**3. Solid oxide fuel cell system fuelled by bioethanol**

As mentioned earlier, utilization of bioethanol by being converted into H2 for electrical power generation via SOFC is recognised. Thus, this section describes the fundamental process of an SOFC system fuelled by bioethanol and the criteria used to define the performance evaluation indicators of this SOFC system as follows:

### **3.1 Process description**

The bioethanol-fuelled SOFC system basically consists of a bioethanol pretreatment unit, preheaters, reformer, fuel cell, and afterburner as illustrated in Figure 2. Bioethanol is purified in the pretreatment unit to achieve a specified ethanol concentration (25mol% ethanol, a suitable stoichiometric ratio for the ethanol steam reforming reaction in Reaction (1)). Then, the steam with a desired ethanol is fed to an external reformer operated under thermodynamic equilibrium condition. Ethanol steam reforming is selected for converting the raw materials into hydrogen rich gas. The reaction is assumed to occur isothermally in the reformer. Finally, the reformed influent stream is fed to the SOFC's anode chamber together with excess air (5 times) preheated and fed to the cathode chamber to produce electricity and thermal energy. The effluent steam containing residual fuel released from the fuel cell is combusted in the afterburner and heat from the fuel combustion is recovered to supply all the heat-demanding units i.e. preheaters, purification unit, and reformer. The final temperature of exhaust gas emitted to atmosphere is specified at 403K (Jamsak et al., 2007). The performance of the SOFC system can be simulated using Aspen Plus software.

Fig. 2. Basic process diagram of bioethanol-fuelled SOFC system

#### **3.2 Mathematical model**

The SOFC model was adapted from the prior literature of Piroonlerkgul et al., 2008 to study the performance of SOFC system. From this model, a constant operating voltage along the cell length and isothermal condition were assumed. Only hydrogen oxidation was considered to react electrochemically within the fuel cell. Oxygen ion electrolyte type was chosen for the SOFC and its electrochemical reactions are described below:

$$0.5\text{O}\_2 + 2\text{e}^\* \Rightarrow \text{O}^{2+} \tag{5}$$

$$\rm H\_2 + O^{2-} \rightleftharpoons \rm H\_2O + 2e^- \tag{6}$$

The validation of this model was in a good agreement with experimental results (Zhao et al., 2005; Tao et al., 2005) at high hydrogen contents (hydrogen mole fraction = 0.97) and (Petruzzi et al., 2003) at low hydrogen contents (hydrogen mole fraction = 0.26). The materials used in the SOFC stack were YSZ, Ni-YSZ and LSM-YSZ for electrolyte, anode and cathode, respectively.

#### **3.2.1 Electrochemical model**

#### **3.2.1.1 Open circuit voltage**

196 Renewable Energy – Trends and Applications

C2H5OH + 1.78H2O + 0.61O2 => 2CO2 + 4.78H2 (4)

As mentioned earlier, utilization of bioethanol by being converted into H2 for electrical power generation via SOFC is recognised. Thus, this section describes the fundamental process of an SOFC system fuelled by bioethanol and the criteria used to define the

The bioethanol-fuelled SOFC system basically consists of a bioethanol pretreatment unit, preheaters, reformer, fuel cell, and afterburner as illustrated in Figure 2. Bioethanol is purified in the pretreatment unit to achieve a specified ethanol concentration (25mol% ethanol, a suitable stoichiometric ratio for the ethanol steam reforming reaction in Reaction (1)). Then, the steam with a desired ethanol is fed to an external reformer operated under thermodynamic equilibrium condition. Ethanol steam reforming is selected for converting the raw materials into hydrogen rich gas. The reaction is assumed to occur isothermally in the reformer. Finally, the reformed influent stream is fed to the SOFC's anode chamber together with excess air (5 times) preheated and fed to the cathode chamber to produce electricity and thermal energy. The effluent steam containing residual fuel released from the fuel cell is combusted in the afterburner and heat from the fuel combustion is recovered to supply all the heat-demanding units i.e. preheaters, purification unit, and reformer. The final temperature of exhaust gas emitted to atmosphere is specified at 403K (Jamsak et al., 2007). The performance of the SOFC system can be simulated using Aspen Plus software.

**3. Solid oxide fuel cell system fuelled by bioethanol** 

performance evaluation indicators of this SOFC system as follows:

**3.1 Process description** 

Feed T=298 K 10 wt%EtOH

**3.2 Mathematical model** 

Bioethanol Pretreatment unit

> Reformer T=1023K

chosen for the SOFC and its electrochemical reactions are described below:

Fig. 2. Basic process diagram of bioethanol-fuelled SOFC system

Heater

The SOFC model was adapted from the prior literature of Piroonlerkgul et al., 2008 to study the performance of SOFC system. From this model, a constant operating voltage along the cell length and isothermal condition were assumed. Only hydrogen oxidation was considered to react electrochemically within the fuel cell. Oxygen ion electrolyte type was

Air T=298K

SOFC

Afterburner

Heat & Electricity

Exhaust gases T=403K

Heater

The open circuit voltage (*E*) is formulated by the Nernst equation given in Eq. (7)

$$\mathbf{E}\_{\rm \\_} = \begin{array}{c c c} \mathbf{E}\_{0} & + & \frac{\mathbf{RT}}{\mathbf{F}} \ln \left( \frac{\mathbf{P}\_{\rm H\_{2}} \mathbf{P}\_{\rm O\_{2}}^{1/2}}{\mathbf{P}\_{\rm H\_{2}O}} \right) \end{array} \tag{7}$$

where *F* is Faraday constant (C mol-1) and *Pi* is a partial pressure of component *i*.

The actual operating voltage (*V*) is less than the open circuit voltage (*E*) due to the presence of various polarizations. Three types of polarization are considered in this model: Ohmic, Activation, and Concentration polarizations as below:

$$\mathbf{V} = \mathbf{E} - \mathbf{r}\_{\text{act}} - \mathbf{r}\_{\text{ohmic}} - \mathbf{r}\_{\text{conv}} \tag{8}$$

#### **3.2.1.2 Polarizations**

#### *3.2.1.2.1 Ohmic polarization*

This ohmic polarization involves the resistance of both ions flowing in the electrolyte and electrons flowing through the electrodes. This resistance loss is regarded as a major loss in the SOFC stack given as:

$$
\eta\_{\text{olmic}} = 2.99 \text{x} 10^{-11} \text{i} L \exp\left(\frac{10300}{T}\right) \tag{9}
$$

where *i* is current density and *L* is thickness of anode electrode

#### *3.2.1.2.2 Activation polarization*

Activation polarization is caused by the loss of electrochemical reaction rate at the electrodes. An operation of SOFC at high temperature can reduce this polarization as the rate-determining step becomes faster. Normally, activation polarization region occurs in the low current density range. This polarization is defined by the Butler-Volmer equation.

$$\dot{i}\_{\perp} = \dot{i}\_{o} \left[ \exp\left(\frac{azF\eta\_{\text{sat}}}{RT}\right) - \exp\left(-\frac{(1-a)zF\eta\_{\text{act}}}{RT}\right) \right] \tag{10}$$

The value of *α* and *z* were specified as 0.5 and 2 (Chan et al., 2001), respectively. Accordingly, the activation polarization at the anode and cathode sides can be arranged into another form as:

$$\left(\eta\_{\text{act,j}}\right) = \frac{\text{RT}}{\text{F}} \text{sinh}^{-1}\left(\frac{\text{i}}{\text{2i}\_{0\text{j}}}\right) \tag{11}$$

where *j* = anode, cathode

The exchange current density (*i*o,*<sup>j</sup>*) for both the anode and cathode sides are expressed as follows:

$$\mathbf{i}\_{\rm o,a} = \gamma\_{\rm a} \left( \frac{\mathbf{P}\_{\rm H\_2}}{\mathbf{P}\_{\rm ref}} \right) \left( \frac{\mathbf{P}\_{\rm H\_2O}}{\mathbf{P}\_{\rm ref}} \right) \exp\left( -\frac{\mathbf{E}\_{\rm act,a}}{\mathbf{RT}} \right) \tag{12}$$

$$\mathbf{i}\_{\rm o,c} = \sqrt{\frac{\mathbf{P}\_{\rm O\_2}}{\mathbf{P}\_{\rm ref}}} \Big|^{0.25} \exp\left(-\frac{\mathbf{E}\_{\rm act,c}}{\mathbf{RT}}\right) \tag{13}$$

where γa and γc are pre-exponential factors for anode and cathode current densities, respectively.

#### *3.2.1.2.3 Concentration polarization*

This polarization arises from the difference in gas partial pressures in the porous electrode region due to slow mass transport. It can be estimated by Eqs. (14) and (15) for the anode and cathode sides, respectively.

$$\mathbf{n}\_{\text{conc},\text{a}} = \frac{\text{RT}}{\text{2F}} \ln \left| \frac{\left(1 + (\text{RT}/2\text{F}) (\mathbf{l}\_{\text{a}}/\mathbf{D}\_{\text{a}\text{ (eff)}} \mathbf{p}\_{\text{H}\_{2}\text{O}}^{\text{I}}) \mathbf{i} \right)}{(1 - (\text{RT}/2\text{F}) (\mathbf{l}\_{\text{a}}/\mathbf{D}\_{\text{a}\text{ (eff)}} \mathbf{p}\_{\text{H}\_{2}}^{\text{I}}) \mathbf{i}} \right| \tag{14}$$

$$\eta\_{\rm conc,c} = \,^1\frac{\rm RT}{4F} \ln\left[\frac{\rm p\_{O\_2}^{l}}{(p\_c/\delta\_{O\_2}) - ((p\_c/\delta\_{O\_2}) - p\_{O\_2}^{l})\exp\Big[(RT/4F)(\delta\_{O\_2}1\_c/D\_{c(\text{eff})P\_c})\hat{\mathbf{i}}\right]}\right] \tag{15}$$

where la and lc are thicknesses of anode and cathode electrodes, respectively, while *δ*O2, *D*<sup>a</sup> (eff) and *D*c (eff) are given by:

$$\delta\_{\rm O\_2} = \frac{\rm D\_{O\_2, k \text{(eff)}}}{\rm D\_{O\_2, k \text{(eff)}} + D\_{O\_2 - N\_2 \text{(eff)}}} \tag{16}$$

$$\mathbf{D\_{a(eff)}} = \left(\frac{\mathbf{P\_{H\_2O}}}{\mathbf{P\_a}}\right) \mathbf{D\_{H\_2(eff)}} + \left(\frac{\mathbf{P\_{H\_2}}}{\mathbf{P\_a}}\right) \mathbf{D\_{H\_2O(eff)}}\tag{17}$$

$$\mathbf{D}\_{\rm c(eff)} = \frac{\xi}{\mathbf{n}} \left( \frac{1}{\mathbf{D}\_{\rm O\_2,k}} + \frac{1}{\mathbf{D}\_{\rm O\_2-N\_2}} \right) \tag{18}$$

$$\frac{1}{\mathcal{D}\_{\text{H}\_{2}\text{(eff)}}} = \frac{\xi}{\mathbf{n}} \left( \frac{1}{\mathcal{D}\_{\text{H}\_{2},\text{k}}} + \frac{1}{\mathcal{D}\_{\text{H}\_{2}-\text{H}\_{2}\text{O}}} \right) \tag{19}$$

The exchange current density (*i*o,*<sup>j</sup>*) for both the anode and cathode sides are expressed as

ref ref

0.25 O act,c

P E

<sup>i</sup> <sup>γ</sup> exp P P RT 

ref

<sup>i</sup> <sup>γ</sup> exp <sup>P</sup> RT 

where γa and γc are pre-exponential factors for anode and cathode current densities,

This polarization arises from the difference in gas partial pressures in the porous electrode region due to slow mass transport. It can be estimated by Eqs. (14) and (15) for the anode

conc,a I

<sup>2</sup>

2

I O

cO cO O O c c(eff) c

RT 1 (RT/2F)(l /D p )i

2F (1 (RT/2F)(l /D p )i

2 22 2

4F (p /δ ) ((p /δ ) p )exp (RT/4F)(δ l /D p )i

2

2 2 2 O ,k(eff)

O ,k(eff) O N (eff)

2 2

2 22

O ,k O N

nD D 

 

where la and lc are thicknesses of anode and cathode electrodes, respectively, while *δ*O2, *D*<sup>a</sup>

D

D D

2 2

H O H a(eff) H (eff) H O(eff) a a

P p 

p p D DD

<sup>ξ</sup> 1 1 <sup>D</sup>

H (eff) 2 2 H ,k H H O <sup>2</sup> <sup>2</sup> 1 ξ 1 1 D nD D

act,j

o,a a

η ln

conc,c I

RT p

2

c(eff)

O

δ

o,c c

2

*3.2.1.2.3 Concentration polarization* 

and cathode sides, respectively.

η ln

(eff) and *D*c (eff) are given by:

where *j* = anode, cathode

follows:

respectively.

η sinh

1

H HO 2 2 act,a

P P E

RT i

F 2i 

0,j

(12)

2

(16)

(17)

I a a (eff) H O

a a (eff) H

(11)

(13)

(14)

(15)

(18)

(19)

 

$$\frac{1}{\mathbf{D}\_{\rm H\_2O(eff)}} = \frac{\xi}{\mathbf{n}} \left| \frac{1}{\mathbf{D}\_{\rm H\_2O,k}} + \frac{1}{\mathbf{D}\_{\rm H\_2-H\_2O}} \right| \tag{20}$$

The relationship between the effective diffusion parameter (*D(eff)*) and ordinary diffusion parameter (*D*) can be described by:

$$\mathbf{D}\_{\text{(eff)}} = \begin{array}{c} \mathbf{n} \\ \hline \xi \end{array} \tag{21}$$

where *n* is electrode porosity and ξ is electrode tortuosity. Assuming straight and round pores, the Knudsen diffusion parameter can be calculated by:

$$\mathbf{D}\_{\rm A,k} = -9700 \sqrt{\frac{\mathbf{T}}{\mathbf{M}\_{\rm A}}} \tag{22}$$

The binary ordinary diffusion parameter in a gas phase can be calculated using the Chapman-Enskog theory of prediction as below:

$$\mathbf{D}\_{\rm A-B} = \left[1.8583 \times 10^{-3} \left(\frac{\mathrm{T}^{3/2} \left(\left(1/\mathrm{M}\_{\rm A}\right) + \left(1/\mathrm{M}\_{\rm B}\right)\right)^{1/2}}{\mathrm{P} \sigma\_{\rm AB}^2 \mathrm{Q}\_{\rm D}}\right)\right. \tag{23}$$

where *σ*AB the characteristic length, *Mi* is molecular weight of gas i, and *<sup>D</sup>* is the collision integral. These parameters are given by:

$$
\sigma\_{AB} \quad = \begin{array}{c} \sigma\_A + \sigma\_B \\ \hline 2 \end{array} \tag{24}
$$

$$\Omega\_{\rm D} = \frac{A}{T\_k^B} + \frac{C}{\exp(DT\_k)} + \frac{E}{\exp(FT\_k)} + \frac{G}{\exp(HT\_k)}\tag{25}$$

where the constants *A* to *H* are *A* = 1.06036, *B* = 0.15610, *C* = 0.19300, *D* = 0.47635, *E* = 1.03587, *F* = 1.52996, *G* = 1.76474, *H* = 3.89411.

#### **3.3 Evaluation of process performance**

The proposed bioethanol-fuelled SOFC system for electrical power generation needs to be evaluated together with any process design adjustments to obtain optimum performance. A number of criteria can be used to define the performance of the system, e.g. economic, 1st and 2nd laws of thermodynamics, environment, etc. Fundamentally, the overall performance evaluation of an SOFC system is defined in terms of electrical efficiency as below:

$$\mathfrak{n}\_{\text{elec,ov}} = \frac{\text{Net electrical power output}}{\text{mol}\_{\text{Fuel}} \,\text{LHV}\_{\text{Fuel}}} \tag{26}$$

The definition of the above equation is energy efficiency based on 1st law of thermodynamics which initially accounts on an ideal energy conservation law. Fuel input term is referred to the lower heating value (LHV). In considering an energy loss from the system which is closer to actual condition, the definition of overall system efficiency is formulated as follows:

$$\mathfrak{n}\_{\text{lelec,ov}} = \frac{\text{Net electrical power output}}{\text{mol}\_{\text{Fuel.}} \text{e}\_{\text{Fuel}}^{\text{o}}} \tag{27}$$

This equation is exergy efficiency which further takes the 2nd law of thermodynamic into account stated that entropy loss occurred in the system with highly irreversible process especially combustion process. The fuel input denominator in Eq. (27) is referred to the standard exergetic potential of fuel. In addition, the analysis in term of exergy can determine the location, source, and amount of actual thermodynamic inefficiencies in each unit. Profound understanding can be perceived from this analysis for solving the process problem correctly.

The criterion mainly considered in this chapter is no external energy demand condition. In the SOFC system, there are units having the roles of both energy consumption and generation. Before investigating and evaluating the system efficiency, energy consumed or generated from the units is allocated within the system until the overall system is under selfsufficient energy condition or *Q*net = 0 as follows:

$$\mathbf{Q}\_{\text{net}} = \mathbf{Q}\_{\text{generation \text{\textquotedblleft}Q}\_{\text{demand}}} \mathbf{\color{red}{\mathbf{Q}}\_{\text{demand}}} = \mathbf{0} \tag{28}$$

where *Q*generation represents the heat from units which can generate thermal energy (SOFC and afterburner) while *Q*demand expressed as the heat from units which consume heat (bioethanol pretreatment unit, heaters, and reformer). The system operated at such a condition can help allocate energy within the process effectively. The exhaust gas released to atmosphere is specified at 403 K (Jamsak et al., 2007). The consideration of *Q*net = 0 associated with the process evaluation has led to the modified efficiency definition:

$$\mathfrak{n}\_{\text{elec,ov}} = \frac{\text{Net electrical power output}}{\text{mol}\_{\text{Fuel.}} \text{LHV}\_{\text{Fuel}} + \text{external heat demand}} \tag{29}$$

In case of incorporating a heat recovery unit such as combined heat and power (CHP) with the SOFC system, the definition of efficiency is adjusted to:

$$
\eta\_{\text{lelec,ov}} = \frac{\text{Net electrical power output + exchanged thermal energy}}{\text{mol}\_{\text{Fuel}}.\text{LHV}\_{\text{Fuel}} + \text{external heat demand}} \tag{30}
$$

#### **4. Process modification for improving performance of the SOFC system**

The fundamental process of the bioethanol-fuelled SOFC system needs to be further developed to utilize bioethanol effectively and achieve higher electrical efficiency. In this chapter, the performance improvement of SOFC systems under consideration is based on selection for appropriate units. The possible units are structurally modified and evaluated for their energy consumption. The process modification of the SOFC system can be divided by two main scopes including adjusting the fuel cell module and improving the balance of plant.

#### **4.1 Solid Oxide Fuel Cell**

Originally, the Solid Oxide Fuel Cell (SOFC) is classified as a high-temperature fuel cell. Due to the demand for high cost materials and fabrication, the intermediate temperature solid

This equation is exergy efficiency which further takes the 2nd law of thermodynamic into account stated that entropy loss occurred in the system with highly irreversible process especially combustion process. The fuel input denominator in Eq. (27) is referred to the standard exergetic potential of fuel. In addition, the analysis in term of exergy can determine the location, source, and amount of actual thermodynamic inefficiencies in each unit. Profound understanding can be perceived from this analysis for solving the process

The criterion mainly considered in this chapter is no external energy demand condition. In the SOFC system, there are units having the roles of both energy consumption and generation. Before investigating and evaluating the system efficiency, energy consumed or generated from the units is allocated within the system until the overall system is under self-

 Qnet = Qgeneration - Qdemand = 0 (28) where *Q*generation represents the heat from units which can generate thermal energy (SOFC and afterburner) while *Q*demand expressed as the heat from units which consume heat (bioethanol pretreatment unit, heaters, and reformer). The system operated at such a condition can help allocate energy within the process effectively. The exhaust gas released to atmosphere is specified at 403 K (Jamsak et al., 2007). The consideration of *Q*net = 0

associated with the process evaluation has led to the modified efficiency definition:

Fuel Fuel

Fuel Fuel

**4. Process modification for improving performance of the SOFC system** 

Net electrical power output <sup>η</sup> mol .LHV external heat demand

In case of incorporating a heat recovery unit such as combined heat and power (CHP) with

Net electrical power output exchanged thermal energy <sup>η</sup> mol .LHV external heat demand

The fundamental process of the bioethanol-fuelled SOFC system needs to be further developed to utilize bioethanol effectively and achieve higher electrical efficiency. In this chapter, the performance improvement of SOFC systems under consideration is based on selection for appropriate units. The possible units are structurally modified and evaluated for their energy consumption. The process modification of the SOFC system can be divided by two main scopes including adjusting the fuel cell module and improving the balance of

Originally, the Solid Oxide Fuel Cell (SOFC) is classified as a high-temperature fuel cell. Due to the demand for high cost materials and fabrication, the intermediate temperature solid

(29)

(30)

Fuel Fuel

Net electrical power output <sup>η</sup> mol .e (27)

elec,ov o

problem correctly.

sufficient energy condition or *Q*net = 0 as follows:

elec,ov

elec,ov

**4.1 Solid Oxide Fuel Cell** 

plant.

the SOFC system, the definition of efficiency is adjusted to:

oxide fuel cell was later developed with the research into novel material technology and thin layer techniques applied in electrolyte and electrodes. Regarding the fuel cell geometry design, it is useful to differentiate the scope into macro and micro geometry configurations. The micro geometry covering the structures of anode, electrolyte, and cathode has direct effects on the electrochemical performance of the fuel cell. The heat transfer mechanisms of convection and conduction through heat exchange areas and the mass transport through active surface areas are influenced by the macro geometry (Nagel et al., 2008). Generally, primary structures of SOFC are tubular, planar, and monolithic as shown in Figures (3), (4), and (5), respectively. The SOFC structure of planar design is more compact than the tubular design and also offers higher ratio of power per volume (Pramuanjaroenkij et al., 2008). For the monolithic design, this SOFC design uses the similar concept with shell-and-tube heat exchanger. It combines the tri-layer of anode-electrolyte-cathode into a compact corrugated structure. This design can assist a thermal energy allocation exchanged between the flow channels and size of the fuel cell to become more compact with the corrugated selfsupporting structure.

Fig. 3. Schematic of tubular SOFC (Source: Kakac et al., 2007)

Fig. 4. Schematic of planar SOFC (Source: Bove & Ubertini, 2006)

Fig. 5. Schematic of monolithic SOFC; (Left) coflow and (Right) cross flow (Source: Minh, 1993)

#### **4.2 Balance of plant**

There are essential units around the fuel cell as supporting units for the overall electrical power generation process. These units can be modified to utilize energy within their system units suitably. Sections in the balance of plant which are potential in improving the efficiency of SOFC system are described as follows:

#### **4.2.1 Bioethanol pretreatment section**

This section has a key role in improving the efficiency of the SOFC system. Originally, bioethanol has a low concentration in a range of 1-7 mol% (Shell et al., 2004; Cardona Alzate & Sanchez Toro, 2006; Roger et al., 1980; Buchholz et al., 1987). In our studies, 10wt% (4.16 mol%) ethanol was specified to represent the range of actual bioethanol concentration. These bioethanol compositions are unsuitable for feeding into the reformer operating under ethanol steam reforming reaction to produce hydrogen because of high water content. Unnecessary thermal energy is required to heat up surplus water within the reformer and the sizes of reformer are larger than necessary. Hence, the bioethanol pretreatment unit plays an important part to purify bioethanol feed into a desired concentration of 25mol%ethanol (46wt% ethanol). A selection of appropriate purification unit for bioethanol conditioning must consider an effective separation with low energy consumption to offer a better performance of the system. In our research (Jamsak et al., 2007), we started with a conventional distillation column used in the bioethanol-fuelled SOFC system as illustrated in Figure 6.

A distillation column is commonly recognised as a high energy consumption unit, but the SOFC released a large amount of exothermic heat. Therefore, it is feasible to apply this unit as a bioethanol pretreatment unit. The results from our simulation studies indicated that there were some operating conditions which can run this system under *Qnet* = 0. However, the overall electrical efficiency obtained from this system was quite low due to high reboiler heat duty consumption and high amount of heat loss in the condenser. Afterwards, among the promising membrane technologies, pervaporation is considered as a replacement for the former purification unit as shown in Fig 7. By the principle of physical-chemical affinity between the membrane material and species, this unit consumes only heat for vaporizing a preferential substance permeated through the membrane. However, it is noted that a pervaporation depends on a driving force generation device, typically a vacuum pump is used to boost up its separation performance. Therefore, part of the generated electrical energy must be consumed to operate the device.

Fig. 5. Schematic of monolithic SOFC; (Left) coflow and (Right) cross flow (Source: Minh, 1993)

There are essential units around the fuel cell as supporting units for the overall electrical power generation process. These units can be modified to utilize energy within their system units suitably. Sections in the balance of plant which are potential in improving the

This section has a key role in improving the efficiency of the SOFC system. Originally, bioethanol has a low concentration in a range of 1-7 mol% (Shell et al., 2004; Cardona Alzate & Sanchez Toro, 2006; Roger et al., 1980; Buchholz et al., 1987). In our studies, 10wt% (4.16 mol%) ethanol was specified to represent the range of actual bioethanol concentration. These bioethanol compositions are unsuitable for feeding into the reformer operating under ethanol steam reforming reaction to produce hydrogen because of high water content. Unnecessary thermal energy is required to heat up surplus water within the reformer and the sizes of reformer are larger than necessary. Hence, the bioethanol pretreatment unit plays an important part to purify bioethanol feed into a desired concentration of 25mol%ethanol (46wt% ethanol). A selection of appropriate purification unit for bioethanol conditioning must consider an effective separation with low energy consumption to offer a better performance of the system. In our research (Jamsak et al., 2007), we started with a conventional distillation

A distillation column is commonly recognised as a high energy consumption unit, but the SOFC released a large amount of exothermic heat. Therefore, it is feasible to apply this unit as a bioethanol pretreatment unit. The results from our simulation studies indicated that there were some operating conditions which can run this system under *Qnet* = 0. However, the overall electrical efficiency obtained from this system was quite low due to high reboiler heat duty consumption and high amount of heat loss in the condenser. Afterwards, among the promising membrane technologies, pervaporation is considered as a replacement for the former purification unit as shown in Fig 7. By the principle of physical-chemical affinity between the membrane material and species, this unit consumes only heat for vaporizing a preferential substance permeated through the membrane. However, it is noted that a pervaporation depends on a driving force generation device, typically a vacuum pump is used to boost up its separation performance. Therefore, part of the generated electrical

column used in the bioethanol-fuelled SOFC system as illustrated in Figure 6.

**4.2 Balance of plant** 

efficiency of SOFC system are described as follows:

energy must be consumed to operate the device.

**4.2.1 Bioethanol pretreatment section** 

Fig. 6. Process diagram of bioethanol-fuelled SOFC system using a distillation column

Fig. 7. Process diagram of bioethanol-fuelled SOFC system using a pervaporation


Table 4. Energy consumption for anhydrous ethanol production from various purification processes (Source: Black, 1980; Jaques et al., 1972; Hala, 1969; Barba et al., 1985; Ligero and Ravagnani, 2003)

However, a pervaporation is still regarded as being the lowest energy consumption unit among the other purification units as shown in Table 4. (Reviewed by Kumar et al., 2010) that gives an example of using various purification processes for anhydrous ethanol production. To emphasize their mentioned data, the simulation results from our studies (Choedkiatsakul et al., 2011) showed the performance of bioethanol-fuelled SOFC system in comparison between two pretreatments; using distillation and pervaporation units. On the basis of purification process operated at 348K, Table 5 presents the classification of energy term in each unit for both purification processes. Although a pervaporation consumed an electrical energy within the unit, it offers an overall electrical efficiency (42%) superior to that of distillation column (34%). However, a hydrophobic membrane material used in the pervaporation required a high ethanol separation factor property as illustrated in Figure 8 but it may be unavailable in real membrane materials.


Table 5. Performance characteristics in comparison between two different purification units based on *Q*net = 0, ethanol recovery (*R*EtOH) = 80%, *V* = 0.7 V, and *T*SOFC = 1073 K (Source: Choedkiatsakul et al., 2011)

Fig. 8. Effect of ethanol recovery in pervaporation on the required ethanol separation factor of hydrophobic membrane

However, a pervaporation is still regarded as being the lowest energy consumption unit among the other purification units as shown in Table 4. (Reviewed by Kumar et al., 2010) that gives an example of using various purification processes for anhydrous ethanol production. To emphasize their mentioned data, the simulation results from our studies (Choedkiatsakul et al., 2011) showed the performance of bioethanol-fuelled SOFC system in comparison between two pretreatments; using distillation and pervaporation units. On the basis of purification process operated at 348K, Table 5 presents the classification of energy term in each unit for both purification processes. Although a pervaporation consumed an electrical energy within the unit, it offers an overall electrical efficiency (42%) superior to that of distillation column (34%). However, a hydrophobic membrane material used in the pervaporation required a high ethanol separation factor property as

illustrated in Figure 8 but it may be unavailable in real membrane materials.

Heat (MW) Bioethanol pretreatment unit Reformer Air preheater Afterburner

Electrical power (MW) Bioethanol pretreatment unit Electrical production Net electrical energy

Choedkiatsakul et al., 2011)

**Ethanol Separation Factor**

of hydrophobic membrane

Energy distribution Purification process configuration

Fuel utilization (%) 92 68 Overall electrical efficiency (%) 42 34

Table 5. Performance characteristics in comparison between two different purification units based on *Q*net = 0, ethanol recovery (*R*EtOH) = 80%, *V* = 0.7 V, and *T*SOFC = 1073 K (Source:

2,301 417 22,575 25,293

453 4,920 4,467

Pervaporation Distillation Column

3,580 421 23,892 27,893

0 3,701 3,701

**134.59**

**67.79**

**45.53**

**34.40**

**65 70 75 80 85 90 95 Ethanol recovery (%) in pervaporation**

Fig. 8. Effect of ethanol recovery in pervaporation on the required ethanol separation factor

**27.72 20.08 23.26**

Consequently, as schematically shown in Figure 9, this problem was solved by having a vapor permeation device installed after a pervaporation (Sukwattanajaroon et al., 2011) to improve ethanol separation performance, an important part of the SOFC system,. The permeate stream of a pervaporation in vapor phase which can be directly fed to a vapor permeator without preheating is a benefit of this technique. From our investigations based on *Q*net = 0, an available hydrophilic membrane of high water separation factor is a suitable choice to be used in a vapor permeation unit. The performance of SOFC system using this proposed purification process obviously overcomes the other two cases as shown in Figure 10.

Fig. 9. Process diagram of bioethanol-fuelled SOFC system using a hybrid pervaporation/vapor permeation process

The overall electrical efficiency can be ranked as: Integrated vapour permeation/pervaporation (45.46%) > pervaporation (36.46%) > distillation column (22.53%), respectively.

Fig. 10. Performance comparison of SOFC system with various purification processes based on self-sufficient condition (Qnet = 0, REtOH = 75%, V=0.75V, TSOFC = 1073K)
