5A.1 VISUALIZATION OF TORNADOES USING FUNNEL AND DEBRIS CLOUDS Minoru Noda*, Minoru Hachiya, Fumiaki Nagao Tokushima University, Tokushima, Japan 1. INTRODUCTION horizontally rotating flow were required to produce a tornado-like flow. In this study, a simple rectangular To estimate the tornadic damage to structures and numerical grid shown in Figure 1 was used. To human beings, it is very important to investigate the generate a horizontally rotating flow, a horizontal shear characteristics of tornadoes. The main method to rate flow was produced by a pair of data points, the wind the intensity of tornadoes is the use of the Enhanced speeds at the inlet and outlet1, with fixed uniform Fujita Scale, which is very reasonable if the degree of values u0 and u1, respectively, arranged as shown in damage is clearly indicated (MacDonald, 2004). The Figure 1. The convergence flow was generated by the Japanese Enhanced Fujita Scale was developed in upwind to outlet2, placed at the center of the ceiling of April 2016 and has been used in Japan to estimate the numerical region, and controlled by the difference wind speeds from damage ever since (Okuda, 2015). in wind speed between inlet and outlet1. On the other This scale includes some damage indices not available hand, dimensions were given as real scales, because in the American Enhanced Fujita Scale, such as the conditions that control the scale of the tornadic automatic vending machines, power transmission line vortex were unclear. This investigation was carried out poles, and asphalt of roads (Noda, 2013). However, under simple configurations, assuming incompressible estimating the strength and scale of tornadoes is flow and no heat transmission, because a tornado-like difficult if the amount of damage available for vortex can be produced experimentally. evaluation is not sufficient. Other methods that depend These calculations were carried out by OpenFOAM on the extent of damage involve direct observation, based on the finite volume method. Table 1 shows the such as the VORTEX2 Project (Wurman, 1995). This is parameters for this calculation. a very suitable method but it is difficult to use in Japan z because of the challenge in moving observational equipment quickly and safely. Therefore, other methods are necessary to estimate characteristics of tornadoes quickly and safely. d x One idea is to use images and movies uploaded to y outlet2 inlet u1 the Internet by the general public. The relationship d u outlet1 between the maximum wind speed of tornadoes and 1 u H the shape of funnel clouds was investigated, using a 0 h two-dimensional Rankine vortex (Dergarabedian, outlet1 1970). However, real tornadic vortices have three- inlet u0 dimensional structures. Therefore, the relationship W/2 between the characteristics of tornadoes and their W funnel clouds should be investigated considering three- W/2 dimensional tornadic flows. In this study, three-dimensional real-scale tornado- Figure 1 Numerical region and Boundary like vortices were produced using the Large Eddy Simulation in OpenFOAM. Moreover, funnel clouds for Table 1 Parameters for this calculation these flows were calculated assuming that only the Inlet wind speed, u0 5 m/s temperature change is adiabatic. In this manner, the Outlet 1 wind speed u1 4 m/s relationship between the characteristics of flows and Height of Numerical Region, H 3 km shapes of their funnel clouds was investigated. Width of Numerical Region, W 12 km Furthermore, debris clouds for these flows were Width of Upwind hole, d 1.2 km (d/W = 0.1) obtained through the three-dimensional motion Height of Inlet, h 3 km (h/H = 1.0) simulation of debris. 2. SIMULATION OF TORNADO-LIKE FLOW BY 2.2 Instant Flow Field LES Figure 2 shows the ensemble-averaged components 2.1 Calculation Configuration of velocity in cylindrical coordinates—tangential, radius, and vertical directions—and pressure at the same distance from the vortex center of the instant flow field, It is well known that a convergence flow and a z = 20 m, t = 30000 s. In this study, to quantitatively * Corresponding author address: Minoru Noda, Tokushima evaluate the strength and scale of the vortex, the core Univ., Graduate School of Science and Technology, radius and the circulation were obtained by fitting Tokushima, 770-8506, Japan; e-mail: [email protected] the Burgers-Rott vortex (Burgers, 1948) to this u.ac.jp tangential wind speed distribution, as given by the following formula: 2.4 Vertical Distribution of the Time-Averaged Flow Field r 2 U 1 exp (1) Figure 4 indicates vertical distributions of vortex t 2 2r parameters averaged over a span of 1000 s. The green lines and red lines indicate the average value and width of the standard deviation, respectively. 1 0.1 Burgers vortex 0.05 The figure indicates that the core radius and 0.8 0 -0.05 circulation increase with altitude, but the maximum -0.1 0.6 -0.15 tangential wind speed remains nearly constant near the -0.2 0.4 -0.25 ut/Utmax ur/Utmax ground. This flow field was used in the visualization of -0.3 0.2 -0.35 tornado by funnel cloud, which is discussed below. -0.4 0 -0.45 0 5 10 15 20 0 20 40 60 80 100 120 r/ε r/ε (a) Ut (b) Ur 0.9 0.5 0.8 ) 0 0.7 2 0.6 -0.5 0.5 0.4 -1 (Utmax) 0.3 ρ uz/Utmax -1.5 0.2 0.1 P/(1/2 -2 0 -0.1 -2.5 0 5 10 15 20 0 5 10 15 20 r/ε r/ε (c) Uz (d) p Figure 2 Distributions of the wind velocity components and the pressure along the radius direction at z = 20 m Figure 4 Vertical distributions of vortex parameters 2.3 Unsteadiness of Flow Field averaged over dt = 1000 s 3. VISUALIZATION OF TORNADO-LIKE FLOW BY Figure 3 indicates the time histories of the core FUNNEL CLOUD radius, circulation, and maximum tangential velocity given by the Burgers-Rott vortex distribution at the z = 20 m horizontal plane. The flow is unsteady, so the 3.1 Definition of Funnel Cloud averaged values of the vortex parameters over dt = 1000 s were calculated at each height. To investigate the relationship between the shapes of funnel clouds and the flow fields in this study, the funnel cloud was defined as the saturated water vapor surface. Assuming adiabatic change and incompressible flow, Teten’s formula gives the saturated water vapor surface, as follows: 7.5T0 7.5T a T 273.15 s0 10 T0 237.3 T 237.3 RH 1 (2) as T0 273.15 where as0 is a water vapor amount at the reference point. as is the saturated water vapor amount at the surface. T is the temperature at the surface and T0 and RH are the temperature and the relative humidity at the reference point, respectively. These formulas indicate that the funnel cloud is controlled by the temperature and relative humidity. Moreover, the temperature is controlled by pressure field as described by the following formula: Pf P0 T T0 d (z z0 ) (3) C p where Pf is the pressure for tornado-like flow, P0 is the reference pressure, Cp is the specific heat at constant pressure, and d is the dry adiabatic lapse rate. Therefore, as demonstrated, the funnel cloud is controlled by the pressure field and the relative humidity. Figure 3 Time history of vortex parameters at z = 20 m To construct the pressure field ptarget, for the target Utmax 1000 maximum tangential wind speed Utmax,target, the 800 600 Utmax=20 pressure of the tornado-like flow was modified to obtain 40 z(m) 400 60 the pressure field for an arbitrary maximum tangential 80 200 105 130 wind speed, using the following formula: 0 -3000 -2000 -1000 0 1000 2000 3000 2 y(m) U t max,target Figure 7 Effect of maximum tangential wind speed on p p (4) target original the shape of funnel cloud U t max,orginal where Utmax,orginal and poriginal are the maximum 4. VISUALIZATION OF TORNADO-LIKE FLOW BY tangential velocity and the pressure of the original flow DEBRIS field, respectively, produced by the calculation shown in the previous section. 4.1 Equation of 3DOF motion of Debris Figure 5 shows an example of a funnel cloud produced under the condition of Utmax = 105 m/s and RH = 60 % by this method. In this study, it was assumed that the debris motion occurred only due to the drag forces. Therefore, the three-degrees-of-freedom equation of motion was used as shown below: 1 mx U(x) x U(x) x C D A mg (5) 2 where m is the mass of debris, x is the vector of spatial coordinates of debris, is the air density, U(x) is the wind speed vector at x, CD is the drag coefficient, A represents the area of debris, and g is the gravity acceleration vector. By normalizing this equation with the representative wind speed and gravity acceleration, the non-dimensional motion can be written as follows: x* C DTa U* (x* ) x * U* (x* ) x * g* (5) where Ta is the Tachikawa number, defined as the ratio Figure 5 An example of funnel cloud produced by the of the aerodynamic force to the gravity force method described in section 3.1 (Utmax = 105 m/s, RH (Tachikawa, 1983). The index * indicates non- = 60 %) dimensional values. This equation indicates that non- dimensional motion is controlled by two parameters, 3.2 Effects of Relative Humidity the drag force coefficient, and the Tachikawa number.