Transient behaviour of a small methanol reformer for fuel cell during hydrogen production after cold start

Transcription

1 Energy Conversion and Management 46 (5) Transient behaviour of a small methanol reformer for fuel cell during hydrogen production after cold start Rong-Fang Horng * Department of Mechanical Engineering, Kun Shan University of Technology, 949, Da-Wan Road, Yung-Kung City, Tainan County, Taiwan 71, ROC Received 13 February 4;received in revised form 28 May 4;accepted 13 June 4 Available online 26 August 4 Abstract The cold start transient characteristics of a small methanol reformer for a fuel cell were investigated. The main parameters studied were the oxygen to methanol mol ratio (O/C), fuel supply rate, heating power and heating temperature. The composition of the gas produced by the reformer was analysed. The main aim of this paper was to determine a favorable combination of the parameters for obtaining rapid hydrogen production during the transient behaviour of the reformer. A small methanol reformer with fuel, air and water injectors, heaters and a catalyst was constructed. Vaporised methanol was injected into the reformer, which then flowed into the catalyst. For the purpose of enhancing the response of the cold start transient reaction, eight glow plugs were mounted at the inlet of the catalyst to control the flow temperature together with the adjustment of the oxygen to methanol mol ratio. The best response from cold start was obtained with 96 W heating power, 8 C heating temperature, 14 ml/min methanol and 7 L/min air supply rates among the experimented parameters. Under this operation setting, hydrogen was produced after 22 s from cold start with the production rate stabilising after 4 5 min and eventually reaching the highest concentration at 35 C. Generally, hydrogen commenced production at a catalyst outlet temperature of 1 C and stabilised at 35 C. Ó 4 Elsevier Ltd. All rights reserved. Keywords: Methanol reformer;partial oxidation;cold start;transient performance;hydrogen production * Corresponding author. Tel.: ;fax: address: hong.rf@msa.hinet.net (R.-F. Horng) /$ - see front matter Ó 4 Elsevier Ltd. All rights reserved. doi:1.116/j.enconman

2 1194 R.-F. Horng / Energy Conversion andmanagement 46 (5) Nomenclature B N S S x t U U RSS b d k e k bias limit repeated measurements standard deviation average precision a parameter;a function of measurement times and average precision measurement uncertainty measurement uncertainty derived by root sum square method bias error total measurement bias precision error Subscripts k source of error RSS root sum square x average value of total measurements 1. Introduction Severe air pollution in cities is a common problem faced by countries worldwide. Although it is generally accepted that electrifying vehicles is one of the most efficient methods in reducing this problem, this method has never quite been popularised due to the difficulties in maintaining a constant supply of electricity such as battery lives, battery recharging times etc., while the vehicle is in transit. The fuel cell, which uses hydrogen to generate electricity, is generally regarded as the most prospective method of powering vehicles. The main consideration is, thus, the source of the hydrogen and the method of storing it. The traditional method of storing gas in pressurised metal cylinders is not ideal as hydrogen is highly combustible. It is only very recently that a hydrogen storage canister was developed, which enabled safe storage of hydrogen and, therefore, the usage of fuel cells in electrical cars. Alternatively, if a small reformer that can produce hydrogen on board a car were available, this would undoubtedly be a more efficient and safer method of powering electrical cars. With a reformer, the main design considerations are its physical dimensions, its response rate after cold start and its transient response during acceleration. Hydrogen can be extracted by the reforming process from methanol, natural gas, fossil fuel etc. The methods of reforming include partial oxidation, auto-thermal, steam reforming and so on [1]. Hydrogen obtained from methanol possesses the desirable characteristics of ease of decomposition and a low reaction temperature of approximately 25 C and is, thus, a suitable method for steam reforming [2]. The Energy and Resources Laboratories of the Industrial Technology Research Institute of Taiwan has published some papers since 1993 on research of reformers for fuel cells. Cheng [3] proposed a set of design criteria and did a literature review of the development of the reformer.

3 R.-F. Horng / Energy Conversion andmanagement 46 (5) Sung [4] conducted a theoretical and experimental investigation on a methanol reformer and revealed that the conversion efficiency of methanol increased with the temperature of the reformer as did the CO level. He recommended the outlet temperature of the reformer be controlled between 18 and 31 C. Sung [5] also investigated a methane steam reformer and discovered that the CO concentration from the reformer fulfilled the level of the fuel cell. He further suggested that the water to methane ratio should not be too high in order to reduce the heat lost to heat the water and proposed that further research be conducted to characterise the behaviour of the reformer at the start and during changes in load. Chen [6] manufactured a small methanol reformer with a metallic substrate combined with a Cu ZnO/Al 2 O 3 catalyst and performed a parametric study from which it was concluded that the reformer performed optimally with a water to methanol mol ratio of 1.8 and an oxygen to methanol mol ratio of.2. Recent years have seen the rapid development of methanol reformers in particular applications involving cars. Numerous papers have been published by many researchers investigating the effect of cold start and the dynamic characteristics of methanol reformers. Hohlein et al. [7] manufactured a methanol steam reformer system operated at 24, 26 and 28 C, respectively, and noted the intimate relationship between the heating mode and the concentration of the emitted CO. They also found that the characteristics of the starting process were less than expectation. Using their newly constructed methanol reformer, Emonts et al. [8] discovered that under low load, the conversion rate of methanol was significantly improved when the operating temperature was increased from 26 to 28 C, with 1% conversion efficiency of methanol. Further increase in temperature, however, resulted in less evident improvement. They further revealed that the higher the temperature, the higher was the concentration of the emitted CO, the level of which was reduced as the load was increased. Wiese et al. [9] investigated the quasi-static and dynamic behaviour as well as the start up condition of a methanol reformer and noted that reaction first occurred after 1 min from start up. As the reaction rate of the fuel supply system was found to influence the dynamic response of the reformer significantly, they recommended that an injection method be adopted to enhance performance. Han et al. [2] built a 3 kw methanol reformer, which provided a thermal efficiency beyond 89% and attained maximum hydrogen production after 15 min from cold start up. Nagano et al. [1] investigated catalytic methanol steam reforming by simulation and experiment. They found that there was a trade off relationship between methanol conversion and CO concentration. The trade off relationship was improved by using an internal corrugated metal heater and an external catalytic combustion heater to enhance the heat transfer. Optimal reaction parameters were also obtained. Takeda et al. [11] used a steam reforming method together with fuel oxidation in a methanol reformer to enhance the vapourisation effect of the heater on the liquid fuel and water to accelerate the reaction of the catalyst. With a Cu ZnO/Al 2 O 3 catalyst, Choi et al. [12] studied the hydrogen emission characteristics with and without adding water to a methanol reformer. When no water was added, 1% reaction efficiency of the methanol could be achieved, while the presence of water resulted in a high concentration of hydrogen and a low level of CO. Holladay et al. [13] fabricated a micro scale methanol reformer and discovered that the temperature of 36 C resulted in a conversion efficiency of 99% using the steam reforming method for a O/C ratio of 1.8. Lindstrom and Pettersson [14] developed a methanol reformer fitted in cars for Volvo. The catalyst in the reformer under a combined partial oxidation and steam reforming method commenced operation after 4 6 min. Loffler et al. [15] included a pre-reformer in their reformer system, which promoted achieving the activation temperature of

4 1196 R.-F. Horng / Energy Conversion andmanagement 46 (5) the catalyst through fuel oxygenation and the main reformer attained the reaction condition by the steam reforming method. This system resulted in a high thermal efficiency and a high purity of the hydrogen produced. From the above review, it is evident that the methanol reformer has a desirable fast response to achieve the reaction temperature and is highly suitable for use with fuel cells to power automobiles. However, more efforts are still needed on the cold start and transient characteristics of the reformer if it is to be used as an on board vehicle fuel processor. The study on the cold start behaviour of the methanol reformer to find the parameters for shortening the start up time is the main aim of this paper. 2. Experimental method and procedure 2.1. Experimental set up The main apparatus was a purposely designed reformer with a catalyst supplied by the Heterogeneous Catalysis Department in the Union Chemical Laboratories of the Industrial Technology Research Institute of Taiwan. The experimental set up is shown in Fig. 1, and the specifications of the reformer are listed in Table 1. Other equipment included a fuel supply system, a temperature controlling unit, a gas sampling system and a gas chromatographer (Agilent Model 685). The catalyst was coated with a layer of honeycombed ceramic whose main components were Pt and the mixed oxide of Cu ZnO/Al 2 O 3. The fuel supply system consisted of a liquid and an air injection sub-system, each equipped with a pressure regulator and a flow meter. Fig. 1. Experimental arrangement of the methanol reformer.

5 R.-F. Horng / Energy Conversion andmanagement 46 (5) Table 1 Specifications of the reformer Main body Reaction chamber Height Nozzle Heater Power supply for heater 124B mm stainless steel 51 mm Fuel, water, air 12 W 8 glow plugs 12 VDC Catalyst Substrate Ceramic Diameter 117B mm Length 5 mm Composition Pt and Cu ZnO/Al 2 O 3 Mesh 64 cells/cm 2 The liquid fuel system had an additional electrical fuel pump and a fuel injector, which injected vapourised liquid methanol fuel, while the air injection system had a single and a multi hole injector. The fuel was well vaporised as it came into the catalyst. The flow temperature control system consisted of glow plugs, a temperature controller and a temperature logger. Each glow plug supplied a heating power of 12 W with a 12 V d.c. battery and a total of eight glow plugs, providing 96 W were employed to control the temperature of the flow entering the catalyst. Five K-type thermocouples were positioned at the front and the back of the fuel spray, the inlet and the outlet of the catalyst and at the smallest radius of the tapered pipe. The temperature at the outlet of the catalyst acted as the feedback to the temperature controller. The flow sampling and analyzing system consisted of a condenser, a sampling pump, sampling bags, a micro syringe and a gas chromatographer Experimental method As the main objective of the paper is to investigate the characteristics of the methanol reformer after cold start, the method of partial oxidation was employed to determine the best experimental parameters and operating procedure. The parameters included O/C ratio, heating power, heating temperature, methanol supply rate and air flow rate. For each transient test, the gas chromatographer was first warmed up before initiating the data logger. Next, the heating system was activated, and air heated by the glow plugs was then directed into the catalyst for pre-heating. After 1 s, the liquid fuel unit commenced supply, and simultaneously, the emitted gas was sampled, and the sample injected into the gas chromatographer for analysing its transient characteristics. This process continued until the composition of the gas achieved stability. To ensure no gas contamination, the sample bag must be vacuumed to remove any gas after each sampling. Between each cold start test, the entire system was shut down for at least six hours so that the temperature of the reformer system returned to room temperature. As mentioned, partial oxidation was used to investigate the stable characteristics of the reformer, and the methanol supply rate was used as the base line for calculating the required air flow using the O/C ratio. The supply rate was fixed as 14, 1 and 6 ml/min, while the O/C ratio was set between.12 and.97.

6 1198 R.-F. Horng / Energy Conversion andmanagement 46 (5) Table 2 Measurement uncertainties : ±1.1% CO: ±.51% CO 2 : ±.51% Temperature: ±2.5 C Current: ±.4 A Voltage: ±.2 V One of the main measuring instruments was the gas chromatographer, which gave a calibration line that determined the measurement accuracy of the gas composition. The calibration line was obtained by feeding three different concentrations of standard gas samples into the gas chromatographer and solving through linear regression for a R 2 (regression coefficient) that was above.999. Subsequent gas concentration produced from the reformer was thus determined based on this line. This system used argon as the carrier gas and the column used was SUPELCO U. The setting of the gas chromatographer was a carrier gas flow rate of 1 ml/min, an inlet temperature of 1 C, an oven temperature of 16 C and a detector temperature of C Measurement error analysis The total measurement error (d k ) includes a bias error (b) and a precision error (e k ) such that [16] d k ¼ b þ e k ð1þ The precision error is determined by taking N repeated measurements. The precision index of the average of a set of measurements is always less than that of an individual measurement according to S x ¼ p S ffiffiffiffi ð2þ N where S is the standard deviation of the N repeated measurements. The bias error is the systematic error, which is considered to remain constant during a given test. The measurement uncertainty (U) with 95% confidence can be given by the followed model: U RSS ¼½B 2 þðts x Þ 2 Š 1 2 ð3þ where U RSS is the measurement uncertainty derived by the root sum square (RSS) method, B is the bias limit, and t is set equal to 2 for large samples (N > 3). According to the above analysis, the measurement uncertainties are estimated and shown in Table Results and discussion In the experiment, different heating powers, coupled with different heating temperatures was studied to establish an operating procedure of the reformer from cold start. From this procedure,

7 R.-F. Horng / Energy Conversion andmanagement 46 (5) the best experimental parameters, such as the methanol supply rate and air flow rate were then determined. Figs. 2 4 compare the time taken for the outlet temperature of the catalyst to reach C, the self initiation temperature of the catalyst, under different heating powers coupled with different heating temperatures. It is noted that the shorter the time taken, the more efficient is the operation. Fig. 2 shows the results for a heating power of 96 W. It is evident from the figure that when the pre-set heating temperature is 8 C, the methanol supply rate is 14 ml/min and the air flow rate is 7 L/min, the best cold start condition is achieved in 22 s. This is followed closely by the setting conditions of 6 C and 4 C heating temperatures coupled with the same air flow rate of 7 L/min, which give 23 and 257 s, respectively. Further, under the same heating temperature of 8 C, but coupled with a methanol supply rate of 1 ml/min and a air flow rate of 6 L/min, an efficient cold start operating condition of 26 s, approximately, can be obtained. Good operating results are also obtained under the conditions of 6 and 4 C heating temperatures coupled with an air flow rate of 6 L/min to give the time taken for the temperature at the outlet of the catalyst to reach C as 27 and 275 s, respectively. However, when the methanol supply rate is 6 ml/min, the time taken after cold start to reach the required temperature is above 38 s, clearly much slower than the former two settings. Nevertheless, when a heating temperature of 8, 6 or 4 C is coupled with any of the above settings, the temperatures at the catalyst outlet all reached C. In other words, the self initiation conditions can be achieved for any of the three heating temperatures. Fig. 3 shows the results for a heating power of 72 W. From the figure, it is clear that the combination of 8 C heating temperature, 14 ml/min methanol supply rate and 7 L/min air flow rate results in the best cold start response, i.e. a total time of 235 s. Other desirable combinations include heating temperatures of 6 and 4 C with an air flow rate of 7 L/min, which result in Time for to reach o C(sec) Heating power: 96W Methanol supply 14mL/min 8 o C 6 o C 4 o C 1mL/min 6mL/min 8 o C 8 o C 6 o C 6 o C 4 o C 4 o C Air supply (L/min) Fig. 2. Comparison of the effect of air supply rate on the time taken for the outlet temperature of the catalyst to reach C under different methanol supply rates and different heating temperatures at a heating power of 96 W.

8 1 R.-F. Horng / Energy Conversion andmanagement 46 (5) Time for to reach o C(sec) Heating power: 72W Methanol supply 14mL/min 1mL/min 6mL/min 8 o C 6 o C 4 o C 8 o C 6 o C 4 o C 8 o C 6 o C Air supply (L/min) Fig. 3. Comparison of the effect of air supply rate on the time taken for the outlet temperature of the catalyst to reach C under different methanol supply rates and different heating temperatures at a heating power of 72 W. Time for to reach o C(sec) Heating power: 48W Methanol supply 14mL/min 1mL/min 8 o C 8 o C 6 o C 6 o C 4 o C 4 o C 6mL/min 8 o C 6 o C 4 o C Air supply (L/min) Fig. 4. Comparison of the effect of air supply rate on the time taken for the outlet temperature of the catalyst to reach C under different methanol supply rates and different heating temperatures at a heating power of 48 W. times of 25 and 265 s, respectively. When the methanol supply rate is reduced to 1 ml/min and the air flow rate is 6 L/min, the time taken to reach C is between 27 and 28 s. The longest time taken is when the methanol supply rate is 6 ml/min. Fig. 4 shows the results for a heating power of 48 W. The shortest time taken is between 245 and 28 s for the test condition of 8 C heating temperature, 7 L/min air flow rate and 14 ml/

9 R.-F. Horng / Energy Conversion andmanagement 46 (5) min methanol supply rate. When the methanol supply rate is reduced to 1 ml/min and the air flow rate is 6 L/min a good cold start condition is also achieved and the time taken for this case is between 28 and s. As with Figs. 2 and 3, a methanol supply rate of 6 ml/min results in the longest time taken to achieve C. From the above experimental results, it is clear that the best setting of those tested is with the combination of a heating power of 96 W, a heating temperature of 8 C, a methanol supply rate of 14 ml/min and an air flow rate of 7 L/min, which resulted in the shortest time of 22 s to achieve the C requirement at the catalyst outlet. This is followed by the combination of a heating power of 96 W, heating temperature of 6 C, methanol supply rate of 14 ml/min and air flow rate of 7 L/min, giving the time required as 23 s. Figs. 5 7 present the temperature distribution at the reformer under the test conditions of 8 C heating temperature, a methanol supply rate of 14 ml/min, an air flow rate of 7 L/min for heating powers of 48, 72 and 96 W, respectively. Simultaneously, the reformer temperature development of previously published work [14] was compared with this study in these figures. T 1 and T 2 are the temperatures in front and behind the fuel spray, respectively, while T 3 and are the temperatures at the inlet and outlet of the catalyst. From the temperature variations of T 1, T 2 and T 3, it is evident that after cold start, the methanol immediately comes into contact with the high temperature air and the heaters to produce significant oxidation, which causes its temperature to rise rapidly. It is noted that the experiment was designed such that the catalyst is heated by the heat in the air flow until the outlet temperature reaches the pre-set 8 C. As soon as the heaters are turned off, the air flow before the catalyst cools down, and therefore, the methanol would not be oxidised in this region. It is activated instead in the catalyst as the elevated temperature of the catalyst reduces the air flow rate. At the same time, T 1, T 2 and T 3 decrease rapidly due to the reduced heat supply from the heaters and the oxidation of the methanol, results from turning 5 4 T 2 T 3 Set heating power: 48 W Set heating temp.: 8 o C Methanol supply: 14 ml/min Air supply: 7 L/min Temperature ( o C) 1 T 1 Reformer temperature of Ref.14 =8 o C Time from cold start (sec) Fig. 5. Temperature time histories at various positions of the reformer during the transient condition after cold start of a reformer for heating powers of 48 W.

10 122 R.-F. Horng / Energy Conversion andmanagement 46 (5) T 2 T 3 Set heating power: 72 W Set heating temp.: 8 o C Methanol supply: 14 ml/min Air supply: 7 L/min Temperature ( o C) 1 T 1 Reformer temperature of Ref. 14 =8 o C Time from cold start (sec) Fig. 6. Temperature time histories at various positions of the reformer during the transient condition after cold start of the reformer for a heating power of 72 W. 6 5 T 3 T 2 Set heating power: 96 W Set heating temp.: 8 o C Methanol supply: 14 ml/min Air supply: 7 L/min Temperature ( o C) 4 1 T 1 Reformer temperature of Ref. 14 =8 o C Time from cold start (sec) Fig. 7. Temperature time histories at various positions of the reformer during the transient condition after cold start of the reformer for a heating power of 96 W. off the heaters and the reduced air flow rate. The rapid increase of is attributed to the achievement of the catalyst light-off temperature. From Fig. 5 it is evident that when the heating power is 48 W, the rate of increase of T 1 and T 2 are similar, while that of T 3 is more gradual. The time taken for to reach the pre-set 8 C is approximately 179 s, to reach C is approximately

11 R.-F. Horng / Energy Conversion andmanagement 46 (5) s and to reach the maximum temperature is approximately 36 s. Fig. 6 presents the results for the case where the heating power is 72 W. As with the previous case, the rate of increase of T 1 and T 2 are the same. The difference is that T 3 also increases at a similar rate. With the higher heating power, the time taken for to reach 8 C is shorter at about 173 s, to reach C is approximately 231 s and the time to reach the maximum temperature is approximately 355 s. When the heating power is further increased to 96 W as shown in Fig. 7, the rate of increase of T 1, T 2 and T 3 are almost identical. In this case, the time taken for to reach 8 C is an even shorter 165 s, to reach C is about 22 s and that to reach the maximum temperature is 345 s, approximately. As compared with the results of the previous work, as shown in the figures, the reformer temperatures of this study exceed those of Lindstrom et al. [14] at approximately s. Furthermore, the time for their reformer temperature to reach C is approximately 27 s. However, as discussed above, the times for reaching the same temperature in our study are 237, 231 and 22 s for heating powers of 48, 72 and 96 W, respectively. That is, the cold start response of this study is comparable to that of LindstromÕs. From the above analysis, it is apparent that heating power significantly affects the rise rate of the inlet and outlet temperatures of the catalyst. In other words, the higher the heating power, the faster is the temperature rise rate in the studied settings, which essentially means that the catalyst reaches its self initiation temperature in a shorter time. Fig. 8 shows the catalyst outlet temperature variation and the concentration of, CO and CO 2 emitted during the transient condition after cold start with a heating power of 96 W and the best settings of those experimented, i.e. a heating temperature of 8 C, a methanol supply rate of 14 ml/min and a gas supply rate of 7 L/min. It is clear that no hydrogen is produced when the outlet temperature is under 8 C. The reason attributed to the increase in the rise rate of the temperature of the catalyst due to the release of a significant amount of heat energy, which occurs when the methanol to air ratio approaches the theoretical air fuel ratio, for heating the, CO, CO 2 (Vol%) Set heating power: 96 W Set heating temp.: 8 o C Methanol supply: 14 ml/min Air supply: 7 L/min CO 2 CO Temperature of catalyst outlet, ( o C) Time from cold start (sec) Fig. 8. Gas composition and outlet temperature of the catalyst with time during the transient condition after cold start of the reformer for a heating power of 96 W.

12 124 R.-F. Horng / Energy Conversion andmanagement 46 (5) catalyst and the main body of the reformer. Consequently, the majority of the methanol will combust due to oxidation. Therefore, an initial level of 5% CO 2 appears, and the concentration of CO and remains zero until the heaters stop operating and the air supply rate reduces to a suitable level such that the severe oxidation of methanol shifts to partial oxidation to start producing. The outlet temperature of the catalyst at this instance is 1 C, approximately and rapidly increases as does the concentration of. The increase in concentration of the CO and CO 2 is, however, more gradual. When the outlet temperature reaches approximately 35 C, the production rate of stabilises and the hydrogen concentration remains at approximately 37%. Figs. 9 and 1 show the gas composition of the emission from the reformer and the temperature variation at the catalyst outlet for heating powers of 72 and 48 W, respectively, during the transient condition after cold start. In general, the outlet temperature of the catalyst and the variations of the concentrations of, CO and CO 2 are similar to those when the heating power was 96 W. The main disparities are the more gradual overall temperature rise, which delayed the reaction of the catalyst, and the slower reaction rate of the catalyst. Fig. 11 compares the effect of different heating powers on the outlet temperature of the catalyst and the concentration during the transient condition of the reformer after cold start. It is clear that under a heating power of 96 W, the temperature rise at the outlet of the catalyst is the fastest as is the rise in the concentration of hydrogen. This is followed by the heating temperature of 72 W and then by 48 W. It is evident that under the three different conditions, the higher the heating power, the higher is the rise rate of the temperature and the production of. Despite this difference, the times taken to achieve stable production of hydrogen for the three cases are almost identical and the concentration of produced are also similar. Fig. 12 shows the relation between the outlet temperature of the catalyst and the hydrogen concentration under the different methanol supply rates and O/C ratios when the reformer stabilises. Because of the methanol supply rates and the limit of the operating temperature of the, CO, CO 2 (Vol%) Set heating power: 72 W Set heating temp.: 8 o C Methanol supply: 14 ml/min Air supply: 7 L/min CO 2 CO Temperature of catalyst outlet, ( o C) Time from cold start (sec) Fig. 9. Gas composition and outlet temperature of the catalyst with time during the transient condition after cold start of the reformer for a heating power of 72 W.

13 R.-F. Horng / Energy Conversion andmanagement 46 (5) , CO, CO 2 (Vol%) Set heating power: 48 W Set heating temp.: 8 o C Methanol supply: 14 ml/min Air supply: 7 L/min CO 2 CO Temperature of catalyst outlet, ( o C) Time from cold start (sec) Fig. 1. Gas composition and outlet temperature of the catalyst with time during the transient condition after cold start of the reformer for a heating power of 48 W. Hydrogen (Vol%) Set heating temp.: 8 o C Methanol supply: 14 cc/min Air supply: 7 L/min Hydrogen 96W 72W 48W Temperature 96W 72W 48W Temperature of catalyst outlet, ( o C) Time from cold start (sec) Fig. 11. Comparison of the effect of heating power on the outlet temperature of the catalyst and the concentration during the transient condition after cold start of the reformer. catalyst, different methanol supply rates would produce different limiting values of the O/C ratio. Further, it is clear from the figure that under different methanol supply rates, the outlet temperature of the catalyst and the concentration of produced reveal similar trends. In other words, as the O/C ratio increases, the outlet temperature of the catalyst and the concentration increase simultaneously. When the methanol supply rate is 14 ml/min and the O/C ratio is.46, the best produced was 41% in volume. The corresponding outlet temperature of the catalyst is 355 C. If the supply rate is reduced to 1 ml/min and the O/C ratio is increased slightly

14 126 R.-F. Horng / Energy Conversion andmanagement 46 (5) Hydrogen (Vol%) Methanol supply 14ml/min 1ml/min 6 ml/min 5 T Methanol supply 4 14ml/min -5 1ml/min 6 ml/min O/C ratio Temperature of catalyst outlet, ( o C) Fig. 12. The relationship between the outlet temperature and the concentration under different methanol supply rates and O/C ratios during the stable condition of the reformer. to.58, the peak value of concentration produced is 36.5% in volume, and the corresponding outlet temperature of the catalyst is 352 C. At an even lower methanol supply rate of 6 ml/min and a O/C ratio of.74, the highest concentration produced falls to a mere 29.3% in volume, and the associated outlet temperature of the catalyst is 32 C, approximately. Beyond this peak value of the curve, even though the outlet temperature continues to rise, the concentration appears to fall. 4. Conclusions A series of experiments were conducted to investigate the transient characteristics of a methanol reformer and the following conclusions were drawn. After cold start, the higher the heating power and the methanol supply rate, the quicker was the cold start reaction. The best operating condition of these tested for the reformer was a heating power of 96 W, a heating temperature of 8 C, a methanol supply rate of 14 ml/min and an air supply rate of 7 L/min. This operating combination required only 22 s until hydrogen was being produced, and the production stabilised between 4 and 5 min. The outlet temperature of the catalyst when hydrogen commenced production was approximately 1 C, and at maximum hydrogen production, the temperature was 35 C, approximately. By combining different heating powers with different heating temperatures, methanol supply rates and air supply rates, the reformer can eventually achieve the state of stable hydrogen production. During the stable condition, it was found that the higher the methanol supply rate, the higher was the outlet temperature of the catalyst as well as the production concentration. Further, with a change of O/C ratio, a peak existed in the concentration of produced. Therefore, a specified O/C ratio must be combined with appropriate methanol supply rates in order to achieve a satisfactory production efficiency of by the catalyst.

15 R.-F. Horng / Energy Conversion andmanagement 46 (5) Acknowledgement The author gratefully acknowledges the support of the National Science Council of Taiwan, ROC, under grant NSC E Thanks are also due to the Heterogeneous Catalysis Department in the Union Chemical Laboratories of the Industrial Technology Research Institute for providing the catalyst in the experiments and technical support. References [1] Cheng YT, Yang CK, Su HT. Electricity generation techniques and development of fuel cell. J Energy 1995;25: [2] Han JS, Kim IS, Choi KS. Purifier-integrated methanol reformer for fuel cell vehicles. J Power Sour ;86: [3] Cheng YT. The design and new development of reformer for fuel cell. J Energy 1993;23: [4] Sung LY. Test of methanol reformer for fuel cell. J Energy 1994;24: [5] Sung LY. Test and analysis of reaction of methane steam reforming for fuel cell. J Energy 1994;24: [6] Chen HC. Study on the hydrogen generation by the methanol reformer for a PEM fuel cell. Master Thesis of Department of Aeronautics and Astronautics of National Cheng Kung University, Taiwan, 2. [7] Hohlein B, Boe M, Bogild-Hansen J, Brockerhoff P, Colsman G, Emonts B. Hydrogen from methanol for fuel cells in mobile systems: development of a compact reformer. J Power Sour 1996;61: [8] Emonts B, Hansen JB, Jorgensen SL, Hohlein B, Peters R. Compact methanol reformer test for fuel-cell-powered light-duty vehicles. J Power Sour 1998;71: [9] Wiese W, Emonts B, Peters R. Methanol steam reforming in a fuel cell drive system. J Power Sour 1999;84: [1] Nagano S, Miyagawa H, Azegami O, Ohsawa K. Heat transfer enhancement in methanol steam reforming for a fuel cell. Energy Convers Manage 1;42: [11] Takeda K, Baba A, Hishinuma Y, Chikahisa T. Performance of a methanol reforming system for a fuel cell powered vehicle and system evaluation of a PEFC system. JSAE Rev 2;23: [12] Choi Y, Stenger HG. Fuel cell grade hydrogen from methanol on a commercial Cu/ZnO/Al 2 O 3 catalyst. Appl Catal B: Environ 2;38: [13] Holladay JD, Jones EO, Phelps M, Hu J. Microfuel processor for use in a miniature power supply. J Power Sour 2;18:21 7. [14] Lindstrom B, Pettersson LJ. Development of a methanol fuelled reformer for fuel cell applications. J Power Sour 3;118:71 8. [15] Loffler DG, Taylor K, Mason D. A light hydrocarbon fuel processor producing high-purity hydrogen. J Power Sour 3;117: [16] Abernethy RB, Benedict RP, Dowdell RB. ASME measurement uncertainty. Trans ASME: J Fluids Eng 1985;17:161 4.