本研究第一年的任務 已完成工作 完成度
高溫供氣系統的發展
已設計一套高溫供氣系統，可分別將氫氣與氧氣 加溫至 800 oC，質量流率至少可以達到 2×10-5 kg/s，並可自由調動並量測流量，偵測壓力
100%
電池單極板的發展與簡易型 流道的設計
已完成經由初步的流場模擬與簡易型平行流道 製作，分為三角形與矩型截面兩種，極板內可裝 置套管以量測溫度與壓力
100%
封裝技術的發展
已完成利用石英玻璃棉封裝之技術，可防止氣體 滲漏與電池組裝漏電，並配合單極板內流道設 計，盡量減低 PEN 所承受之應力，減少材料破 壞之可能性
100%
研發成果資料表
日期：2005 年 1 月 18 日 計 畫 名 稱 ： 固態氧化物燃料電池高溫供氣系統與電池介面之發展與整合 計畫主持人：孫珍理
計畫編號：NSC 93-2623-7-011-006 -ET
期刊 論文
研討會
"Flow Channel Design of Interconnect for Planar SOFC – A Numerical
Investigation," ASME 2005 Summer Heat Transfer Conference, San Francisco, CA, U.S.A., July 17-22, 2005. (摘要已被接受)
技術報告
申請 獲得 專利
應用 與產業界、研發
機構互動成果
與核能研究所內進行氧化物燃料電池開發研究之團隊，針對電解 質、電極與 interconnect 等關鍵材料，單電池數值模擬進行討論交 流。
可利用之產業 及 可開發之產品
燃料電池供氣系統與測試平台
技術特點
利用特殊設計避免高溫材料膨脹翹曲所引起之滲漏與安全性問 題，供氣管件之絕緣處理，溫度迴授控制系統、氣體流量與電池功 率之匹配等。
推廣及運用的價值
目前燃料電池的測試平台皆以外國廠商為主，本計畫自行建構之高 溫供氣系統與測試平台，提供了關鍵技術的突破。目前國內研究燃 料電池的團隊越來越多，可為有需求的研究機構節省外購的經費。
Proceedings of HT2005 2005 ASME Summer Heat Transfer Conference July 17-22, 2005, San Francisco, California, USA
HT2005-72624
FLOW CHANNEL DESIGN OF INTERCONNECT FOR MICROSCALE SOFC – A NUMERICAL INVESTIGATION
Shien-Chih Ou and Chen-li Sun Department of Mechanical Engineering Nation Taiwan University of Science and Technology
Taipei,106 Taiwan [email protected] [email protected] ABSTRACT
In this study, the impacts of flow channel design of interconnect on the performance of planar SOFC (Solid Oxide Fuel Cell) are evaluated through cell-level simulation with commercially software package CFD-ACE. Four different flow pattern archetypes are examined: serpentine, double serpentine, staggered cylinder, and diagonal rib. Co-flow configuration is applied for fuel delivery in this investigation.
To keep the Ohmic loss in the same order for all four archetypes, the ratio of interconnect contact area (rib) to channel projection area is approximately 0.5. The influences of the different flow pattern designs on the distributions of the reactant concentrations, temperature, and current density are discussed. The deviations of the polarization curves are less than 2% for all four archetypes despite the pressure reduction is significantly smaller for staggered cylinder and diagonal rib flow channels. This is ascribed to the uniform gas concentration in microscale SOFC irrespective of flow pattern archetype. The variations of the pressure reduction are mainly caused by different flow characteristics of the flow channel designs and are less revealing in pointing out the performance variations of fuel cell. The pressure reduction of the serpentine channel is found to have the largest pressure reduction in the cell.
Keywords: solid oxide fuel cell (SOFC), performance analysis, gas manifold design
INTRODUCTION
In the last ten years, environmental concerns about global warming and the need to reduce carbon dioxide emissions provides the stimulus to seek alternative methods of power generation. Especially under the influences of the Kyoto Protocol [1], the urge of pursuing new technologies that can ultimately achieve zero emission becomes apparent. Among other possible solutions, fuel cell has drawn numerous researchers’ attention due to its stability and high efficiency.
Among different types of fuel cell, SOFC has the advantages of excellent efficiency, low emission of pollutants, and structure simplicity. Due to its high operational temperature, normally around 600 to 1000oC, SOFC does not require expensive catalysts nor to worry about water management [2]. In general, an operational fuel cell system is consists of a stack of single cells to augment the output power. A single cell has a sandwiched structure of two interconnects and PEN (Positive-Electrolyte-Negative) layer. The interconnects not only electrically connect the anode of one cell to the cathode of the adjacent cell, but are also serve as fluidic manifolds to circulate reagents and eject product steam. A poor manifold design may result in uneven fuel distribution, fuel crossover, cell-to-cell electrical shorts, and degeneration of the cell-to-cell performance.
Many researches were conducted to discuss the effects of manifold design on cell performance. Recknagle et al. [3]
employed 3D thermo-fluid electrochemical simulation for co-, counter-co-, and cross-flow stack designs of SOFC. The numerical results showed that co-flow case corresponded to the most uniform temperature distribution and smallest thermal gradients for similar fuel utilization and average cell temperature. The cross-flow counterpart however, displayed a hot spot toward air outlet and localized fuel depletion. Iwata et al. [4] studied the effects of gas re-circulation ratio, operating pressure on current and temperature distribution for planar SOFC. Simulations for co-, counter-, and cross-flow types were conducted with adiabatic boundary conditions. The temperature distribution was uniform regardless of flow type under the boundary condition of radiative exchange between outer interconnect and electric furnace surface.
Ferguson et al. [5] presented a 3D mathematical model to examine the local distribution of electrical potential, temperature and concentration of the chemical species for a unit of SOFC. The effects of geometry parameters, such as electrode thickness and rib width, are explored to demonstrate using the code as a design tool. The outcome revealed that wider rib gave less Ohmic losses but caused poor species
diffusion underneath the rib. Similar conclusions were also obtained by Tanner and Virkar [6], and Lin et al. [7]. Yakabe et al. [8] used finite volume method to construct a simulation model that included steam-reforming, water-shift reaction, and gases diffusion in the porous electrodes. Fuel reforming was found to induce a steep thermal gradient near inlet and radiation on internal surface of manifold would make 10% of the differences in temperature.
Nevertheless, the shortcoming of the aforementioned studies is lacking of discussing the influences of flow pattern archetype in interconnect on cell performance. To fulfill this need, the present work reports on a numerical investigation to comprehend the roles of flow channel design on performance characteristics of microscale SOFC in conjunction with pressure drop in microchannels.
MODEL DEVELOPMENT 1. Assumptions
In this study, the following common simplifying assumptions are made.
Fuel Cell operates at steady state.
The reactant gases obey ideal gas behaviors.
Flow field remains laminar.
The physical properties of all materials are isotropic and homogenous.
Gravity is negligible.
Fig. 1 Cell geometry and simulation domain model 2. Model Geometry and Physical Properties
The simulation domain is a single SOFC cell consisted of anode, electrolyte, cathode and two interconnects. As shown in Fig. 1, the single cell is 1 cm wide and 1 cm long, with a thickness of 50 μm for electrodes, 200 μm for electrolyte, and 1000 μm for interconnects. The depth of flow channel is fixed at 500 μm. All models of different flow pattern archetypes have the same geometry for each layer and share similar physical conditions. The physical properties used in the simulation are summarized in Table 1.
To level the Ohmic polarization across different flow channel designs, interconnect contact area is kept at 63% to 66% of the active area. Nevertheless, the widths of ribs vary
from 100 μm for serpentine and double serpentine channels, to 200 μm for diagonal rib channel. The width of the central rib that separates two serpentine channels in double serpentine design is 700 μm. The diameter of the columns in stagger cylinder channel is 460 μm.
Table 1 Physical Properties of Fuel Cell Model Component Electrolyte Anode Cathode Interconnect
Material YSZ Ni/ZrO2 LSM LaCrO3
The electrochemical reactions on each electrode are expressed as:
Anode: 2 H2 + 2 O2- → 2 H2O + 4 e- (1) Cathode: O2 + 4 e- → 2 O2- (2)
A single set of governing equations is employed for all computational domains, similar to the approaches of Cha, et al.
[9].
3. Boundary Conditions
A mass flow rate of 1 × 10-8 kg/s of pure hydrogen at 1000K and 8 × 10-8 kg/s of pure oxygen at 1000K are assigned to anode inlet and cathode inlet, respectively to keep the molar flow rates equivalent for both reagents. Outlet pressures are set to 1 atm. The corresponding Reynolds numbers are smaller than 10 for both reagents to validate the laminar flow hypothesis in the simulation. The upper, lower, and peripheral surfaces of the cell are assumed to be adiabatic, which corresponds to the central cells in a multi-cell stack. Fixed potential are imposed at the current collection surfaces.
4. Grid Sensitivity
Grid studies are carried out to determine the optimal strategy for mesh generation. Two main types of grid are tested: structured and mixed grid which contains structured and prismatic unstructured meshes. The result of the grid test is showed in Fig. 2 with average grid size spanning from 2.8 × 105μm3 to 2.5 × 106μm3. Current densities are found to be within 1% deviation by using structured grid with cell size under 106 μm3 for serpentine, double serpentine, and diagonal rib archetypes, while similar result is obtained by using mixed grid for staggered cylinder archetype. In this study, the unit: μm
average grid sizes are 6.4 × 105 μm3, 5.1 × 105 μm3, 4.3 × 105 μm3, and 4.4 × 105 μm3 for serpentine, double serpentine, staggered cylinder, and diagonal rib archetype, respectively.
Average Grid Size (μm3)
105 106
Current Density (A/cm2)
0.74
Fig. 2 Grid sensitivity for various flow pattern archetypes.
RESULTS AND DISCUSSION 1. Temperature and Flow Profile
Figure 3 shows the velocity profiles in the anode side channel at 0.3 V overpotential. Velocity is found to be maximized at the outlet for channel of anode side. This is contributed from the steam produced by the electrochemical reaction at anode. On the contrary, velocity at outlet of cathode side channel is not magnified. When electrochemical reaction takes place, oxygen is consumed at cathode and the flow rate is declined.
Current Density (A/cm2)
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Pressure Drop (Pa)
0
Fig. 4 Pressure reduction at anode side channel vs.
current density
The variations of pressure reduction at anode side channel with current density are shown in Fig. 4 for all flow pattern archetypes. In the serpentine design, the long single channel results in relative high velocity and pressure drop is as high as 1800 Pa at a current density of 0.78 A/cm2. For double serpentine, the bifurcation of the channel reduces the pressure drop greatly to 408 Pa at 0.71 A/cm2. The pressure drops for staggered cylinder and diagonal rib designs are even lower, ranging from 16 Pa to 56 Pa. In microscale SOFC, feature sizes are normally shorter than the penetration length [7] and spices diffusion is quite uniform. There is little influence of concentration overpotential and the major cause of the pressure reduction is the flow pattern designs. The results indicate that the serpentine design needs more power for fuel delivery than
(a) (b)
(c) (d)
Velocity (m/s)
Fig. 3 Velocity profile at anode side channel for
(a) Serpentine, (b) double serpentine, (c) staggered cylinder, (d) diagonal rib archetype.
staggered cylinder and diagonal rib designs to achieve the same supply flow rate.
The temperature profiles are found to be very uniform inside the cell for all four flow pattern archetypes. With adiabatic boundary condition and supply reagents at 1000K, mean cell temperatures range from 1200K to 2000K with current density spanning from 0.13 A to 0.78 A/cm2. Double serpentine design shows a lower cell temperature at high current density, 1899K at 0.708 A/cm2. The differences of mean cell temperature for four flow pattern archetypes vary from 16K at low current density to 137 K at high current density.
2. Species Distribution
Figure 5 shows the mass fraction distribution of hydrogen at the interface between anode and electrolyte. A combination of Fig. 3 and Fig. 5 highlights the effects of flow pattern designs on fuel transport and flow. In Cha’s [10] research, species distribution is uniform on electrode with 100 μm wide ribs. Since the width of rib is shorter than the penetration length [7], there is no noticeable species gradient underneath the rib. However, the fuel concentration shows variation across fluidic channel and contact rib with 500 μm wide ribs.
In our study, the width of ribs spans from 100 μm to 460 μm for all archetypes. No obvious concentration gradient is observed under the ribs except the central rib (600 μm) in double serpentine design.
For serpentine design, mass fraction of hydrogen is reduced gradually along the channel. Similar phenomenon is observed for double serpentine design. For staggered cylinder design, mass fraction of hydrogen decreases drastically near the inlet region. Molar fraction of hydrogen remains around 0.2 to 0.3 in most of the active area for staggered cylinder archetype.
For diagonal rib design, higher molar fraction of hydrogen is
observed along the inlet-outlet diagonal and results in low molar fraction of hydrogen regions at the two opposite corners.
The depletion of hydrogen indicates that product steam is accumulated at the upper right and lower left corners.
3. Cell Performance
The current density distributions at the current collection surface of cathode are presented in Fig. 6. Maximum current densities are observed around the periphery of cell for all flow pattern designs due to low local Ohmic loss. Similarly, high current density in the central rib region is prominent for double serpentine archetypes. Since the channel spacing is small (100 μm) for serpentine and double serpentine designs, current density distributions are relatively uniform in the center region.
For staggered cylinder design, current density under the cylinders is 33% more of that in the surrounding region. For diagonal rib design, spots of low current density are located near the two ends of ribs.
Current density distribution is found to be correlated to the contact area of interconnect and electrode. When the width of rib is smaller than 200 μm, rib collects less current and most electrons conducts through the edge of the interconnect plate.
Despite the variation of current density distributions, cell polarization curves are almost identical for all four flow pattern archetypes, as shown in Fig. 7. This is mainly because the rib contact area remains nearly consistent for all designs and keeps the Ohmic loss equivalent. Moreover, high cell temperature results in low activation overpotential and the polarization curves are nearly linear at low current density. Since pure hydrogen and oxygen are used in the simulation, in addition to the small feature sizes in microscale SOFC, concentration overpotential is small. According to Fig. 7, the cell performance of double serpentine archetype is found to be the best. The comparable low cell temperature of serpentine
(a) (b)
(c) (d)
Mass Fraction
Fig. 5 Molar fraction distributions of hydrogen at the interface between anode and electrolyte for (a) Serpentine, (b) double serpentine, (c) staggered cylinder, (d) diagonal rib archetype.
archetype contributes to a higher open circuit voltage (OCV) at high current density.
Current Density (A/cm2)
0.0 0.2 0.4 0.6 0.8 double serpentine diagonal rib staggered cylinder
Fig. 7 Model cell polarization curves
Figure 8 compares the power density for all flow pattern archetypes. Maximum power density of 0.194 W/cm2 is achieved at 0.348 A/cm2 for double serpentine design. For current density smaller than 0.3 A/cm2, power density is augmented with increasing the current density since more electrons are produced at large current density. Nevertheless, mean cell temperature rises with the increase of current density and causes a reduction of OCV. For current density larger than 0.4 A/cm2, power density decreases with the increase of current density. To achieve the maximum power, the cell should be operated around 0.56 V.
Current Density (A/cm2)
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Peak pPwer Density (W/cm2)
0.00
Fig. 8 Comparison of power density CONCLUSIONS
This study presents a cell-level simulation of microscale SOFC for serpentine, double serpentine, staggered cylinder diagonal rib archetypes. Mean cell temperature is found to play an intriguing role on the output power. Although high operational temperature helps to reduce the activation overpotential, it also decreases the open circuit voltage (OCV).
At high current density, the reduction of OCV overwhelms the benefit of decreasing the activation overpotential and results in low power output. Moreover, the numerical results show that if the rib is narrower than 500 μm, and interconnect contact area remains consistent, the cell performance is nearly independent of the flow pattern designs. Regardless of the cell performance, pressure reduction is significantly smaller for staggered cylinder and diagonal rib designs. This indicates
(a) (b)
(c) (d)
Current Density (A/m2)
Fig. 6 Current density distribution at current collection surface of cathode for (a) Serpentine, (b) double serpentine, (c) staggered cylinder, (d) diagonal rib archetype.
that with proper flow pattern design, the power consumption of fuel delivery can be minimized for microscale SOFC at the same level of output power.
ACKNOWLEDGEMENT
This work is supported by National Science Council of Taiwan under grant No. NSC 93-2623-7-011-006 -ET.
REFERENCES
[1] 1997, "Kyoto Protocol." United Nations Framework Convention on Climate Change, Kyoto, Japan.
[2] Issacci, F., 2003, "Thermal Management and Transport Phenomena in Fuel Cell System - Practical Issues,"
Proceedings of the The 6th ASME-JSME Thermal Engineering Joint Conference, March 16-20, Torrance, California.
[3] Recknagle, K. P., Williford, R. E., Chick, L. A., Rector, D.
R., and Khaleel, M. A., 2003, "Three-Dimensional Thermo-Fluid Electrochemical Modeling of Planar SOFC Stacks,"
Journal of Power Sources 113(1), pp. 109-114.
[4] Iwata, M., Hikosaka, T., Morita, M., Iwanari, T., Ito, K., Onda, K., Esaki, Y., Sakaki, Y., and Nagata, S., 2000,
"Performance Analysis of Planar-Type Unit SOFC Considering Current and Temperature Distributions," Solid State Ionics 132(3), pp. 297-308.
[5] Ferguson, J. R., Fiard, J. M., and Herbin, R., 1996, "Three-Dimensional Numerical Simulation for Various Geometries of Solid Oxide Fuel Cells," Journal of Power Sources 58(2), pp.
109-122.
[6] Tanner, C. W. and Virkar, A. V., 2003, "A Simple Model for Interconnect Design of Planar Solid Oxide Fuel Cells,"
Journal of Power Sources 113(1), pp. 44-56.
[7] Lin, Z., Stevenson, J. W., and Khaleel, M. A., 2003, "The Effect of Interconnect Rib Size on the Fuel Cell Concentration
[7] Lin, Z., Stevenson, J. W., and Khaleel, M. A., 2003, "The Effect of Interconnect Rib Size on the Fuel Cell Concentration