Determine of velocity field with PIV and CFD during the flow around of bridge

The article describes the processing of specific junior research FAST-J-11-51/1456 which dealt with physical and CFD of the velocity field during the flow around of bridge piers. Physical modelling has been carried out in Laboratory of water management research in Institute of Water Structures in Brno University of Technology Faculty of Civil Engineering. To measure of the velocity field in profile of bridge piers were used laser measuring method PIV (Particle Image Velocimetry). The results of PIV served as a basis for comparing experimental data with CFD results of this type of flow in the commercial software ANSYS CFX.


Introduction to the issue
a If we place in the path of the flowing fluid barrier, the fluid begins to wrap her.The most famous examples of real objects include wrapping, such as water in a river flows around the piers or flowing air flows around objects and bodies on the earth's surface (air wrapping around static objects and cars or aircraft).The body wrap also occurs when the fluid at rest and the body moves in the space -as a ship floating in the sea.When you wrap any objects created water flow tangential stress strain their surface, which gradually degrades the material from which they are formed.Wrap may be either pressure or vacuum.During this hydraulic phenomenon leads to boundary layer separation, which is a thin layer along flowarounding object, and the wakes of different shape in dependence on the flow velocity (a type of flow -laminar and turbulent) and the shape of the body itself.In the boundary layer creates a strong deceleration current, which in some cases leads to a discontinuity of the current and the formation of vortices.Flow separation and wake are always associated with a large loss of energy and intense pulsation speeds and pressures, and therefore we try to remove it and most appropriate treatment shaped structures [1].a mistrova.i@fce.vutbr.czb daniel.picka@seznam.cz

Objectives of the project
The aim of the project was measured by PIV velocity field for three selected shapes gridiron pillars and three angles of deflection of the pillars of the longitudinal axis of flow.The pillars are arranged in the profile trough so that they are always at least three pillars located in the central part of the profile above the glass tray at the bottom of the hole (Fig. 1).

Fig. 1. 3D model placement pillars in luminous efficacy trough
The project was divided into two parts, and physical and numerical modeling.The findings and outputs obtained from physical modeling served to show basic characteristics of the flow along the pillars, the pillars of the influence of the shape and influence the direction of deflection of the pillars on the flow in the space behind the pillars.The main objective of this research project was to obtain enough quality data for the initial calibration and subsequent verification of CFD model.

Hydraulic trough -measuring track
Model bridge piers was installed in a hydraulic flume width of 1 m, length 12 m and maximum achievable flow 160 l•s -1 .Construction is reinforced concrete trough with vertical glass walls.Inflow of water into the trough was channeled through perforated plate with circular holes that measured in specific area to ensure, where possible, the most consistent flow and created a stable flow profile.Drain water from the trough was controlled using segmented plate cap fitted at the end of the trough.

Models of piers
Overall, it was necessary to measure the velocity fields for three shapes gridiron pillars (Fig. 2), the three angles of deflection of the pillars of the longitudinal axis and the filling stream: 0°, 30° and 45°, and for three flow rates: minimum (small) -from 1.3 to 1.6 l•s -1 , medium (middle) -from 24.0 to 24.5 l•s -1 and the largest (large) -from 40 to 43 l•s -1 .Flow rate was dependent on the water level in the settling trough levels in the range of 245 to 255 mm from the bottom of the trough -see table 1 and table 2  For fixing pillars into the flowing water in the trough was made structure consisting of stainless steel with welded hollow metal around the edges, increasing its rigidity.The shorter side of plate were also welded to the vertical hollow with holes for mounting screws because of the possibility strutting beam between gutter walls (Fig. 3).

Measuring technique used to measure the velocity field -measuring kit for PIV
The main part of this kit includes a double pulse laser type Nd: YAG, manufacturer New Wave Research Gemini PIV type with adjustable repetition rate of 0-15 Hz and a maximum energy of 120 mJ in a flash and camera FlowSense 2M, the Danish manufacturer Dantec resolution 1600 × 1200 pixels.To synchronize all elements of the set in the measurement was used Central System Controller Hub also made by Dantec.In the selected horizontal cut channel set was recorded 100 instantaneous velocity field states with a sampling frequency of 1 Hz in the first set and a frequency of 5 Hz in the second set of measurements.Estimating the value of the expanded uncertainty water velocity in the measured field, with a confidence level of 95% is less than 10%.PIV method is not intended to measure the point velocity, but to detect the spatio-temporal context.In Figure 4 it is possible to see the overall view of the measurement technique for measurement by PIV spaced around the track density in the laboratory.Before measurement of velocity field by PIV were carefully considered ways to illuminate space when wrapping the pillars of the flowing water and shooting options illuminated field camera to the measurement results as much as possible to account for the fact that the flow in the selected area.It was decided that the area of interest will be recorded in a single horizontal cut (incision parallel with the bottom trough) and at a distance of about 125 mm from the bottom of the trough.Illuminate the area of interest was carried out through a mirror mounted in a special sealed aquarium which could be placed directly into the gutter flooded area (downstream for a number of pillars).From the top of the aquarium open to him to be able to insert the head Laser source (Fig. 5).Scanner for laser illuminated area of interest (horizontal cut) was deposited on the floor of the laboratory under the bottom of the trough.Image capture area of interest, therefore, takes place through a mirror set sloping (ie the inclination angle of 45° from the bottom of the trough) under the glass hole in the bottom of the trough.In Figure 7 we can see the size and scope of the rectangular gridiron wake pillars, pillars of diversion from the longitudinal flow direction by 30 ° and the flow is 42.9 l.s -1 .Here we see, compared to the vector field in Figure 6a, much clearer in the wake flow back along the side wall pillars and the area downstream of it.The maximum speed measured in the direction of flow (flow at 42.9 l.s -1 ) was 0.45 m.s -1 and the velocity of reverse flow in the wake 0.19 m.s -1 .For comparison, at the flow rate 1.3 l.s -1 , the value of the speed of 0.018 m.s -1 and 0.008 m.s -1 .Computational domain of size 1.0 × 1.91 m was divided into three sub required by grid computing density (Fig. 12).The area around the pillars were covered densely.Discretization of the computational domain was using tetrahedra, whose size ranged from 2.0 mm to 7.0 mm.Total computational grid contains 383,000 nodes and 01065-p.5 1,141 million cells.The field was applied boundary conditions: wall (wall -no slip) at the bottom of the pillars and the longitudinal walls of the trough.At the inlet was administered condition known velocity obtained from physical measurements using PIV.The outlet was applied hydrostatic pressure distribution corresponding to the discharge space.Reference pressure of 0.1 MPa was assumed [4].

The results of numerical modeling
Outcomes of CFD (streamlines from numerical module) and physical measurements of velocity field in the wrap pillars are similar (see Figure 6b to 11b).Other outputs from numerical modeling can be seen in Figure 13 to 15. CFD method used proved to be satisfactory.In the future, it would be a good idea to wrap the pillars numerically modeled in 3D space.This method is particularly challenging for several computing power and time [4]. .

Fig. 2 .
Fig. 2. Scheme of the three pillars of shapes gridiron

Fig. 3 .
Fig. 3. View of the spacer structure and method of fixing the glass wall trough

Fig. 5 .
Fig. 5. View of the aquarium with a mirror and the general disposition of the model in the trough5 Results of physical measurements using PIVAfter getting the huge amount of data during the measurement in the laboratory, it was decided to evaluate the collected data.Two-images were further processed using the software FlowManager v4.71.Of vector fields, which are the output PIV measurement methods laid down in a horizontal plane (see Chapter 3) can best track the location of flow separation of water from the wall surface of the reference object of flowaronding pillar.Fig.6ashows the time-mean values of image projections of the velocity vector in the horizontal plane with a rectangular gridiron pillars, pillars of diversion from the longitudinal flow direction by 30° and the flow is 1.3 l•s - 1 .We can see the range wake in the space around the pillars and the stream behind the pillar.At such low flow rates is a reverse flow is very small.

Table 1 .
Overview of the measured states -flow rates, type of gridiron and depth of water in the trough -recorded during the measurement Flow [l.s -1 ] Water level [mm]Rectangle 45

Table 2 .
Load cases -the values of Re, the corresponding velocity and flow rate -calculated before physical modeling