Flow at an open-channel confluence is characterized by a free shear layer formed between converging flows and a separated shear layer formed from the downstream side of the junction. These shear layers are closely related to the formation of secondary flows at the confluence, which is of great importance in river engineering. In this study, three-dimensional turbulence characteristics of the flow at a right-angled open channel confluence accompanying these flow structures are investigated by turbulence measurements and a visualization of the near-wall flow by an oil-film method. From the experimental results, existences of a descending secondary flow along the free shear layer and a strong helicoidal flow accompanying a weak counter-rotating secondary flow along the separated shear layer are verified.

View full abstract

Laboratory experiment in 60-degree confluence was conducted using wellgraded sand under active bedload condition for a discharge ratio of 0.6. The equilibrium channel bed was fixed and detailed measurements of mean flow were then carried out. Thereby, bed topography, two component velocity measurements, bed shear stress distribution and two counter rotating secondary flow cells are reported. Also, flow mechanism and bed morphology development in a confluence are explained. Bed shear stress maxima appear over the upward slopes along which bedload is transported away from the scour hole.

View full abstract

It is one of the most important subject in river improvement works to predict bed erosion. However, there have been few theoretical studies and the qualitative mesurements on bed erosion especially in actual rivers. In this paper, the relationship between hydraulic factors and the bed erosion is studied using river data. It is found that the tendency of the bed erosion can be expressed by two factors, which are the hight of bar and degree of meander. Using this expression, the bed erosion can be predicted very easily even at local offices at the site.

View full abstract

Many failures of levees due to running water are started with the erosion of the surface of levees by running water in an early stage of the failure. From the results of analysis of disaster cases and the results of experiments, it was clarified that sodded levees having vegetation on the surface of levees have much higher strengths against erosion due to running water compared to earth levees having no vegetation. Therefore, one of the important factors governing the safety of levees against failures by running water is the erosion resistance of vegetation on the surface of earth levee. However, previous studies were little made on the erosion chracteristics of the earth levees having vegetation both domestically and overseas nevertheless the importance of this theme.
This paper will present a method of on-site erosion test using a new device for checking the erosion mechanism and characteristics of earth having vegetation, the erosion characteristics of sod on the slope of levee determined by said method, and a method of predicting the erosion-resisting strength based on the results of the above. Finally, the applicability of the prediction method will be investigated on the basis of survey of sodded levee erosion by a large flood.

View full abstract

The bed Scour near the outer bank in the curved reach of natural rivers is generally caused by the secondary flow and it tends to become exaggerated by the formation and movement of alternate bars, Experimental and theoretical works are conducted in order to investigate that vane works provide a useful means to reduce the scour depth even in such a curved channel. The lateral bed profile obtained by the theory is found to be in a fairly good coincidence with observed one and the PWRI vane functions well so as to reduce the scour depth produced by the secondary current and alternate bar formation

View full abstract

The vane work is a new hydraulic structure for protecting river bank from erosion and it is important to establish reasonable vane design method for proper use of vane works. The influence of the attack angle of secoundary flow towards vanes on scour depth of the bed near outer bank is examined and a correction factor f for it is introduced reasonably to determine the longitudinal spacing ofvanes based on the existing PwRI's design method. Existing two design methods of vanes, PYRI method and IIHR method, are compared from hydraulic and economical stand points on the basis of natural river data. It became clear that PwRI design method is more reasonable and economical than IIHR one, judging from all the aspects of vanes.

View full abstract

The optimum arrangement of the groyn has been determined mainly by hydraulic model experiments. This method has a weak point not to be able to determine groin arrangement to river courses with arbitrary alignment. At the atandpoint that the numerical model and the hydraulic model test are complemental to each other, three dimensional numerical model was developed to decide the optimal arrangement of the groin This model includes the influence of forces exerted by groins on the flow of rivers. The result of calculation corresponded so much to that of existing experiments with regard to the state of the three dimensional flow and the river bed change, and also this model enabled us to decide the optimum arrangement of the groin to the river with arbitrary alignment.

View full abstract

Two dimensional experiments with permeable groynes were carried out and theoretical analysis was tried in order to know functions of permeable groynes. We found the fact there are close connection among three factors of mean velocity (U₁), permeable velocitythrough groyne (v_Q) and separate length (L). v_Q is depending on U₁ - it takes two fold values correspoding to the same U1 under the certain condition for the porosity of groyne. On the other hand, increase of v_Q. reduces L, and L is weakly related in U₁.

View full abstract

Drag and lift forces acting at man-made blocks used in river bed and their surface resistances are studied experimentally. The results give relation between the surface resistance and placement way for each block. Regarding to the drag and lift, the values acting at single blocK differ from those at blocks composing the river bed.The study indicates effect of the blocks adjacent to the specific block substantially.

View full abstract

Sediment yield is characterized by many regional factors such as geology, geomorphology, climate and vegetation. The object of this study is to establish a method of surveying sediment yield process in a bare slope, and of analysis and evaluation of the characteristics of the slope. They clarified the characteristics of sediment yield: in particular that the erosion rate is affected by slope gradient, compressive strength of the bare slope, rainfall intensity and the freezing and thawing effect.

View full abstract

We had a localized torrential downpour near the Fukui-Gifu prefectural border in September, 1965 (we call this Okuetsu Rainstorm). Especially, Nishitani village in Fukui prefecture (after this disaster, residents in this village moved to another place and now this district is included in Ohno city) suffered crushing damage from this storm. Namely, in Nakajima district which was the main part of the Nishitani village, before the disaster the number of households was 154, but after the disaster, 58 households were swept away and 86 were buried. A distinguishing feature of this disaster was that the bulk of the houses in the main part of the Nakajima district situated on the debris fan which had been formed by the sediment runoff from the Kamata river was buried second floor deep in sediment but not swept away. Judging from the fact-finding survey and data analysis, these houses were buried by the sediment which outflowed and deposited from the Kamata river taking the type of the sediment flow rather than the debris flow.
In this paper, in order to reproduce the burying process of the Nakajima district by the sediment flow, sediment hydrograph of the Kamata river is evaluated by the calculation of the bed level variation at first, then by using this sediment hydrograph as a input boundary condition, simulation of the horizontally twodimensional bed level variation in Nakajima district is carried out, and it is found that the calculated results are comparatively in good agreement with actual phenomena.

View full abstract

A method to simulate reservoit sedimentation in steep channels, where a flow changes from super-critical flow to sub-critical flow, is developed in the consideration of non-equilibrium suspended load. Hydrauiic jump section is the junction point of forward and backward differences in the calculation of channel bed level. A method to determine this level is discussed. The deposition rate of non-equilibrium suspended load downstream of the delta of sedimentation is calculated by a Fourier series solution of the diffusion equation. The bed profiles calculated by using these methods are in good agreement with the experimental results.

View full abstract

Velocities and bottom shear stresses are measured, by an LDV, in flows over fixed-bed models, which correspond to instantaneous shapes of a sand bar during scouring. Two bed load tran sport formulas in non-equilibrium state are obtained in terms of the measured shear velocity with corrections of the effect of slope; one for erosion and the other for deposit. An empirical coefficient K_B in the equation of motion for sand is determined to be 0.02cm^-1.

View full abstract

This paper deals with three problems on the slope failure by rainfall. The variation of the profile of hill-side ground water with time is discussed by considering the mass conservation with the ground water and rainfall infiltrated into soil. The occurrence condition of slope failure is able to introduce by considering the slip safety factor obtained to use the simple divided calculation method on the arc slip surface. Then, a method to calculate the amount of sediment yields by slope failure is proposed.

View full abstract

The occurrence number of landslides due to heavy storms per unit area (250m*250m) is found to follow a negative binomial distribution with parameters k and p, in which the variance is greater than the mean. Taking note of the reproducing property of the parameter k against a same value of p, and taking account of the linear relationship between the mean number of landslides and the parameter k, an exponential regression model of k is proposed. The coefficients of the explanation variables of the regression model can be calculated by the maximun likehood method, and the confidence intervals of expectaions for the occurrence number and area of landslides are theoretically derived.

View full abstract

Lava flows are regared as Bingam fluid and its property is explained by rate process which is shown by Arrhenius formula. Then, viscosity and yield steress are shown by exponential functions of temperature.
The 1986 lava flows extruded by the fissuer eruption at lzu-Ooshima were repoduced by numerical calculation of two-dimensional unstady flow. The result of calculations fit closely actual phenomenon.

View full abstract

Two subjects are chosen for discussions; one is the flow characteristics in a meandering compound channel with flat rigid and equilibrium rigid beds, and the other is the resistance of flows in compound channels.
Experimental results show that the flow parttern, the profiles of water surface and velocity are strongly affected by energy or momentum transfer between in-bank and over-bank flows.
An energy conservation equation is presented in order to evaluate the effects of interaction between the flows in main channel and flood plain to the flow resistence. A method is developed to predict the flow-depths, the mean velocities or the flow-discharges of main channel and flood plain. Its applicability is verified by flume data as well as by sensibility analysis.

View full abstract

In a compound open channel, the interaction between main channel and flood plain flow is the most important problem to be clarified. This interaction affects the resistance to flow and sedimentation on the flood plain.
To carry on the research of the interaction, it is necessary to clarify the flow structure in this region. Strong secondary flow near the boundary, which was showed by Sellin and another researchers, is investigated in experimental compound open channels in this paper. In spite of elusiveness of the secondary flow, many attractive photographs are obtained by the refined techniques for flow visualization. Characteristics of the secondary flow and interaction between main channel and flood plain flow are discussed here.

View full abstract

A series of experiments are performed to measure the mean velocity, turnbulence intensity and boundary shear stress distributions in compound channels with rough floodplains. Three different floodplain widths, main channel widths, and heights are used to understand the effects of varing channel geometry, size, the ratio of flow depth to the main channel depth and roughness elements. From these results the strong relation is observed between the mean velocity structure, turbulence structure and shear distribution. It is found, especially, that the roughness element on the floodplains strongly effects the flow structure in compound channel.

View full abstract

A new model of discharge assessment for compound channel flows is proposed based on the characteristics of wall shear stress. It is originally assumed that there is no shear acting on imaginary interface plains between main channel and floodplains. Both subsections hold the equilibrium between gravity force and boundary shear force. The boundary shear force is a function of the floodplain width, main channel width and height, and the ratio of Manning's roughness on the floodplain to that in the main channel. It is confirmed that the proposed model can reproduce not only the discharge-stage curves obtained by authors' experiments but also another researchers' ones.

View full abstract

A mothod for calculating the velocity field in open channel flows for which boudary shear is an important factor. It is based on the assumption of a scalar eddy viscosity that is proportional to the shear velocity on the actual flow boundary. The 2-dimensional monentum equation produced by this turbulence closure is solved numerically. In a channel with a rectangular cross section, the flow structure depends on both the width- depth ratio and the wall-bed roughness ratio. For such channels, figures to evaluate the boundary shear stress distribution on the bed are provided. The validity of the theory, when applied to sediment transporting flows, is demonstrated using William's (1970) experimental data.

View full abstract

Flow resistance on lower regime river beds is discussed with the data of Colorado University given by Guy et. al. The Engelund method to estimate the flow depth in an alluvial channel for a given flow discharge is modified for the lower flow regime. The hydraulic conditions to classify flat, ripple, and dune beds are given by the relationship between dimensionless bed shear stress and Reynolds number with respect to sand diameter. Moreover relationships between total and effective bed shear stresses in a dimensionless form are modified for dune bed in the Engelund diagram, and newly proposed for ripple and flat beds, which were not presented in the diagram. The accuracy of flow depth estimation for a given flow discharge with the modified diagram is much improved than without modification.

View full abstract

A transient process of the Reynolds-stress distribution due to a change of the bed shear stress in an open channel flow is investigated and a relaxation model, or a convolution-integral model with a concept of an impulse response, is proposed. The impulse response of the Reynolds-stress distribution for the change of bed shear stress is determined by a classic wind-tunnel data for air flow under transition due to an abrupt change of bed roughness, and it is successfully applied to the open channel data of flow with an abrupt change of bed roughness without any modification of a relaxation parameter of the model. Furthermore, the applicability of the proposed model is inspected by the transient properties of the velocity profile and the turbulence intensities obtained from the flume experiments for open channel flows with spatially acceleration or deceleration.

View full abstract

Velocity measurements, using fiber-optic laser-doppler anemometer, were conducted in steep open-channel flows on smooth and incomplete rough beds. Bed slope was changed from 1/500 to 1/25, which covers Froude number from 0.5 to 3. In this condition, the law of the wall and turbulence intensity distribution were examined. The von Karman constant Kis probed to universal, but the integral constant A decreases as the bed shear stress becomes larger. This decrease of A can be expressed by the decrease of the damping factor B in the van Driest's mixing length distribution in the near-wall region. Turbulence intensity u' normalized by friction velocity U* decreases from the semiempirical universal formula with the increase of the bed shear stress, in the same manner as the velocity distribution.

View full abstract

There are many empirical and some theoretical formulas on fluid friction. However, the application limits of these formulas are not always evident and some of these formulas are very complex. Why are many formulas having very different forms like this proposed and why are those often very complex? The reason is that the existent friction factor does not become constant and the formula of wall shear stress is not always reasonable.
Hence, in this paper, the new equations of the friction and the definitions of its factor are proposed and it is shown that many phenomena are explained successfully by these new equations.

View full abstract

A X-type hot film anemometer has been used to measure the secondary flow and turbulent structure in a gradually expanded rectangular open channel flow. The width B of the flume expands linearly in the flow direction x (dB/dx=1/12.5). Velocity measurements have been done in cross sections where aspect ratio Ar (=B/H; H: water depth) equals 2, 4, 6 and 8. The following results in each cross section are shown: isolines of primary mean velocity U, secondary flow pattern, distribution of local shear velocity u_*b on the bed, isolines of turbulence intensity u', v', w' and Reynolds stress-uv.

View full abstract

Coherent structures in turbulent open channel flow are investigated using flow visualization and their images processing techniques. In this method, tracers distribution on Z-T plane, Y-T plane are processed from the streamwise view of flow visualization using fluorescent dye illuminated by a sheet of laser light. Through taking pictures of them by CCD camera, the images are acquired. After the images get binary processing, they are analyzed statistically. As the result, instantaneous images, time-averaged images, and hierarchy of the coherent structures in turbulent open channel flow are considered.

View full abstract

On sand bed rivers the author made simultaneous observations of the river-bed formation along witb tbe vertical distribution of turbidity at the time of flood flow, using the echosounder and turbidity meter, As a result of his observations it has become clear that there is a strong current curling up from the river-bed when the water-level is rising and the river-bed flattens out, whereas, when the water-levei is lowering and sand-waves with considerable height are developing, the same current does not appear but becomes rather a quiet flow.

View full abstract

The purpose of this experimental study is to measure the fluid force acting on the square timber laid on a revetment. By the orifice installed through the square timber, the heads of fluid, surface pressure, are measured to calculate the drag force and shearing stress. It's ascertained the drag coefficient C_D is almost constant on it throughout entire span. So, it's possible to estimate the drag force and flow resistance by determining the drag coefficient.

View full abstract

The properties of flow in vegitated channels are investigated through laboratory experiments. The water plants are simulated by a group of cylinder with the same height, and the turbulence measurements are conducted. The equilibrium velocity in the vegitated region of water depth is determined by the bed slope, the concentration and the diameter of plants. The free surface flow over the vegitation is faster that the equilibrium velocity for the inside of vegitated region, and some mechanism plays a role of momentum exchange to accelerate the flow in the vegitated region near the interface with the free-surface flow region. In order to clarify this mechanism are inspected the mixing length, the correlation coefficient of the turbulence motion, the spectral charactersitics of velocity fluctuation and so on.

View full abstract

The paper consists of three parts. I) Estimation model for projecting rock height distribution on the river bed; To normalize the figure of the bed which is extremely varies with river location, the projecting height distribution is calculated by astocastic model. II) Similitude for roughness coefficiant; To keep the kinematical similitude, it is necessary to satisfy, the following five restic condions, t_r=x_rh_r^-1 (Froude rule)(1), Q_r=x_rh_r^3/2 (discharge rule)(2), l_b=h_rx_r^-1 (geometric rule)(3), τ_or/ρ_r=R_rh_rx_r^-1 (boundary shear stress rule) or n_r=R_r^2/3x_r^-1/2 (Mannings rule)(4), α_r=1 (velocity distribution rule)(5), (r is the.ratio of the model to the proto type, ) If F_r, I_br and n_r are satisfied, it was substantiated to get Manning's n for the prototype. III) Hydraulic stracture of roughness: To estimate the represent roughness length forthe distributed diameter, a hydraulic stracture of the roughness was discussed.

View full abstract

Flow around large roughness elements in open channel flow is studied by flow visualization using hemisphere and hemicylinder in a flow. When water depth to roughness height ratio (H/R) equals 3, shed vortices reach the water surface and there occurs ‘boiling phenomenon’. The flow around a two-dimensional roughness (hemicylinder) was found similar to that of the hemisphere, and boiling phenomenon appeared in 5 rows. Near the reattachment point, the mean flow is upward in the “vortices head's” section and downward in the “leg's” section. Results show the effect of vortices' motion on the mean velocity profile.

View full abstract

The purpose of this paper is to study the charactristics of fluid forces acting on spherical roughness elements which placed densely and sparsely. By measuring the velocity distribution above the roughness elements, it is possible to determine the shearing stress on the rough wall with the aid of the logarithmic formula and hence the equivalent roughness length. On the other hand, Drag Force and Lift Force were calcurated from the surface pressure of the roughness elements which were measured directly. The integrated mean shear values have been compared with those estimated from directry measured Drag Force and the energy gradient.

View full abstract

The drag and lift forces acting on spheres protruding from a rough boundary are directly measured. The flow velocity at the top level of a sphere is most pertinent as the relative velocity to the drag and lift coefficients, C_D and C_L. C_D is constant at 0.4 in deep flows, whereas it is affected by the depth in shallow flows. C_L is about 0.08. The non-dimensional standard deviation of drag fluctuation varies with the protrusion ratio; however that of lift is roughly constant at 0.4. The probability density and the autocorrelation are well approximated with Gaussian and exponential functions, respectively. The fluctuations of the drag and lift forces are not greatly correlated.

View full abstract

lt is very important in Hydraulic Engineering as well as, in Fluid Hechanics to reveal the response of velocity profiles and bed shear stress to abrupt change of bed roughness in open chamlel flows. ln the present study, accurate velocity measurements over abrupt1y changed k-type roughness elements were conducted in open channel flows by making use of two-component laser Doppler anemometer. An inmer boundary layer develops from the boundary point (x=0) between smooth and rough beds. The velocity and turbulence distrlbutions in the inner boundary layer rapidly respond to abruptly changed roughness. of particular significance is that the bed shear stress overshoots immed iately downstream of the abrupt roughness change.This finding is very important to construct bed protection works in rivers.

View full abstract

The effects of a step change in bottom roughness on the water surface profile are discussed from the results of experiments using three kind of bed roughness and those of a numerical analysis based on the two step Lax-Wendroff scheme. Both results show that the transition of flow depth is classified into three types of flow patterns, mainly depending on the Froude number. Furthermore, the adaption length of turbulence intensity is definded by the calculated results of the k-ε turbulence model, being compared with experimental ones.

View full abstract

In addition to meandering of the water course, abutments, bridgepiers, rivetsent blocks, spur dikes and water intake facilities etc. are installed in the river channel. The flow in this situation is thought as that with detached large roughnesses. In the region where we have continuous roughness, for instance, piles, the velocity of the flow decreases, and the tractive force also reduces. Therefore in this region the suspended material deposite. But in case of discontinuous roughness, occasionally the bed is severely scoured in a pile region. In this study bed configuration and flow features with a detatched large roughness are experimentally investigated. Main flow variation in the longitudinal and the lateral directions is theoretically and experimentally examined.

View full abstract

The characteristics of turbulent flows over the two-dimensional dunes has been investigated in two methods, namely measurements with the aid of hot-film anemometer and numerical simulation by k-ε turbulent model.
The computed resuls of k-ε model agrees well with the measured data of mean velocity and turbulence intensities profiles.
Based on the calculated flow field, the motion of the suspended particle traced by Monte-Carlo simulation, in which Markov process model is adopted.
Finally the non-equilibrium dispersion process of suspended particle over the dunes is predicted based on the proposed simulation model. The profiles of the suspended particle concentration over dunes agrees well with the results of experiment, which indicates the applicability of present simulation to the sediment transport in the flow over non-uniform boundaries.

View full abstract

In this paper, the velocity distribution of secondary currents and their effects on suspended sediments concentration are discussed. The analysis of sencondary currents was developed on the basis of streamwise mean vortex equation by using non-slip condition as boundary condition. The theoretical curves show a fairly good agreement with experimental results. The spanwise variation of suspended sediments concentration in a open channel flow with stable secondary currents was clarified to maintain the similar form from experimental data and the information made the theoretical analysis easy. The theory predicted the spanwise undulation in suspended sediment concentration and it's vertical distribution in both upflow and downflow regions.

View full abstract

The availavility of the ω-ψ method to simulate the flow field in an open channel with gradually varied bed slope was studied. But the eddy viscosity, included in the simulation and very important factor in hydraulics, could not be well-considered at that stage, since, for such flow, the boundary condition was dominant rather than the eddy viscosity. In this paper, the fundamental feature of that factor is studied through the simulation of the two-dimensional boundary layer flow on the flat plate in an open channel. For the turbulent boundary layer flow, the relation between eddy viscosity as a mean value of the flow field and mesh size or Reynolds number is made clear. On the other hand, for the laminar flow, the apparent relation among such factors is not found.

View full abstract

This paper deals with the flow behaviour and associated dispersion at a river confluence by the two-dimensional depth-averaged models. Basic mathematical models are derived in the curvilinear coordinate system and their finite difference schemes are presented. The developed models are applied to the analysis of the transverse mixing of contaminants in the Y river system. Its validity is successfully examined by comparing the numerical results with measured ones, and the flow behaviour and the spreading of concentration under some flow regimes are obtained by the numerical experiments. The relationships between flow behaviours and bathymetry are discussed, and the mixing process in the study area are also disclosed.

View full abstract

Turbulent diffusion mechanism in the modelled turbulence field is considered by using the relation between Lagrangian correlation coefficient and Eulerian space-time correlation one proposed by Corrsin. As the results, the longitudinal turbulent diffusion coefficient is related to the transverse one in terms of the ratio of the turbulent intensities of each direction. The result is applied to the estimation of the distribution of diffusion coefficient in open channel flows with the Eulerian turbulent characteristics calculated by using the turbulence model in this paper.

View full abstract

Sediment transportation in pipes can be divided into two modes; individual and massive movements as well known in the case of open channels. Disscusions are made on the criterion dividing the regime between these modes, and on the relation among flow discharge, energy gradient and sediment concentration, with attension forcused on the mechanisms of sediment transportation. The formulas which are developed are tested by flume data and compared with the results obtained from several empirical formulas.

View full abstract

This paper concerns with the scale-up method of pressure loss in coarse solid-liquid two-phase flow in horizontal pipes. In general, for economic reasons, the transport velocity is determined in the range of the flow with a moving bed. Therefore, the published data on the flow with a moving bed were analyzed, and an empirical equation to scale up the pressure loss (method I) was obtained. It was confirmed that calculated results agreed with the experimental ones within 30%. Furthermore, a scale-up method of pressure loss (method II) was established based on the moving bed model. This method gave the accurate estimations of pressure loss. However, it involves two unknown variables, so the experiments have to be conducted using two different kinds of pipes. The previous method for fine particles (so called is method) was modified to be applicable regardless of the particle diameter (method III).

View full abstract

Friction facter and boundary shear stress distributions are ditermined for rectangular channels of a wide renge of Reynolds number and a renge of width-to-depth ratios (B/H) 0.5 to 10. The influence of wall effect on the boundary shear and maximum-velocity point occuring below the water surface are examined. The boundary shear distributions of open channels are discussed indetailed. By using the maximum-velocity point, the calculation method of the boundary shear distributions are proposed.

View full abstract

A series of simplified experiment were carried out in order to study a dynamic effect of a free surface. The behaviors of vortex pair near the free surface was compared with that of nearthe fixed boundary. The velocity fields were measured by Laser Doppler Anemometer, from which the vorticity distributions were calculated. A visualization technique was also utilized to observe the vortex moving. The following was confirmed; the presence of the free surface not only restricts the movement of vortex like a fixed wall, but also results an energy conversion from to to waves.

View full abstract

In numerical simulations of a rapidly varied turbulent flow in an open channel, there are several difficulties to be solved., They are;(1) determination of the position ofthe free surface, (2) method of expressing the bottom boundary of a complicated shape, (3) modeling of the turbulence, and (4) effect of the curved streamlines on flow and tubulence characteristics. The Navier-Stokes equation, the Grid Generation Method, and the Reynolds equation with the K-ε model are applied. Numerical results are compared to the experimental data obtained with an LDV. Mean velocity profiles and bottom shear stresses are simulated reasonably well whereastc ands are not partly due to the accuracy of the measured data.

View full abstract

During a flood, it is very difficult to control a dam properly. In this paper, the authors propose the practical expert system to support centraized dam control using the artificial intelligence which contains a fuzzy inference engine. The system consists of twosupporting functions, i. e. one is to make a schedule of dam release and the other to decidea gate opening of next step. Asking veteran dam operators their know-how of dam operation, the structure of knowledge base has been determined. Part of the functions have been tested with the real data concerning the dams on Nahari River in Shikoku-Island. The results of inference approximately coincide the real operations, and are satisfactory.

View full abstract

The hydrologic information networks such as FLICS has been recently expanded across the country. The effective use of these informations becomes a serious matter of great concern.
This paper shows two fuzzy inference methods to predict runoff. First methodapplies the following conditional propositions.
If X is A and Y is B then Z is C.
If X is A' and Y is B' then Z is C'.
A, B, C, A' and B' are given fuzzy concepts and C' is an unknown one. Second method uses the linear regression model which include fuzzy concepts. These two methods are examined in a small basin and good results are obtained.

View full abstract

We design a flood control support environment, which is composed of the Inference System (IS) and the Procedural Knowledge System (PIS). PKS deals with problems which can be solved mathematically, while IS takes charge of problems which needs a knowledge-based approach. IS is composed of self-contained knowledge-based systems, each of which is an expert system to deal with a partial issue of flood control problems. Knowledge-based systems are classified into two types according to the problem type to be treated: a system which has a deterministic inference process based on a production system and one with a fuzzy inference process.

View full abstract