【正文】 ctional areas open to flow in the subscript directions。 Savage et al., 2020). They showed that there is reasonably good agreement between the physical and numerical models for both pressures and discharges. Especially, Kim (2020) investigated the scale effects of the physical model by using FLOW3D. The results of numerical simulation on the series of scale models showed different flow discharges. Discharge and velocity of larger scale models has shown larger value than the smaller scale models. Existing studies using CFD model mostly deal with the model’s applicability to discharge flowrate, water surfaces, and crest pressures on the spillway. In this study, flow characteristics such as flowrate, water surfaces, crest pressures on the spillway, and vertical distributions of velocity and pressure in consideration of model scale and surface roughness effects are investigated in detail by using mercial CFD model, FLOW3D, which is widely verified and used in the field of spillway flow analysis. The objective of this study is to investigate quantitatively the scale and roughness effects on the flow characteristics by analyzing the putational results. 2. Scaling and Roughness A hydraulic model uses a scaled model for replicating flow patterns in many natural flow systems and for evaluating the performance of hydraulic structures. Shortings in models usually are termed scale effects of laboratory effects. Scale effects increase in severity as the ratio of prototype to model size increases or the number of physical processes to be replicated simultaneously increases. Laboratory effects arise because of limitations in space, model constructability, instrumentation, or measurement. Generally, steady nonuniform flow characteristics in open channel flow with hydraulic structures can be explained as a following relationship (ASCE, 2020). 9 where Sw is water surface slope, So is channel bottom slope, h is water depth, k is roughness height of solid boundary, V is flow velocity, g is gravitational acceleration, and vare dynamic iscosity, density, surface tension of water, respectively. Eq. (1) states that water surface profile is expressed as bottom slope, relative roughness height, Froude number, Reynolds number and Weber number. Similarity of variables in Eq. (1) between scaled model and prototype is maintained for the hydraulic model to properly replicate features of a plicated prototype flow situation. Generally, geometric similarity (So) is achieved and experiments are carried out by using Froude number similarity in the hydraulic Table 1. Approximate Values of Roughness Height, k model on the open channel flow and hydraulic structures. Water is used to analyze the flow characteristics of scaled model, thus modeling accuracy is promised because the properties of water are not scaled. So, a small scale model may causes a failure to simulate the forces attendant to fluid properties such as viscosity and surface tension, to exhibit different flow behavior than that of a prototype. Moreover, relative roughness height of the scaled model cannot be exactly reproduced because materials of experiment are limited. Previous study on the scale limits of hydraulic models leads to some guidelines. The Bureau of Reclamation (1980) used length scale ratios of Lr = 30~100 for models of spillways on large dams. And model flow depths over a spillway crest should be at least 75 mm for the spillway’s design operating range. The average roughness height for a given surface can be determined by experiments. Table 1 gives values of roughness height for several kinds of material which are used for construction of hydraulics structures and scaled models (Hager, 1999). To