Isothermal chamber is filled with a certain density of high thermal conductivity porous material, which has a wide range of applications in the flow measurement of pneumatic field, and its isothermal characteristic is critical to its application results. In this paper, in order to improve the numerical calculation accuracy of the isothermal characteristics during discharge process, a discharge model of isothermal chamber with fractal effective thermal conductivity (ETC) for determining the isothermal characteristic is reported. Firstly, the stuffer in isothermal chamber is considered as porous random fiber bundle, and an ETC prediction model of anisotropic porous random fiber bundle in both vertical and horizontal directions is established by fractal theory. This model with two directions is in good agreement with the experimental results, and the relative root mean square errors (RRMSE) are 3.94% and 9.85%, respectively. Secondly, the discharge model with fractal ETCs is built, and the isothermal characteristics of isothermal chambers with three different porosities are determined by numerical simulation. Finally, experiments to determine the isothermal characteristic are carried out. The numerical simulation results are in good agreement with the experimental results, and the relative errors are less than 3%. It could be concluded that accurately determining the ETC of the stuffer in isothermal chamber can improve the numerical calculation accuracy of the isothermal characteristic. Moreover, compared with the experimental method, numerical method is energy-saving and timesaving.

The Soret effect is a phenomenon in which the components of mixed fluids are separated by a temperature gradient. This study investigated a simple and low-cost technique for extracting high concentration hydrogen (H₂) by utilizing the Soret effect in a H₂-carbon dioxide mixed gas. Previous research by the authors attempted to improve the efficiency of H₂ separation by applying a temperature difference to a network of separators with a continuous structure (the Burgers cascade) and enabling a mixed gas throughflow. One of the problems in that study was the impairment in separation performance that occurred due to the remixing of the components separated by flows within the device cells. Therefore, to resolve this problem, the present study conducted numerical simulations for gas flow, temperature distribution, and H₂ concentration and studied the effect of inserting an additional structure (a partition) to prevent gas remixing at first. Based on the simulation results, experiments were conducted using the Burgers cascade in a device with partitioned cells. The results confirmed that the partitions improved the separation of H₂ by up to about 1.6 times compared with conventional devices.

This study aims to investigate the hot-smoke dispersion behavior released from a chimney in a cross flow experimentally using specially designed wind tunnel. An artificially obtained quasi-isotropic turbulence was generated using an active turbulence generator developed by Makita. A heated jet and an unheated jet with smoke are injected into the cross flow from the vertically-oriented chimney installed in the test section. Smoke motion was captured by high-speed camera to obtain instantaneous patterns of the smoke dispersion. Six-kinds of the featured patterns are clearly identified, such as; (I)(II) bifurcated vortex tubes with and without a strong mutual interaction, (III) connected hairpin-type vortices, (IV)(V) the mixture of the coherent and turbulent vortices, (VI) downwash-type structure. These smoke pattern in the downstream field from the chimney are found to be depended on buoyancy, turbulent motion, and inertia forces. Under the quasi-isotropic turbulence, the smoke dispersion prefers to exhibit a meandering pattern (Mode V) under wide range of adopted flow velocities, which is hardly observed using the grid turbulent test device. Interestingly, as compared to the unheated jet, the meandering smoke structure with heated jet ejection was also observed even at the lower jet velocities and the higher cross flow velocities, suggesting that the buoyancy force shall play an important role on appearance of meandering motion and control the smoke dispersion. Direct smoke exposure case (Mode VI) is preferred to be observed when the quasi-isotropic turbulence is imposed, although the trend of appearance depending on the jet temperature can be predicted even using grid turbulent device. It is concluded that using grid turbulence would not be suitable to predict on the smoke dispersion problem in the actual scale. The time averaged smoke concentrations profiles are analyzed at locations along the cross flow direction and it is revealed that the effective diffusion becomes stronger when the quasi-isotropic turbulence is imposed in the cross wind. Further, it is confirmed that the smoke dispersion behavior can be well-characterized by the existing prediction method based on the point source model.

The level-set approach extended for its viscosity solution is investigated to derive a relation to the conservation law of fluid phenomena and the phase-field approach based on the free energy theory. This mathematical approach is useful to consider approximate models for fluid interface problems. Here, this approach is applied to the thickened interface model of the combustion flame problem to derive a new mathematical formulation by the viscosity solution of the level-set equation, which can maintain a flame propagation speed of global solution. This approach reveals a mathematical incompatibility of empirical thickened flame models and proposes some improvements.

An experimental investigation on effects of aluminum particle size on the combustion characteristics of aluminized composite propellant was performed. Reducing the agglomeration size on the burning surface is important to reduce slag in the solid rocket motor and increase the combustion efficiency. Four propellants (HTPB 15 wt%, Al 17 wt%, and AP 68 wt%; AP distribution is trimodal) with the same composition except for different in aluminum particle sizes were examined. Average particle sizes of 30 μm, 10 μm, 5 μm, and 2 μm were considered. Combustion tests were examined with an N₂-flashed strand burner with two windows. A high-speed imaging technique was used to observe the formation and ejection of agglomerations on the propellant burning surface at the atmospheric pressure. The agglomeration sizes and its distributions were measured. All distributions were monomodal, and the mean and peak diameters of the agglomerations reduced with decreasing diameter of ingredient aluminum particles. The propellant strands were burned at pressures ranging from atmospheric pressure to 5 MPa. The burning rate of propellants were measured on the basis of the movement of the combustion surface acquired by a video camera. The propellant with 2 μm aluminum particles showed a smaller agglomeration size and higher burning rate (17% higher at 5 MPa) than did the other propellants. 2 μm aluminum particles seem to be effective for decreasing agglomeration size.

This study deals with different machine learning algorithms modeling of small-scale burner to predict the temperature from the top and bottom of the flame from combustion of ethanol-diesel fuel blends. The data used for training and testing of the proposed algorithms was acquired by combusting ethanol-diesel fuel blends at different fuel mixing proportions, inner diameters of quartz tubes, flow rates of air and volume flow rates of blend in small-scale burner. Three models for machine learning algorithm based on back propagation (BP), generalized regression neural network (GRNN) and support vector machine (SVM) for small-scale burner were established employing the experimental data for training and testing. The feasibility of these algorithms in predicting the flame temperature of ethanol-diesel combustion in small scale were contrasted by performance correlation coefficient (R) and mean absolute percentage error (MAPE). The results showed that all three algorithms could well identify the complicated nonlinear relationship between the flame temperature and related variables. In addition, for the top temperature, the R value of the SVM model was the largest, which was 0.98513, and the MAPE value was the smallest, which was 1.60%. And, for the bottom temperature, the R value of the SVM model was the largest, which was 0.98135, and the MAPE value was the smallest, which was 1.84%, similarly. Therefore, the SVM model could predict the flame temperature of small-scale burner well, and its performance was the best among the three machine learning algorithms.

Renewable energy-based microgrid systems are widely being studied as electrification methods for rural communities in developing countries. Waste heat generated by the components of the microgrid systems, such as the biogas driven generators (BDG), presents the potential of utilizing the low-grade heat in a way that can contribute to the sustainability of such energy systems. From the points of view of affordability, local manufacturability, and applicability for agriculture, thermally driven pumps (TDP) may be attractive for coupling with such microgrid systems. Therefore, the current study has focused on the development of a new type of thermally driven pumping system as a potential waste heat utilization component for microgrid applications in rural areas. A liquid piston-type TDP concept without moving parts, except few valves, was developed and parametric experimental investigations were carried out. The performance and characteristics of the system were studied, which revealed that the proposed system has a superior performance compared to the literature. It was also found that the system performance strongly depends on the heat addition rate and delivery capacity of the system, which are suitable characteristics for the intended application. Hence, the experimental data were used to estimate whether the proposed system can pump enough water that needs to be supplied for the biogas production to supply a 10 kW BDG unit of a microgrid. It was found that 87 - 93% of the total pumped water (13 - 27 m³) would be available for agricultural and other purposes while only 6 - 13% would need to be fed to the biogas digester. Generally, the results seem to be promising, and yet there are potentials for the optimization and improvement of the proposed system, hence they have been pointed out.

Microscale hydrogen (H₂) combustion is one of the promising technologies for renewable miniaturized heat sources. This study analyzes the oxygen combustion of H₂ in small-scale counterflow burners, with carbon dioxide (CO₂) added for safe hydrogen treatment (flame visualization and reduction of flame propagation velocity). The effects of burner inner diameter, burner gap, and gas flow rate on the flame shape/size (thickness and diameter) are measured through flame image analysis. The experimental results show that the flame thickness and diameter monotonically decrease with a decrease in the burner inner diameter, burner gap, and H₂ flow rate. The flame thickness decreases with an increase in the flame stretch rate, and the approximate curve representing this relationship varies depending on the burner inner diameter and H₂ flow rate. Accordingly, the flame thickness normalized by H₂ flow velocity and burner inner diameter is newly proposed, which strongly correlates with the flame stretch rate and converges on a single line, i.e., inverse of the square root of the flame stretch rate. These findings are also applicable to biogas (CH₄-CO₂ mixture)-O₂ micro counterflow diffusion flames with the same CO₂ concentration in the fuel gas and apparent equivalence ratio.

An oscillating heat pipe (OHP) is a highly efficient cooling system for densely integrated electronic and electric devices operating at high frequencies with high heat generation densities. However, because of the complicated internal flow with phase changes, it is difficult to predict the heat transport performance of OHPs accurately. Such predictions are needed to understand the fundamental phenomena in heat transport and optimize the OHP design parameters. The objective of this study is to predict the three prediction targets comprising the internal flow pattern, wall temperature difference between the cooled and heated sections, and heat transport rate of the OHP through machine learning in recurrent neural networks. Experiments on OHP with ethanol were performed for the heat input range of 62-125 W to obtain time series data of the internal flow pattern images, wall temperatures, and cooling water temperatures. The internal flow pattern images were processed by semantic segmentation and subsequently used for training the models for each prediction target. The internal flow patterns were recursively predicted using the trained model. The predicted internal flow patterns were then input into the wall temperature difference and heat transport rate models to predict these two prediction targets. The predicted and experimental time series data for each prediction target were compared, and the prediction ability of the machine learning-based procedure was demonstrated by the quantitative agreement between the experimental and predicted statistical values.

The mixing process is often used not only to homogenize the materials, but also to bring out better functions by mixing different materials. In the mixing process involving chemical reactions, it is also important to quickly remove the by-products of the reaction. In this study, a flow field around a single rotor rotating in a partially filled chamber was numerically simulated using a 3D model to examine the interface renewal characteristics of highly viscous fluids. The fluids examined were glycerin, which is Newtonian fluid, and aqueous solution of carboxymethylcellulose (CMC), which is Non-Newtonian fluid. A three-wing rotor was chosen as the mixing rotor. The evaluation of interface renewal was performed using history particles. Also, the flow states were classified into rotational, shear, and extensional flows using the Flow number to analyze the flow field. The velocity distribution obtained by the simulation was in good agreement with the experimental results under the same geometry and conditions. For distributive mixing and interface renewal, CMC results were superior to glycerin. Flow number-based evaluations showed similar flow conditions around a single rotor, regardless of the physical properties of the test fluid. However, in the case of glycerin, fluid mixing between adjacent spaces separated by the rotor wing was suppressed and interface renewals were reduced. This is because, compared to CMC, glycerin does not have a decrease in viscosity due to shear-thinning properties, and the flow rate passing through the rotor tip is relatively small. It was also shown that there was a positive correlation between the flow rate passing through the tip part of the rotor and the interface renewal, regardless of the characteristics of the mixing fluids.

The efficiency, reliability, and safety of aircraft have been improved by substituting hydraulic and mechanical systems with electric systems. Furthermore, in future electric aircraft, where fan-driving engines are partially replaced by electric motors, solving the thermal management of heat generation from the motor-controlling power-devices will become a crucial problem. In this study, three-dimensional steady thermal network analysis (TNA) was carried out to analyze the temperature field in a heat sink, for motor-controller air cooling, which is expected to be used for the future electric aircraft. The numerical procedure used was verified by comparing the results of TNA with those of three-dimensional fluid-solid conjugate heat transfer analysis using finite volume method. After the verification, TNA was carried out for an aircraft flight scenario, where the optimum geometry of the heat sink was investigated, minimizing the following three objective functions: the pressure loss of air flow through the heat sink, the maximum local wall temperature of the heat sink at the surface of motor controller, and the weight of the heat sink. In this multi-objective optimization, the design of experiments technique was used to decide the space-filling points for the six design parameters. The three objective functions at each sampling point were calculated using TNA. A response surface was created for each objective function. A multi-objective optimization on the response surfaces, using a genetic algorithm, was iteratively carried out to find the Pareto optimal solutions for the air-cooled multi-stage heat sink. From the results of the Pareto optimal solutions, the effects of the design parameters on the objective functions were discussed. In addition, multiple regression analysis was performed for quantitative evaluation of the results of the Pareto optimal solutions, and the dominant design parameters for each objective function were identified.