensible heat is used to generate

high pressure steam to be sent to a steam turbine Rankine

cycle for the production of electrical energy.

However, the molten salt temperature of 550 �C seems to

be suitable for the enrichment of a methane stream by

producing the hydrogen through the SMR process. The

proposed process scheme is shown in Fig. 1. Solar energy

stored in the molten salt can power the hydrogen production

process (reformer heat duty, reactant steam generation, pre-

heating of the reactants), and the residual heat can be used for

the electrical energy production. By this way, a co-generation

plant can be developed, able to produce an HCNG stream (to

be sent in the NG pipeline or to be stored for the NG-ICE

application), and clean electrical energy.

In the present paper, the behaviour of such a plant is

simulated by modeling the SMR reactor and the heat exchange

phenomena between the molten salt stream and the packed

bed reactor. A reformer able to produce an HCNG-17 stream is

designed, and the electrical power output by the plant is

evaluated.

Finally, the above achievements are extended to the design

of a solar plant able to supply enriched methane and to fulfil

the electrical energy requirements (for example) of a small

Italian town. Some technical issues are discussed as well.

2. Process description

Fig. 2 shows a scheme of the solar SMR plant for the produc-

tion of enriched methane.

The reactant mixture, essentially composed of methane

and steam, is the feedstock of a heat exchanger-shaped steam-

reforming reactor (Fig. 3), in which the 550 �C molten salt

stream, heated up by solar energy in the CSP plant, is sent into

the shell. Reactions (1)–(3) generate a gas mixture composed of

CH4, H2O, H2, CO and CO2. The CO is converted into CO2 and H2

in a WGS reactor operating in two-steps (high and low

temperature), and the final abatement of the residual CO traces

Fig. 1 – Simplified scheme of CSP plant and molten salt loop, with two-tank heat storage, coupled to SMR and electrical

power generation plants.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9100

Author's personal copy

is made by a PROX reactor. The steam is separated by

condensation, while the CO2 content is eliminated by absorp-

tion in a typical regenerative MDEA unit. The final stream

leaving the process is composed of methane and a content of

H2 depending on the reformer performance; this mixture can

be sent to the NG grid or to a compression station for storage.

The molten salt stream supplies the heat required by

reforming reactions, process steam generation and reactants

pre-heating; the residual sensible heat is used to generate

further steam to be sent to a steam turbine for electricity

production (Fig. 1).

The core of the process is the steam-reforming reactor

heated up by the molten salt. Due to the 17 vol% hydrogen

content required in the outlet gas mixture, methane conver-

sion in the reactor has to be about 5%. The physical–chemical

phenomena inside the reaction environment, and the heat

exchange between the molten salt stream and the reactant

mixture, are taken into account in the two-dimensional

mathematical model described in the next section.

3. Mathematical modeling

The methane steam reformer heated up by the molten salt

stream is modelled in detail by a two-dimensional mathe-

matical model based on mass, energy and momentum

balances with the intrinsic kinetic equations reported by Xu

and Froment [13].

The following assumptions have been made:

� steady-state conditions;

� negligible axial dispersion and radial convection;

� ideal gas behaviour;

� a single tube representative of any other tube;

� pseudo-effectiveness factors h1, h2 and h3 for the reactions

(1)–(3) independent of local conditions and fixed at 0.02 as an

average value of those reported in the literature [14,15].

The reactor is schematized as shown in Fig. 4.

Equations of the packed bed reactor, together with

boundary conditions, are:

3.1. Mass balances

v�~uz~ci

�v~z

¼ dpL

PemrR2i

v2�~uz~ci

�v~r2 þ 1

~r

v�~uz~ci

�v~r

!� h:rcatð1� 3ÞL

uz;0cCH4Or1 (4)

where Ri and L are the internal radius and the length of the

reformer, c, ~z and ~r are the axial and radial dimensionless

coordinates ð~z ¼ z=L;~r ¼ r=RiÞ, ~uz and ~uz;0 are the dimensionless

Fig. 2 – Process layout.

Fig. 3 – Heat exchanger-shaped steam reformer reactor. Fig. 4 – Schematization of the reactor.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9 101

Author's personal copy

gas velocity and its inlet value, ~ci and cCH4 ;0 the dimensionless i

component concentration and the inlet methane concentra-

tion, ri is the reaction rate of the component i according to Xu-

Froment kinetics model [13], h is the effectiveness factor, rcat is

the catalyst density, 3 is the bed void fraction, dp is the catalyst

particle diameter.

Pemr¼ uzdp/Der is the mass effective radial Peclet number

calculated according to Ref. [16].

3.2. Energy balance

v~TR

v~z¼ lerL

ðuzCtotÞcp;mixR2i

v2 ~TR

v~r2 þ1~r

v~TR

v~r

!

þrcatð1� 3ÞL

P3j¼1 hj

�� DHj

�rj

ðuzctotÞcp;mixTR;0(5)

where ~TR and TR,0 are the dimensionless and the inlet reactor

temperature, cp,mix is the gas mixture specific heat, hj,(�DHj)

and rj are the effectiveness factor, the enthalpy and rate of the

reaction j.

The effective radial thermal conductivity, ler, i.e. the

thermal conductivity of the pseudo-homogeneous phase (gas

mixtureþ solid catalyst particles) that allows for the radial

dispersion, is calculated according to Ref. [17] that, although

concerning pseudo-homogeneous models, result widely suit-

able for the simulation of packed bed reactor heat transport.

3.3. Energy balance in the molten salt shell

d~TMS

d~z¼ � UL

wMScp;MS

�~TMS � ~TR

�2pRi (6)

where ~TMS, wMS and cp,MS are the dimensionless temperature,

the mass flow-rate and the specific heat of the molten salt

stream. The overall heat transport coefficient U is calculated as:

U ¼�

1hMS

Ri

R0þ tmet

kmetþ 1

hw

��1

(7)

where R0 is the external radius of the reformer tube, hMS is the

heat transport coefficient in the molten salt side given by [18],

tmet and kmet are the metal tube thickness and conductivity,

respectively; hw is the wall-to-fluid heat transport coefficient,

evaluated as the usual heat transport coefficient of an

unmixed layer near the wall [19] and evaluated according to

Ref. [20].

The one-dimensional nature of the energy balance in the

heating fluid zone is assumed because of the negligible radial

temperature profiles that are foreseeable.

3.4. Momentum balance

d~Pd~z¼

fGmgL

rgd2pP0

ð1� 3Þ2

33(8)

where ~P and P0 are the dimensionless and the inlet pressure in

the reaction zone. The friction factor f is calculated by the

well-known Ergun equation. The one-dimensional nature of

the momentum balance depends on plug-flow assumption

that imposes the same gas velocity in every point of the

reactor section.

The boundary conditions needed to solve the PDE set of

Eqs. (4)–(6) and (8) are:

� Conditions on inlet section�

~z ¼ 0;c~r�; (9)

~uz~cCH4¼ 1

~uz~ci ¼uzci

uz;0cCH4 ;0ði ¼ H2O;H2;CO;CO2Þ

~TR ¼ 1

~P ¼ 1

~TMS ¼TMS;in

TR;0Co-current configuration

� Conditions on outlet section�

~z ¼ 1;c~r�

~TMS ¼TMS;in

TR;0Counter-current configuration; (10)

� Conditions on reformer tube wall�

~r ¼ 1;c~z�; (11)

v�

~uz~ci

�v~r

¼ 0

lerv~TR

v~r¼ qrRi

TR;0¼ URi

TR;0

�TMS � TR=Ri

�

where qr is the heat flux from the molten salt to the reactor

packed bed.

� Conditions on reformer tube axis�

~r ¼ 0;c~z�; (12)

v�

~uz~ci

�v~r

¼ 0

v~TR

v~r¼ 0

The two-step WGS reaction system is not modelled in detail,

since the traditional and consolidated technology is applied to

the process scheme. From industrial data, WGS reactor is

assumed to convert more than 99% of the CO present in the

stream coming from the reforming reactor.

The Preferential Oxidation reactor (PROX) is inserted in

order to eliminate the CO residue: the PROX technology has

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9102

Author's personal copy

been developed in the last years as a safety system to be

applied upstream a PEMFC stack, which requires a CO content

lower 20 ppm for a right operation [21]. Therefore, it is

assumed that after the PROX reactor the gas stream does not

contain CO. The PROX is inserted after the water steam

condensation step since it operates at temperature lower than

the steam condensation temperature at operating pressure.

After its use as heating fluid in the reforming process, the

molten salt stream is sent to a Rankine cycle (Fig. 1), where its

residual sensible heat is converted into electrical energy. By

this cycle, conversion efficiencies up to 38% are claimed to be

attainable in medium-sized plants (tens of MWs) [12] whereas

only 28–32% are achievable in small-sized plants (few MWs).

In the process proposed in the present work the steam turbine

size is small (<1 MWel) and, therefore, a 28% Rankine cycle

efficiency has been assumed in the following simulations.

3.5. Thermal recoveries

The molten salt stream has to supply the heat duty of the

process, i.e. the heat to drive the steam reforming reaction, to

pre-heat the inlet gas mixture stream and to generate the

reactant steam.

Of course, the possibility of thermal recoveries has to be

taken into account since they reduce the thermal load

required from the molten salt stream, allowing a higher exit

temperature of the molten salt and, consequently, a greater

electricity production from the Rankine cycle. Important

contributions to the process overall heat requirements should

be given by:

� The process stream cooling between the reformer (about

500–540 �C) and the first WGS step (450 �C), and between the

first and the second WGS reactor WGS (at about 450 �C and

250 �C, respectively).

� The process steam condensation.

In the following calculations, a conservative assumption is

made about the aforementioned thermal recoveries which are

considered to be just enough to heat the reactant water up to

saturation conditions and to pre-heat the gas reactant

mixture up to the inlet reformer conditions.

3.6. Numerical solution and simulations

In order to solve the set of partial differential equations of the

two-dimensional model, the radial coordinate is made

discrete by means of central second-order differences: the

resulting ODE system is solved using a Runge–Kutta method

with variable step.

Reformer geometry and catalyst properties (see Table 1) are

kept at fixed values whereas methane conversion, enriched

methane obtained and electrical power produced are calcu-

lated in each simulation. The effect of the following variables

and operating conditions has been analyzed:

- Residence time

- Steam-to-carbon ratio

- Operating pressure

- Temperature of the feed

- Co-current vs. counter-current mode

- Number of reformers assembled in the shell

The final hydrogen content specification (17 vol%) is

obtained by adding the required flow-rate of pure methane to

the SMR plant outlet stream (Fig. 2).

4. Results and discussion

Reactor conditions used in our simulations are reported in

Table 1. Catalyst pellets and packed bed properties are typical

of industrial steam reformers, while length and diameter of

the reactor are much lower than in industrial tubes which are

10–15 m long and with an internal diameter of about 10–13 cm.

The higher compactness of the reformer applied in this work

is due to the lower methane conversion required in the

process.

We assumed, as design point, a flow-rate of 4 kg/s for the

molten salt stream at 550 �C sent to the shell (Fig. 3). The

molten salt flow-rate in the solar collector plant is adjusted to

obtain 550 �C at the module exit, depending on the direct solar

radiation availability (that is characterized by seasonal, daily

and instantaneous variability). By means of the high temper-

ature heat storage tank (Fig. 1) it is possible to match the

steady state running chemical plant (molten salt flow-rate

fixed at 4 kg/s) with the CSP plant where molten salts flow at

variable rate. The use of properly sized thermal storage

systems and, possibly, of a backup heat source (e.g. methane

or biomass) is essential to ensure steady state running of the

chemical plant.

In each simulation, methane conversion

XCH ¼FCH4 ;0 � FCH4 ;ex

FCH4 ;0(13)

is calculated and the amount of methane to be added to reach

the hydrogen content specification of 17 vol% is evaluated.

Then, the enriched methane (HCNG-17) production rate and

the electric power obtained exploiting the molten salt residual

sensible heat are assessed.

4.1. The effect of the residence time

The effect of the residence time on the plant performance is

shown in Figs. 5 and 6 for the co-current heating fluid

configuration. Residence time is calculated as the ratio

between the mass of catalyst in the packed bed and the total

inlet molar flow-rate:

W=F ¼ rcatð1� 3ÞpR2i L

Ftot(14)

Table 1 – Geometric features and physical properties.

L (m) 0.5

Ri (m) 0.0381 (1 1/200)

dp (m) 0.011

rB (kg/m3) 1016

3 0.5

h 0.02

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9 103

Author's personal copy

In simulating the effect of, W/F the steam-to-carbon ratio, i.e.

the ratio between the reactant steam and the methane in the

inlet stream, is set at 2.5; gas inlet temperature and operating

pressure are 500 �C and 10 bar, respectively.

According to Eq. (14), the higher the total inlet molar flow-

rate, the lower the resulting residence time.

Fig. 5 shows the effect of the residence time on steam

reformer methane conversion. Increasing the W/F ratio has

a positive effect on the methane conversion because of the

longer contact time of the gas mixture with the catalyst

packed bed; moreover, increasing the W/F ratio the reformer

outlet gas mixture temperature increases (TR,ex¼ 464.8 �C at

W/F¼ 1 kgcat s/mol, TR,ex¼ 467.2 �C at W/F¼ 10 kgcat s/mol).

On the other hand, increased residence times obtained

through lower feed (and methane) flow-rates lead to lower

amounts of hydrogen produced. As shown in Fig. 6, the total

flow-rate of enriched methane produced is strongly affected

by the residence time, being reduced to less than one-fifth

increasing the residence time from 1 to 10 kgcat s/mol. As

a consequence, by increasing the residence time, the steam to

be generated for the reaction and the total thermal duty

required to the molten salt stream are reduced; additionally,

the temperature of the molten salt leaving the reactor and the

steam generator (TMS,ex¼ 508.5 �C at W/F¼ 1 kgcat s/mol,

TMS,ex¼ 535.5 �C at W/F¼ 3 kgcat s/mol and TMS,ex¼ 545.3 �C at

W/F¼ 10 kgcat s/mol) is higher at longer residence times, while

the electrical power generated by the steam turbine exploiting

the residue molten salt sensible heat increase (see Fig. 6).

In conclusion, increasing the gas residence time leads to

a higher methane conversion in the steam reformer and to

a higher electrical power output, but reduces the enriched

methane production rate. The optimal residence time has to

be fixed on the basis of the amounts of enriched methane and

electrical power required by the users and plant design

optimization.

Counter-current configuration has been simulated as well

by setting the sign ‘‘minus’’ in Eq. (6) and using the boundary

condition (10).

The reformer performance is throughout similar to that

one obtained in the co-current configuration. The reason is

that the molten salt temperature profile (Fig. 7) is almost flat,

both for co-current and counter-current layout, since the

amount of heat supplied to the reactant mixture is only

a small fraction of the total sensible heat stored in the salt

stream.

Also temperature profiles of the gas near the tube wall, in

the middle section and on the reformer axis, are reported in

Fig. 7. Strong radial temperature gradients can be observed

with quite low temperatures in the zone around the axis.

However, because of the low methane conversion required for

this application, the thermal level reached in the reformer is

enough for the plant specifications. An issue to be tackled

would be the Ni–Al2O3 catalyst activity at these temperatures,

but many papers focussed on membrane-assisted ‘‘low’’

Fig. 5 – Methane conversion at various residence times; gas

inlet conditions: 500 8C, 10 bar, steam-to-carbon [ 2.5.

Fig. 6 – Amount of enriched methane and electrical power

produced at various residence times; gas inlet conditions:

500 8C, 10 bar, steam-to-carbon [ 2.5.

Fig. 7 – Temperature profiles of molten salt stream and of

gas inside the reformer (r is the radial distance in the gas-

side reformer tube of internal radius Ri: 0 < r < Ri); W/

F [ 3 kgcat s/mol; gas inlet conditions: 500 8C, 10 bar,

steam-to-carbon [ 2.5.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9104

Author's personal copy

temperature methane steam reforming reactors [22–25],

which work at this thermal level, attest the capability of the

catalyst to support the reactions at temperature within the

range 350–450 �C, i.e. at a thermal level even lower than

the one considered in this work.

4.2. The effect of steam-to-carbon ratio

Fig. 8 shows the effect of the inlet steam-to-carbon ratio S/C on

the steam reformer performance. Obviously, increasing the

amount of steam in the reactant mixture leads to an increase

of the methane conversion since both steam reforming (1) and

water gas shift (2) reactions are supported. On the other hand,

the larger the S/C ratio the lower is the resulting plant

performance in terms of enriched methane and electrical

power production rate (Fig. 9). In fact, fixing both the residence

time and the reformer volume means that the total inlet flow-

rate is constant for all simulations (2.72 kmol/h for each

reformer). Then, increasing the fraction of steam in the inlet

flow-rate leads to the reduction of the amount of inlet

methane flow-rate. Consequently, a lower amount of

methane reaches a higher conversion degree thanks to the

larger steam-to-carbon ratio, but globally the hydrogen

produced is lower. Concerning the electrical power produced,

when S/C is increased the larger amount of heat required for

reacting steam generation reduces the residual molten salt

heat for the thermoelectric conversion system.

Therefore, S/C ratio has to be as small as possible but larger

than the stoichiometric value in order to avoid carbon coke

formation. S/C¼ 2.5 can be considered a fair value for our

process.

4.3. The effect of the operating pressure

The effect of the inlet pressure P on the methane conversion

XCH4 at constant S/C is reported in Fig. 10.

Since the steam reforming reaction (1) occurs with an

increase of the gas volume, the pressure affects negatively the

equilibrium conversion, so that both XCH4 and the enriched

methane production results unfavourable at high pressure

levels (Fig. 11).

On the other hand, increasing the operating pressure leads

to a slight enhancement of the electrical power production.

The main reason is that the latent heat required to generate

the reactant steam slightly decreases as the pressure

increases (37.8 MJ/kmol at 5 bar vs. 34 MJ/kmol at 20 bar).

Altogether, a convenient solution appears to be setting the

inlet pressure at a level as low as possible consistently with

the pressure drops which, however, in the ranges of operating

conditions used in this work, are always quite small.

4.4. The effect of inlet temperature

A high inlet temperature Tin has a positive effect both on the

reformer and on the overall plant performance, as shown in

Figs. 12 and 13, thanks to the endothermicity of the reforming

reaction.

Obviously, the inlet temperature cannot be higher than the

molten salt inlet temperature (550 �C), unless an additional

heat source is used.

Fig. 9 – Amount of enriched methane and electrical power

produced at various steam-to-carbon ratios; gas inlet

conditions: 500 8C, 10 bar, W/F [ 3 kgcat s/mol.

Fig. 10 – Pressure effect on reformer methane conversion;

W/F [ 3 kgcat s/mol; gas inlet conditions: 500 8C, S/C [ 2.5.

Fig. 8 – Methane conversion vs. steam-to-carbon ratio; gas

inlet conditions: 500 8C, 10 bar, W/F [ 3 kgcat s/mol.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9 105

Author's personal copy

In this work, an inlet temperature of 500 �C is assumed,

since no fossil fuels are used for the process heat duties.

4.5. The effect of the number of reformers

The effect of the number of reformers (nreformers) assembled

inside the reactor shell (Fig. 3) is analyzed at the fixed value of

molten salt flow-rate of 4 kg/s. Results reported in Fig. 14 show

a linear increase of the enriched methane production rate with

the number of reformers at constant residence time W/

F¼ 3 kgcat s/mol, due to the small effect of nreformers on molten

salt temperature profile (the exit molten salt temperature from

the reformer is 548.2 �C at nreformers¼ 4 vs. 546.5 �C at

nreformers¼ 8). Differently, the electrical power output decreases

as a consequence of the lower final temperature of the molten

salt (TMS,ex¼ 534.8 �C at nreformers¼ 4, TMS,ex¼ 519.7 �C at

nreformers¼ 8) due to the larger amount of steam to be generated

increasing the number of reformers. Therefore, although the

temperature of molten salt exiting the reformer is quite inde-

pendent from the number of reformers assembled in the tubes-

and-shell reactor, the larger amount of reactant steam to be

generated has a strong effect on the final molten salt temper-

ature and consequently on the electric power produced (422 kW

at nreformers¼ 4 vs. 395.5 kW at nreformers¼ 8).

As for the residence time, the number of reformers is fixed

on the basis of the enriched methane and of the electrical

power required by the users.

5. Plant design

In this section, we consider the main design features of a solar

enriched methane plant for supplying the enriched methane

and the electrical energy required by a small Italian

municipality.

Fig. 12 – Reformer conversion at various gas mixture inlet

temperatures; W/F [ 3 kgcat s/mol; gas inlet conditions:

10 bar, S/C [ 2.5.

Fig. 13 – Amount of enriched methane and electrical power

produced at various inlet temperatures W/F [ 3 kgcat s/mol;

gas inlet conditions: 10 bar, S/C [ 2.5.

Fig. 14 – Amount of enriched methane and electrical power

produced vs. number of reformers in configuration of

figure; W/F [ 3 kgcat s/mol; gas inlet conditions: 500 8C,

10 bar, S/C [ 2.5.

Fig. 11 – Amount of enriched methane and electrical power

produced at various operating pressures; W/F [ 3 kgcat s/

mol; gas inlet conditions: 500 8C, S/C [ 2.5.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9106

Author's personal copy

The yearly pro-capite household consumption of methane

and electricity, according to the Italian Statistics Institute

(ISTAT) [26] are:

QCH4 ¼ 429:1m3

y

QE:E: ¼ 1228:7kWh

y

The methane consumption has to be transformed to

equivalent enriched methane consumption on the basis of the

different heating values (LHV basis):

Qen�CH4¼ 486:9

m3

y

The volumetric flow consumption is higher for the enriched

methane since the methane is characterized by a lower

heating value per unit of volume greater than the hydrogen

one (10.7 MJ/Nm3 for hydrogen, 35.7 MJ/Nm3 for methane).

Contrarily, the mass flow is strongly lower, since hydrogen is

the fuel with the greatest heating value per unit of mass

(120 MJ/kg vs. 50 MJ/kg about for the methane).

According to the results of our simulations, the following

operating conditions have been assumed:

� steam-to-carbon ratio, S/C¼ 2.5;

� operating pressure, P¼ 5 bar;

� inlet gas mixture temperature, Tin¼ 500 �C.

Moreover, it has been assumed that the plant operates in

continuous mode (8760 h/year), thanks to the use of a properly

dimensioned molten salt hot storage (550 �C) tank.

Residence time W/F and number of reformers nreformers are

fixed, assuming that each user requires the above Qen�CH and

QE.E. values of enriched methane and of electricity,

respectively.

Fig. 15 reports the values of W/F and nreformers required to

satisfy the energy demand of various numbers of users. For

instance, 2930 users can be supplied by enriched methane and

electricity using a plant with nreformers¼ 4 and W/

F¼ 2.9 kgcat s/mol; likewise, 3000 users require nreformers¼ 6

and W/F¼ 4.85 kgcat s/mol, while 3020 users require

nreformers¼ 8 with W/F¼ 7 kgcat s/mol.

To obtain the size of the solar plant required, a linear

concentrator has been considered, of the model developed

and tested at ENEA [27]. This plant is based on the solar trough

technology, with a NaNO3/KNO3 (60/40 w/w) molten salt

mixture as heat transfer fluid and storage medium working in

the 290–550 �C temperature range; a two-tank heat storage

system (Fig. 1) has also been adopted. The assumed optical

efficiency is 0.8, while the thermal efficiency has been calcu-

lated by a thermal simulation of the collector string (it varies

with the irradiation); 5% thermal losses due to piping and

storage systems have also been considered. Given the ‘‘active

area’’ A of the solar field (i.e. the effective collection area of the

mirrors), the capacity S of the heat storage system, and the

irradiation sequence, a time simulation of the heat supplied

by the CSP plant can be performed, using the calculated effi-

ciency value. As irradiation sequence, the measured hourly

sequence at the site of Priolo Gargallo (Sicily) in the year 2003

has been adopted; the calculated average solar-to-thermal

efficiency with this radiation sequence is about 58.5 %. An

external backup heat source, i.e. a natural gas or biomass

fuelled molten salt heater, is often applied in CSP plants in

order to guarantee a constant-rate heat delivery. The heat

storage system with high capacity S can considerably reduce

the amount of integration, but cannot eliminate it completely:

as a matter of fact, a solar heat storage can compensate day–

night cycles and also short cloudy periods, but cannot

compensate protracted cloudy periods. Thus, an additional

energy source is sometimes required, especially in the winter.

Simulations at various values of A and S have been carried

out, assuming a nominal constant output salt flow-rate of

4 kg/s as design point (1.58 MWth of thermal power), and the

plant working from 20 January to 15 November.

Fig. 16 shows the variation of the backup integration with the

heat storage capacity S and with the active solar field area A. It

can be observed that the CSP plant backup duty can be drastically

reduced by increasing the heat storage capacity S up to a certain

value, after which just a slight backup decrease can be obtained

Fig. 15 – Number of enriched methane and electricity users

supplied by the solar enriched methane plant at various

feedstock residence times.

Fig. 16 – Relationship between amount of backup heat

source as % of nominal CSP plant capacity (1.58 MW), heat

storage capacity S, and the required solar field active area A.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9 107

Author's personal copy

by increasing the storage capacity S. In fact, while short-time

fluctuations (night–day cycles or short cloudy periods) can be

easily compensated by a relatively small storage, fluctuations on

longer periods (the order of several days or weeks) cannot be

balanced even by a heat storage with large capacity.

Various criteria can be adopted to choose the best config-

uration. If the priority is the complete exploitation of the

produced solar energy, an under-sized solar field (sized on the

summer radiation peak) of 10,000 m2 (active area) with

thermal storage of 40 MWh can be considered; however, such

a solution requires a high energy backup, that is about 39% of

the nominal yearly solar plant thermal power (Fig. 16).

Differently, adopting larger solar fields (e.g. sized on the

average spring radiation), the heat collected by the solar plant

in summer can saturate the heat storage (enriched methane

and electricity production): in this case an alternative use of

the exceeding solar energy should be planned (i.e. summer tri-

generative operation), or the solar plant should be partially

shut down when the available radiation exceeds the heat load

and the heat storage is full.

Another sizing strategy is limiting the integration to

a planned value; for example, to obtain a backup lower than

the 10% of nominal CSP plant capacity (i.e. 158 kWth on yearly

average), a solar field active area of 22,000 m2 is required, with

thermal storage of 160 MWh (Fig. 16); in this case, during the

summer the solar heat production exceeds by the 50% of the

nominal production.

An intermediate choice can be to match the average

production of solar energy with the power of the plant; in this

case, the total yearly solar energy production is equalized to

the total yearly energy load (i.e. the energy required by the

users) and the resulting solar field active area is 16,000 m2,

with thermal storage of 80 MWh and winter backup of 20%

that is the same as the summer energy excess. This last choice

can be considered a good compromise when land availability

and backup heat resource are both limited.

Considering that the effective surface occupied by the solar

field is about twice the active area A to avoid shadows, these

results are summarized in Table 2.

From these data we can conclude that the solar enriched

methane plant can be usefully applied for small municipalities,

while for towns with more than 20,000 inhabitants the space

needed could be a drawback. Moreover, the technology

proposed seems to be suitable for big hospitals, sport centres,

hotels, etc., wherever a free space is available for the CSP plants.

Furthermore, a rough estimation of greenhouse gas emis-

sion reduction applying the solar enriched methane plant can

be made on the basis of the following remarks.

Each inhabitant emits about 1.45 ton/y of CO2 for domestic

requirements of gas (842 kg/year) and electricity (606 kg/year).

Applying the proposed technology, the electricity generation

does not emit GHG, since it is produced by solar energy. The

burning of 17 vol% enriched methane (486.9 m3/year) in place

of pure methane (429.1 m3/year) leads to about 793 kg/year of

CO2 emission at the final user, while the solar SMR process

produces about 40.6 kg/year of CO2 (for 486.9 m3/year of

HCNG-17 produced) that can be sent to suitable CO2 stable

disposal systems.

Hence, GHG emission reduction of about 45.3% can be

achieved by this co-generative solar application. Of course,

additional CO2 emissions should be accounted if a fossil fuel

based CSP backup system is applied; for example, for the solar

plant configuration with 316 kWth, yearly average backup

(Table 2) by means of a methane heater with 80% efficiency,

CO2 emission reduction of about 32.0% can be achieved, that is

still a reasonable gain. In a more exhaustive life-cycle

assessment of this technology, also the ‘‘indirect’’ CO2 emis-

sions should be considered, i.e. those deriving from the plant

construction and salt production; on the other hand, the

impact of CSP plant ‘‘carbon footprint’’ on the overall GHG

emissions can be considered negligible.

6. Conclusions

The performance of a novel hybrid plant for the production of

a 17 vol% H2–CH4 gas mixture has been assessed. The steam

reforming heat duty is supplied by a molten salt stream

heated up by a concentrating solar power (CSP) plant. A two-

dimensional model of the reactor has been used to simulate

the effect of some operating conditions, as residence time,

steam-to-carbon ratio, operating pressure and inlet

temperature.

Calculation results show that a CSP plant with an active

area of about 16,000 m2 coupled with a tube-and-shell reactor,

with four reformers, is able to supply the electricity and

enriched methane to about 2930 domestic users.

However, some important issues have to be faced yet,

mainly dealing with technical-economy assessment of this

novel co-generative plant design. The effective surface occu-

pied by the solar field and methane consumption, as well as

the resulting GHG emissions, all depend on the design strategy

and, hence, on the resource accessibility (free land areas,

biomass availability, etc.).

The proposed technology seems to be suitable for munic-

ipalities, hospitals, hotels, sport centres, etc. and its wide-

spread application would be a crucial step towards the

achievement of the Kyoto specifications.

r e f e r e n c e s

[1] Haeseldonckx D, D’haeseleer W. The use of natural-gaspipeline infrastructure for hydrogen transport in a changingmarket structure. International Journal of Hydrogen Energy2007;32:1381–6.

[2] Bauer CG, Forest TW. Effect of hydrogen addition on theperformance of methane-fueled vehicles. Part I: effect of S.I.

Table 2 – Solar plant design specifications to supply thesolar enriched methane plant (1.58 MWth heat load) forabout 3000 users located in South Italy.

Backup heatconsumption(kWth, yearly average)

Heat storagecapacity (MWhth)

Occupied fieldarea (m2)

616 40 20,000

316 80 32,000

158 160 44,000

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9108

Author's personal copy

engine performance. International Journal of HydrogenEnergy 2001;26:55–70.

[3] Bauer CG, Forest TW. Effect of hydrogen addition on theperformance of methane-fueled vehicles. Part II: driven cyclesimulations. International Journal of Hydrogen Energy 2001;26:71–90.

[4] Orhan Akansu S, Dulger Z, Kaharaman N, Veziroglu TN.Internal combustion engines fuelled by natural gas–hydrogen mixtures. International Journal of HydrogenEnergy 2004;29:1527–39.

[5] Ortenzi F, Chiesa M, Scarcelli R, Pede G. Experimental tests ofblends of hydrogen and natural gas in light-duty vehicles.International Journal of Hydrogen Energy 2008;33:3225–9.

[6] Winter CJ, Sizmann RL, Vant-Hull LL, editors. Solar powerplants. New York: Springer-Verlag; 1991.

[7] Mills D. Advances in solar thermal electricity technology.Solar Energy 2004;76:19–31.

[8] Herrmann U, Kearney DW. Survey of thermal energy storagefor parabolic trough plants. ASME Journal of Solar Energy andEngineering 2002;124:145–51.

[9] Pacheco JE, Showalter SK, Kolb WJ. Development ofa molten-salt thermocline thermal storage system forparabolic trough plants. ASME Journal of Solar Energy andEngineering 2002;124:153–9.

[10] Herrmann U, Kelly B, Price H. Two-tank molten salt storagefor parabolic trough solar power plants. Energy 2004;29:883–93.

[11] Kearney D, Herrmann U, Nava P, Kelly B, Mahoney R,Pacheco J, et al. Assessment of a molten salt heat transferfluid in a parabolic trough solar field. ASME Journal of SolarEnergy and Engineering 2003;125:170–6.

[12] Sargent & Lundy LLC Consulting Group. Assessment ofparabolic trough and power tower solar technology cost andperformance forecasts. NREL Subcontractor Report, NREL/SR-550–34440; October 2003.

[13] Xu J, Froment G. Methane steam reforming, methanationand water-gas shift: I. Intrinsic kinetics. AIChE Journal 1989;35(1):88–96.

[14] Rostrup-Nielsen JR. Production of synthesis gas. CatalysisToday 1993;18:305–24.

[15] Xu J, Froment G. Methane steam reforming II: diffusionallimitations and reactor simulation. AIChE Journal 1989;35(1):97–103.

[16] Kulkarni BD, Doraiswamy LK. Estimation of effectivetransport properties in packed bed reactors. CatalysisReviews: Science and Engineering 1980;22(3):431–83.

[17] De Wasch AP, Froment GF. Heat transfer in packed beds.Chemical Engineering Science 1972;27:567–76.

[18] Miliozzi A, Giannuzzi GM, Tarquini P, La Barbera A. Fluidotermovettore: dati di base della miscela di nitrati di sodio epotassio, ENEA Report; ENEA/SOL/RD/2001/07.

[19] Tsotsas E, Schlunder E. Heat transfer in packed beds withfluid flow: remarks on the meaning and the calculation ofa heat transfer coefficient at the wall. Chemical EngineeringScience 1990;45:819–37.

[20] Li C, Finlayson B. Heat transfer in packed beds – a reevaluation.Chemical Engineering Science 1977;32:1055–66.

[21] Dudfield CD, Chen R, Adcock PL. A carbon monoxide PROXreactor for PEMFC automotive application. InternationalJournal of Hydrogen Energy 2001;26:763–75.

[22] Lin Y, Liu S, Chuang C, Chu Y. Effect of incipient removal ofhydrogen through palladium membrane on the conversionof methane steam reforming: experimental and modelling.Catalysis Today 2003;82:127–39.

[23] Oklany J, Hou K, Hughes R. A simulative comparison of denseand microporous membrane reactors for the steamreforming of methane. Applied Catalysis A: General 1998;170:13–22.

[24] Madia G, Barbieri G, Drioli E. Theoretical and experimentalanalysis of methane steam reforming in a membranereactor. Canadian Journal of Chemical Engineering 1999;77:698–706.

[25] Shu J, Grandjean B, Kaliaguine S. Methane steam reformingin asymmetric Pd and Pd-Ag porous SS membrane reactors.Applied Catalysis A: General 1994;119:305–25.

[26] ISTAT, Indicatori ambientali urbani: Anni 2004–2005; 22novembre 2006.

[27] The ENEA Working Group. Solar thermal energy production:guidelines and future programmes of ENEA. ENEA Report,ENEA/TM/PRESS/2001-07; June 2001.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 9 8 – 1 0 9 109