To understand the flow maldistribution phenomena inside the honeycomb structure of a regenerative burning system, the results from an experimental study of flow distribution in the honeycomb were compared with computation fluid dynamics (CFD) results by means of the Lattice Boltzmann method, which is convenient for simulating fluid flow in complex boundaries. The experimental results of flow distribution agreed well with the CFD results. Although porous plates are often placed in front of the honeycomb structure in order to decrease the maldistribution, our simulation results suggested that porous plates have a limited or opposing effect on flow uniformization. To improve the effect on flow uniformization, it is important to adequately control the pore size distribution on the porous plate while keeping total porosity constant. Calculation of running costs of the regenerative burning system showed that the optimized pore plate can raise the efficiency, while the efficiency of the current honeycomb system was lower than that of the saddle system.
To resolve the flow maldistribution phenomena inside the honeycomb structure of a regenerative burning system, turbulent flow in the honeycomb structure was numerically simulated by the Lattice Boltzmann method. A new technique was proposed to automatically design the pore size distributions of the pore plate in front of the honeycomb while keeping the total porosity of the plate constant. In consequence, the technique had an effect on flow uniformization, especially in the case of horizontal inlet of flow with an adequate entrance region. The proposed technique also showed that the coefficient of variation, which represents the uniformity of the flow velocity, could be reduced to 0.3–0.4, while our target value for the coefficient of variation for 3/4 saddle was 0.5.
Flow patterns were visualized by the LIF method in order to elucidate the mechanism of advective mixing in a non-element mixer. Branch flows were injected periodically into the main flow through perpendicular branch pipes. The effects of the number of branch flows and the period of injection on the stretching and folding were investigated experimentally. The boundary length between two fluids L was evaluated in the cross section parallel to the main flow. It was observed that branch flow injection caused a backward flow immediately after the injection and it suppressed the increase in L, but after that L increased due to the increase in the folding. It was also found that the time variation in L depended on the period of injection. Though injected fluid flowed more slowly as the number of branches increased, L increased more due to the remarkable effect of stretching and folding.
Measurement of the particle size distribution of isopycnic floating particles using a solid–liquid fluidized bed was experimentally investigated. Experiments were conducted with a Glassbubbles-water system with a solid–liquid density ratio of 0.125, and a paraffin–24 wt% sodium chloride solution system with a solid–liquid density ratio of 0.760. Despite incidences of spillage of small particles from the fluidized bed, there was a close match between particle size distributions measured by this measurement method and size distributions of particles remaining within the fluidized bed measured by the laser diffraction/scattering method. The ranges of particle sizes which can be measured are approximately 3 : 1 or 5 : 1 in terms of the ratio between the maximum and the minimum particle sizes. By controlling the solid–liquid fluidized bed in a state of uniform fluidization, the particle size distribution of floating particles can be measured.
Wakkanai siliceous shale (WSS) is a natural meso-porous material. This study aimed to improve its capacity to adsorb humidity. The shale's surface was supported by chlorides by being dried at 150°C after impregnation with a chloride solution. The change in specific surface area, pore size distribution and the water sorption/desorption isotherm of the supported shales were investigated. From the water sorption isotherms, the amount of water sorption of WSS supported by LiCl 10% solution increased to 150 mg/g, while that of the natural WSS was 25 mg/g in the range of relative humidity from 30 to 60%. Similarly, WSS supported by NaCl adsorbed water about 7 times more water than natural WSS did. It was suggested that this improvement resulted from water transport into NaCl caused by capillary condensation of water in the WSS pores.
In a basic investigation of the processing of emission gas containing volatile organic compounds by vacuum pressure swing adsorption without purge gas, ethyl acetate and methyl ethyl ketone were employed as volatile organic compounds and mesoporous activated carbon and microporous activated carbon as adsorbents, and the effects of feed gas concentration on the lean gas concentration and of the column diameter on the time required to reach the cyclic steady state were measured. Feed gas concentration was found to have little influence on lean gas concentration, which remained constant after the cyclic steady state had been reached. This is attributed to the variation in the average adsorption ratio of the absorbent in proportion to the feed gas concentration. This phenomenon was more marked with mesoporous activated carbon than microporous activated carbon The effect of column diameter as evaluated from the mass-transfer coefficient in the column was found to be greater in the desorption step than the adsorption step, indicating the existence of an optimum column diameter.
The effect of pH of a nitrite solution on the depletion rate of nitrous acid was experimentally investigated. The depletion reaction of nitrous acid proceeded when the initial pH of a nitrite solution was less than 4.6. When the nitrous acid concentration was greater than 10−3 mol·l−1, the ratio of nitrite consumption to nitrate production was approximately 3. However, when the nitrous acid concentration was smaller than 10−3 mol·l−1, the ratio of the nitrate production to nitrite depletion became greater than 3 due to desorption of N2O4.
Catalytic activity in methanol synthesis from gas containing CO2 and H2, is very low compared with that in synthesis from gas with high CO content provided by steam reforming of natural gas. Therefore, it is necessary to develop a catalyst with high activity. When the molecular pore size of the methanol synthesis catalyst is smaller than the mean free path of the diffusing gases, Knudsen diffusion is dominant. Because the Knudsen diffusion coefficient is proportional to the molecular pore size, studies to improve the methanol synthesis catalyst were conducted by increasing the molecular pore size of the catalyst. The reaction rate and the apparent specific gravity of an improved catalyst with larger pore sizes and larger pore volumes than the original catalyst (main components: CuO–ZnO–Al2O3–Ga2O3–MgO) were investigated. The reaction rate per unit weight of catalyst was almost 1.2 times higher, and the apparent specific gravity per unit weight of catalyst was nearly 0.7 times lower for the improved catalyst compared with the original catalyst. DME synthesis using a hybrid catalysts consisting of the methanol synthesis catalyst and a methanol dehydration catalyst was undertaken using a bench-scale test plant. It was determined that: (1) the yield of improved catalyst was the same as that of the original catalyst, (2) the stability of the improved catalyst was the same as that of the original catalyst during 200 h of bench-scale testing, and (3) use of the improved catalyst contributed to a reduction of catalyst consumption, leading to reduced operating costs.
AlPO4 which was prepared by neutralization of an aqueous solution of Al nitrate was amorphous and has a large specific surface area as much as 130 m2/g after calcination at 1000°C. After the calcination at 1100°C, AlPO4 was slightly crystallized and the crystallite size of AlPO4 was about 16 nm and the catalytic activity was higher than that calcined at 1000°C. But AlPO4 calcined at 1200°C was more crystallized and the activity was decreased. On one hand, AlPO4 promoted by 10 mol% of Ce which was calcined at 1000°C showed a maximum catalytic activity and crystallinity was the same as that of pure AlPO4 calcined at 1100°C. Ce promoted AlPO4 after calcination at 1100°C was well crystallized, and the catalytic activity was decreased. Therefore, it was concluded that Ce as a promoter, creates small AlPO4 particles which was conveniently crystallized to CCl2F2 decomposition by the calcination at 1000°C which is 100°C lower than that in case of pure AlPO4. By treatment with concentrated CCl2F2 at high temperatures such as 800 or 900°C for 8 h, both AlPO4 and Ce promoted catalysts were much crystallized and SSA and the catalytic activity of the catalysts were decreased. The catalytic activity when it will be ultimately inactivated state was estimated using T50% approached at zero SSA, and 550°C will be required to obtain 100% CCl2F2 conversion for Ce promoted AlPO4 catalyst. AlPO4 was crystallized by a steam treatment at 900°C to the same level of that calcined at 1100°C in air and increased the catalytic activity. In case of Ce promoted AlPO4, steam treatment at 1000°C promoted the crystallization of CePO4 but AlPO4 and increased the catalytic activity. The effects of CCl4, CF4, and CCl2F2 treatment were studied and found that AlPO4 was the most crystallized by the treatment by CCl2F2. This may be explained by the formation of chloridefluorides such as AlCl3–nFn and PCl5-mFm which can be sublimated only in the decomposition of CCl2F2.
Mass transfer of an acetic acid-methanol system in a microchannel was measured by laser Raman spectrometry. The Raman peak intensities of the molecules were proportional to the concentrations for each species, and hence the concentrations could be quantitatively determined from the Raman spectra. By measuring Raman spectra at specific spots in the microchannel placed on the X–Y–Z moving stage, we obtained two-dimensional concentration profiles in a microreactor made of a Y-shaped channel. The observed concentration profiles of acetic acid and methanol in the microreactor with a large aspect ratio (defined as width/depth) were coincident with the calculated ones at several flow rates. We confirmed the reliability of both our measurement technique and CFD analysis method. The CFD calculation gave the concentration profiles in microreactors with different aspect ratios. The CFD analysis revealed that the deformation of the fluid interface strongly depends on the aspect ratio of the microreator. Deformation of the fluid interface was small for the aspect ratio >3 and <0.3, and large for the aspect ratio ≅蹅. We found that Y-shaped channels with aspect ratios larger or smaller than unity are suitable for measurement of mass transfer in microreactors by laser Raman spectrometry.
A correlative method for predicting a batch-type drying rate curve of coated film with a single volatile component upon drying with hot air at high velocity, with high humidity and in double-sided mode was proposed by using a batch-type drying rate curve for hot air of the same temperature with low velocity, low humidity and in single-sided mode. The proposed method was composed of the following three patterns: (1) changing hot air velocity, (2) changing supply mode of hot air, and (3) changing hot air humidity. No fitting parameter was included in patterns (2) and (3), while a parameter n was included in pattern (1). No transport property within the coated film was required in these applications of the correlative method. The correlative method was examined in comparison with numerically simulated results based on a numerical strict method previously proposed. Two examples were examined: a Fick-type mass transfer case and a non-Fick type case considering an accelerative effect of mass transfer caused by drying stress in the solidified part of the coated film during drying in addition to a Fick-type mass transfer. A predicting procedure and calculation were tried to obtain a batch-type drying rate curve for drying with hot air at high velocity, with high humidity and in double-sided mode from a batch-type drying rate curve for hot air of the same temperature with low velocity, low humidity and in single-sided mode. Results calculated by the correlative method proposed in this study agreed well with numerically experimental results. Accuracy of the approximation for the Fick-type case was higher than that for the non-Fick type case. A common value of the fitting parameter n was obtained from the both Fick-type and non-Fick type cases in pattern (1) of the correlative method, and the value may be correct in other cases.
Lotus-type metal is a newly developed porous material containing numerous unidirectional cylindrical pores, which is produced by unidirectional solidification of the melted metal in a pressurized gas atmosphere. To effectively employ lotus-type porous metal as a heat sink, it is necessary to clarify not only the heat conductivity but also the fluid permeability, which depends upon the porous structure. The Darcy permeability was measured from the water permeation test of lotus-type porous copper under extremely low pressure conditions. The true permeability determined by accounting for the effect of the penetration ratio of pores was well evaluated by applying the Hagen–Poiseuille equation to the fluid flow, using the fourth root mean fourth power value as the average pore diameter in view of the pore size distribution. Moreover, the effect of the pore size distribution on the permeability was analyzed in the case of the log-normal distribution, and it was shown that the permeability increased with increasing standard deviation.
We previously found that minute linear and cloud-shaped particles of mineral matter in coal were difficult to liberate from organic matter. Thus, quantitative information on the shape of mineral matter in coal is considered to be significant to the development of pulverization-ash removal processes. The present study proposes a new method to obtain quantitative data on the shape and size of mineral matter in coal using a common imaging tool, a photomicroscope, which can be used on-site easily. The mineral distribution in coal could be measured by distinguishing organic matter and mineral matter using luminance distribution curve obtained from gray scale images. As the shape index of mineral matter, the ratio of diameter maximum length to diameter was found to be appropriate. Samples of pulverized coal of different sizes were prepared. The effect of coal particle size, size of mineral particles in coal, and the shape index of minerals on mineral removal efficiency were evaluated by sink-float separation. The present indices were found to play significant role in mineral removal from coal.
Dry-type technology for recovery of fluorine, a precious resource for Japan, from exhaust gases containing hydrogen fluoride has been developed by using calcium absorbents. The gas–solid reaction properties of chlorofluorocarbon decomposition gases and calcium absorbents at low temperature were investigated with the type of calcium absorbent and the grain size as parameters. The reactive gases obtained by HCFC22 thermal decomposition, which were contain about 30% HF and about 15% HCl, were reacted with calcium carbonate or calcium oxide with grain sizes of 1–2 mm and 2–4 mm in the temperature range from 473 to 673 K. Under these conditions, the superficial velocity ranged from 4 to 6 cm/s and the space velocity (SV) from 1600 to 2300 h−1. The fluorine recoveries with calcium carbonate and calcium oxide differed greatly according to the reaction conditions and were greatly dependent on the reaction characteristics of each absorbent. It was suggested that formation of fine pores by CO2 generated with advance of the fluorination reaction contributed greatly to the formation of high purity fluoride. When calcium carbonate of grain size 1–2 mm was used as solid absorbent, fluorine recovery exceeded 95%. On the other hand, the maximum recovery of fluorine with calcium oxide was 80%.