In this work, we propose a new unified-solution algorithm for solids and fluids using generalized simplified marker and cell (GSMAC)-finite element method (FEM). We use the Lagrangian description for displacement vector fields in solids, and, the Eulerian description for velocity vector fields in fluids. We carried out the analysis for static beam bending problems and dynamic viscoelastic response problems in order to the verify the validity of the present scheme for the solid phase. Additionally the analysis of self-excited oscillations of a cylinder in uniform flow is performed as a coupled vibration problem between the fluid and solid.
The appropriate timing to start controlling a dynamic stall was studied to find an optimum timing to start control just before separation and a way to reduce the total amount of energy to suppress separation. The dynamic stall of an airfoil in pitching-up motion was adopted because the time when separation occurs can be determined beforehand. The timing of control was defined as the interval from the start of control to the onset of separation without control. A wall-jet ejected from a thin slit near the leading edge of the airfoil at a constant velocity was used to suppress the separation. Many combinations of the timing of ejection and velocity of the jet were tested to determine the optimum amount of energy required to suppress the stall. It was found that, within the limits of our experimental conditions, there exists an optimum combination.
The performance of an active feedback control with wall deformation was tested by direct numerical simulation of a fully developed turbulent channel flow. The local wall movement is determined based on the flow condition detected by virtual sensors distributed in the vicinity of the wall, i.e., the wall velocity is made opposite to the wall-normal velocity at y/δ=0.1. The turbulent friction drag is reduced by 10% on average. The gain in the pumping power reaches about 28 times the energy consumption for actuating the wall. The primary mode of the resultant wall velocity distribution, which should be most effective in the present drag reduction scheme, is found to have wavelengths of about 300 and 50 viscous wall units in the streamwise and spanwise directions, respectively, while the time period is of the order of the time scale of the quasi-coherent vortical structure of near-wall turbulence. The effects of active wall deformation on quasi-coherent structures are investigated by a conditional averaging technique. In the present control scheme, the Q2 vortex is displaced away from the wall and the wall-normal and spanwise velocity fluctuations associated with the Q2 event are decreased. On the other hand, the location of the Q4 vortex from the wall remains unchanged while the vorticity of the Q4 vortex is substantially decreased. In the region downstream of the Q2 event, the wall is deformed in the shape of a shallow groove, which stabilizes the near-wall streaky structures along the groove.
Three-dimensional Eulerian air velocities and Lagrangian particle trajectories are numerically simulated to describe the effect of particle existence on a high Re number (Re=104) gas-particle turbulent jet using two-way coupling and Large Eddy Simulation in which the effects of particle existence on subgrid-scale flows are taken into account. The calculated results of air and particle turbulence characteristics (mean velocity distributions and turbulence intensity distributions) are in good agreement with experimental data obtained using a laser Doppler anemometer. Comparison of the instantaneous air vorticity isocontours of gas-particle and clean air jets reveals the production of vortices and eddies in both initial and transitional regions and the reduction of air turbulence in the developed region by the presence of particles. Based on the model for the effects of particle existence on subgrid-scale flow, states that reduce or enhance air turbulence in high Re number gas-particle flow are discussed.
In single phase flow, the fluid forces may be considered to be fully correlated along the tube span. This is not the case for two-phase flow. For tube lengths much longer than the typical linear dimension of the two phase flow unit structure (slug, bubble etc.), forces along the tube span are only partially correlated. In this paper, an unsteady flow model taking into account partial spanwise correlation of the fluid forces is presented. The model is applied to fluidelastic instability analysis of a long span heat exchanger tube model. It is shown that the effect of partial correlation is to reduce the effective net fluid forces which thus raises the effective instability threshold. Another important result is the introduction of additional coupling between modes that would otherwise be uncoupled in the fully correlated case. This has important implications for heat exchanger tubes which have closely spaced modes. Lastly, partial correlation is found to eliminate multiple instability boundaries at low values of the mass-damping parameter; this is a practically significant result.
Transitions of flow in periodically grooved channels and pressure drop characteristics are numerically investigated by assuming two-dimensional and fully developed flow fields. It is confirmed that a self-sustained oscillatory flow occurs at a critical Reynolds number from the steady-state flow as a result of Hopf bifurcation due to instability. The critical Reynolds numbers are obtained for various channel geometries. The ratio of the pressure drop of the grooved channel to that of the parallel-plate channel is also investigated. It is shown that the ratio is less than unity for the expanded channel geometries for the subcritical Reynolds numbers, whereas it increases above unity for the supercritical values. On the other hand, it always increases above unity for the contracted channel geometries.
Erosion wear at various locations inside the volute casing of a centrifugal slurry pump has been measured for the flow of solid-liquid mixtures. The measurements show that the wear increases all along the volute periphery with increase in the amount of solids suspended in the mixture. It is also observed that the wear is smaller when the pump operates near the best efficiency point flowrate (BEP) compared to that at lower flowrate.
The violent motion of cavitation bubble near a solid boundary is treated by a newly developed numerical code based on the CIP (Cubic Interpolated Propagation) method. This is the first study to successfully simulate the repetitive collapses and rebounds of bubbles with the penetration of microjets until the second collapse while maintaining toroidal shape. The obtained gas-liquid interfaces show similar tendency to bubble shapes observed in past experimental studies, although phase change, surface tension and viscosity are neglected. The key factor of this flow field is the density difference between gas and liquid. The velocity of microjet and resulting water hammer pressure are calculated successfully. However no shock wave appears from the collapsed bubbles in the present study, which suggests that an appropriate model for highly-compressed gas is important for tempestuous two-phase flow. This study indicates a high possibility of using the Euler-type numerical code to calculate such complicated two-phase problems and also suggests which factor is dominant in the cavitation bubble dynamics.
This study deals with a method to simulate transient responses of tapered fluid lines with rigid pipe walls. The fundamental equation employed in this analysis is a transfer matrix based on a fluid line model in which the frequency-dependent effects of viscosity are considered. For the convenience of simulation, we employed a method for modeling irrational transfer matrix elements of the line by modal approximation. The resulting models were estimated as highly accurate by comparing them with the corresponding exact solutions in frequency domain. By applying these approximation models to a block diagram representation, simulated results of transient responses were calculated. The simulated results of pressure and flow rate responses were compared with the corresponding analytical solutions. As a consequence, the validity and applicability of the proposed method were well confirmed.
Porous medium has got a great interest especially in the recent years, for its wide application in geophysics, petroleum, and air conditioning. Many studies related to porous medium were performed, and most of them are dealing with constant porosity. For the fact that porosity is non-uniform, a great concern has been directed toward the variable porosity studies. In this study, the effects of a double layer porous medium on the free convection along vertical plate embedded in this porous medium were investigated. The governing partial nonlinear differential equations were transformed into a set of ordinary differential equations, which have been solved by the fourth-order Runge-Kutta method. The results are obtained for different layer permeability K, and layer length, L. One case is compared with the previous study, and the result is found to be in good agreement with the result of previous study. Results show that permeability ratios greater than one tend to increase in Nusselt number, and it is valuable to use a high permeability ratio layer in the range of 0.2η to get higher heat transfer rate instead of using constant permeability medium.
Convective heat transfer from the wall surface of a cavity to the external stream is experimentally studied in the Reynolds number range from 10000 to 100000. Temperature distribution of the fluid in the cavity is almost uniform, except in the thin region immediately adjacent to the wall surface of the cavity. The temperature of the fluid in the cavity is represented by the average temperature of the fluid in the cavity. Namely, convective heat transfer from the wall surface of the cavity to the external stream is treated with the heat transfer model comprising the following two phenomena: one is the heat transfer from the wall surface to the fluid in the cavity, and the other is the heat transfer from the fluid in the cavity to the external stream. The vertical wall surface temperature of the cavity decreases from the bottom to the opening of the cavity. The effects of temperature nonuniformity on the wall surface of the cavity for convective heat transfer are also reported.
This paper describes the detailed heat transfer distributions of an atomized air-water mist jet impinging orthogonally onto a confined target plate with various water-to-air mass-flow ratios. A transient technique was used to measure the full field heat transfer coefficients of the impinging surface. Results showed that the high momentum mist-jet interacting with the water-film and wall-jet flows created a variety of heat transfer contours on the impinging surface. The trade-off between the competing influences of the different heat transfer mechanisms involving in an impinging mist jet made the nonlinear variation tendency of overall heat transfer against the increase of water-to-air mass-flow ratio and extended the effective cooling region. With separation distances of 10, 8, 6 and 4 jet-diameters, the spatially averaged heat transfer values on the target plate could respectively reach about 2.01, 1.83, 2.43 and 2.12 times of the equivalent air-jet values, which confirmed the applicability of impinging mist-jet for heat transfer enhancement. The optimal choices of water-to-air mass-flow ratio for the atomized mist jet required the considerations of interactive and combined effects of separation distance, air-jet Reynolds number and the water-to-air mass-flow ratio into the atomized nozzle.
A numerical study was made to identify the most appropriate criterion to describe the spontaneous ignition of an externally heated solid material. Several ignition criteria were examined to determine whether they can represent correctly the effects of the gravity and ambient oxygen concentration on the ignition characteristics. Gravity significantly affects the ignition behavior since the buoyancy induced flow plays a key role in the transient mass and heat transfer processes, leading to two distinct types of ignition. Seven ignition criteria, which were used in a 1-D analysis of ignition process, were examined in this study. It was found that some of the criteria based on temperature profiles lead to erroneous predictions of ignition. This is because complicated transport phenomena caused by the buoyancy induced flow control the transient temperature profile, producing a situation where the ignition criteria were apparently satisfied. The present study shows that the criterion based on the local reaction rate is the universal and most appropriate one.
For achieving a stable premixed combustion, there is a device termed the cyclone combustor, which consists of a cylindrical chamber and fuel nozzles installed tangentially on the sidewall. In this combustor an extremely stable flame can be obtained in the swirl flow, formed along the inner wall of the combustor. The authors utilized this combustor as a flame holder, to burn a high velocity jet flowing axially in the central part, and termed this new combustor a cyclone-jet combustor. In the present study, an excellent flame stability is shown for the cyclone-jet combustor and, by comparing premixed, non-premixed and partially premixed flames, the low NOx combustion characteristics were experimentally examined for this combustor.
The introduction of inexpensive cylinder pressure sensors provides new opportunities for precise engine control. This paper presents a spark advance control strategy based upon cylinder pressure in spark ignition engines. It is well known that the location of peak pressure(LPP) reflects combustion phasing and can be used for controlling the spark advance. The well-known problems of the LPP-based spark advance control method are that many samples of data are required and there is loss of combustion phasing detection capability due to hook-back at late burn conditions. To solve these problems, a multi-layer feedforward neural network is employed. The LPP and hook-back are estimated, using the neural network, which needs only five output voltage samples from the pressure sensor. The neural network plays an important role in mitigating the A/D conversion load of an electronic engine controller by increasing the sampling interval from 1° crank angle (CA) to 20° CA. A proposed control algorithm does not need a sensor calibration and pegging (bias calculation) procedure because the neural network estimates the LPP from the raw sensor output voltage. The estimated LPP can be regarded as a good index for combustion phasing, and can also be used as an MBT control parameter. The feasibility of this methodology is closely examined through steady and transient engine operations to control individual cylinder spark advances. The experimental results have revealed a favorable agreement of optimal combustion phasing in each cylinder.