Improvement of a System for Catchment, Pretreatment, and Treatment of Runoff Water Using PIV Tests and Numerical Simulation

Copyright © and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders.


INTRODUCTION
The pollution of runoff water is an important problem which seriously affects the environment.
The poor quality of storm water which is filtered by the terrain causes contamination known as "diffusive pollution".Several solutions for this problem have been studied since the nineteenth century (Caltrans 2010; Campbell et al.2004;Castro Fresno et al. 2005;Dolz and Gómez 1994;Novotny 2003).In these previous studies, BMPs (Best Management Practices) were developed and the efficiency of the new devices was discussed.As a conclusion of these projects, it was proved that the systems used are not enough to improve the water quality.In this sense, Fernández Barrera in his Doctoral Thesis in 2009 and other authors developed a new System of Catchment, Pre-treatment and Treatment (SCPT) of the pollutants of the runoff water.This new system was patented (Castro Fresno et al. 2010).
The SCPT consists of two sections divided by a flat panel, which is also called the screen.The function of the first of these sections is as a hydraulic plug where it is possible to retain oils.
Then the water flows under the flat panel and goes inside the second section which works as a gravity settler.In this second section, geotextil layers can be included in order to filter the water.This system allows the reduction of the environmental impact due to the contaminants of the runoff water, one of the most important problems of the storm water.
Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro-Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE) IR.1943-4774.0000749Summarizing, the SCPT is an important contribution to environmental engineering because it is able to treat storm water and clean it before it filters into the terrain.For this reason, an in depth study of the system's behaviour has been developed in this paper in order to increase its efficiency.The methodology used to complete the SCPT study is based on a combination of the experimental and numerical studies.This technique is known as Hibrid Engineering (HE) and has been used in many other fields too (Flamant et al. 2004;Mayurkumar et al. 2012).
In this work, the influence of the flat panel of the SCPT on its efficiency has been studied.A 1:4 scale structure of the SCPT, made of methacrylate, has been studied.This structure is 0.2m wide, 0.35m.long and 0.

Techniques used
Two different techniques have been used in this study.On the one hand, the PIV technique has been used to measure the speed of the fluid inside the SCPT scale model.Several parameters can be obtained using this technique which has been studied by different authors in several fields (Melling 1997;Raffel et al. 1998;Schroeder et al. 2008;Wang et al. 2009).On the other hand, a numerical method based on FVM has been used to analyze the behaviour of the system.Furthermore, DOE has been used to optimize the position of the screen and improve the efficiency of the SCPT.The combination of FVM and DOE has been used for years (Zienkiewicz et al. 2005;Madenci and Guven 2007;del Coz et al. 2007).

EXPERIMENTAL TESTS
The PIV technique was used in order to study the performance of a small scale SCPT.This method has allowed the measurement of the fluid velocity inside the SCPT.
The equipment used in the laboratory tests was a laser PIV system "H41" ( Etalon Research 2009).The PIV technique has been used for measuring the fluid speed at multiple points throughout a 2D plane.The fluid is seeded with particles which follow the fluid movement inside a volume.It is possible to deduce the movement of the underlying flow by tracking the Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro-Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE)IR.1943-4774.0000749motion of these particles.This technique is well explained and studied by other authors (Wereley and Gui 2001;Van Hooff et al. 2012).
A laser illuminates the SCPT, and the camera, which is normal to the illuminated plane, captures the particle velocity in that section.An electronic control allows pulse separation control between 100 µs and 5 s, and width pulse control between 10 µs and 32 ms.The software processes the velocity map by cross-correlation using capture imaged in each laser pulse.. Furthermore, the equipment has an optical standard system which is able to make a light plane which is 3x200x250 mm (width x height x length).This optical system can be used with two angular options, 22º or 45º; in this study only 45º opening has been used.The resolution of the camera is 640 x 480 with digital output format of 8-bit and the pixel size is 6 x 6 µm (8.5•10 -2 m format).
The frequency of this camera is 16 Hz, with a standard photographic lens of 1.25•10 -2 m, 1:1.4,and is able to register distances between 0.3 and 2 m.
The images captured are processed by "rtControl Software", which was provided by the equipment manufacturer.This software takes the captured data, divides the images in subwindows and then analyzes consecutive images by cross-correlation.The time between subwindows is known and so, the velocity map can be obtained.It is important to take into account the influence of external factors which can reduce the quality of the images, such as the kind and size of the seed particles, the reflective surfaces, etc.In this sense, due to the complexity of the PIV technique the experience of the authors is very important to obtain success results.
This software can also provide other parameters like the velocity vector (in real time) of the particles in a fluid, and statistical data like vorticity, average speed, etc.
Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro-Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE) IR.1943-4774.0000749Set up of the laboratory tests The configuration of the PIV tests has taken into consideration the following aspects:  The seed particles, as suggests the PIV equipment manufacturer, are polyamide particles of 100 µm, whose density is the same as water density.The polyamide particles have suitable optical properties in order to be illuminated by the laser.
 The superposition of images is not possible.So, the captured plane by the camera must be smaller than the illuminated plane by the laser.In this case, the maximum plane size that the camera can capture is 0.5 m. long.In this case, a 1:4 scale SCPT has been used, whose 2D plane captured is 0.35 m long.
 The scale model inlet is a longitudinal groove in a cylindrical pipe whose volume of water is about 0.1 l/s in order to obtain a non turbulent inflow.The inlet is immersed to avoid the turbulence and obtain a regular flow inside the system.Besides, the seed particles and the water have been mixed before the inlet of the SCPT to make it easier to follow the water movement.
 The equipment calibration must be done before starting the measures by PIV.This calibration assigns units to the results of the software.It is important to know that only the results which have been taken during the same calibration can be compared.
 Several positions of the flat panel have been studied to obtain experimentally the influence of the screen displacement over the SCPT efficiency, see Figure 1.
The laboratory test assembly is shown in Figure 1,where the laser illuminating the SCPT model is shown on the left side.Also, the camera normal to the illuminated plane and the software used are shown on the right side.
Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro-  In conclusion, the main result obtained in the experimental test was the average vorticity value, W. This result was calculated as the scalar component of the angular velocity vector or punctual vorticity,  , for a two-dimensional input, in this case H and V, which corresponds to the plane illuminated by the laser.The following equation (3) was used (Nezu and Nakagawa 1993): Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro- Where: u and v are the velocity components with respect to x and y directions.
m and n are the grid in x and y directions, respectively.
The scalar punctual vorticity,  , was calculated at each grid point, m n  , of a fine mesh of 3.5 mm size that fill the capture of the illuminated plane.In order to avoid local outlet suction, which may change the flow path, the experimental data were selected.In this way, we have removed 50 mm at the end of the output chamber, corresponding to 12 to 14 horizontal grid points approximately.

NUMERICAL SIMULATIONS
The numerical simulation presented in this paper has been carried out by the CFX module in the ANSYS Workbench software v.12.1.This work has been supported by the authors' experience in the field of simulation and non-linear analysis.Over several years, the authors have worked using finite element methods (FEM) and finite volume method (FVM), (del Coz et al. 2007;del Coz et al. 2011;García Nieto et al. 2010).These methods provide great results when the models are correctly compared and validated.
The relationship between the horizontal position of the flat panel and the particle collection efficiency has been predicted in In this sense, the input parameters previously used have been studied in both experimental and numerical ways.In summary, the contribution of this paper is important in order to prove the ability of the PIV and numerical methods to verify the SCPT performance.

Finite volume model
In this sub-section the finite volume model used in this investigation is explained: This finite volume model has been solved by an iterative process to obtain the solution of the problem.

Numerical simulation based on DOE
The influence of the screen displacement on the SCPT performance has been studied in order to improve the efficiency of this system.In this sense, an optimization procedure based on the Design of Experiments technique (DOE) on the FVM models has been carried out in order to know the best position of this screen from the efficiency point of view (del Coz et al. 2013;Montgomery 2001).
The input parameters of the DOE technique have been the horizontal and vertical positions of the flat panel.The output parameter in the DOE has been a derived parameter, which indicates the performance of the system.The efficiency of the system has been obtained from the reduction of the average vorticity inside the SCPT.If the vorticity of fluid flow in the input chamber is high, the majority of the solid particles will stay inside the input chamber (Penko and Calantoni 2013).The few particles that could go to the output chamber will precipitate at the bottom of the SCPT due to this chamber working as a gravity settler.A large difference in the average vorticity between the input and the output chamber means that the SCPT performance is good.
In this case, the punctual vorticity,  , was calculated by means of the equation (3) at each finite volume taking into account the meshing parameter between 2•10 -3 and 6•10 -3 m.Next, the average vorticity, W, was obtained as the mean value of the punctual vorticity for all finite Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro-Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".The reduction of the vorticity is defined by equation ( 4).
 

Laboratory test results
The results of the laboratory tests provide information about the fluid trajectory inside the SCPT and the velocity map of the fluid.
The PIV equipment gives results in several formats.In this case, MATLAB ® files and .avifiles were evaluated.The.avi files provide qualitative results and the Matlab files provide quantitative results.
In Figure 2, the results from the laboratory tests of the scale prototype are shown.The fluid inlet is on the right and the outlet on the left.Different colours are used in these representations to show the different velocities, from blue for lower velocity to red for higher velocity.As it is shown in Figure 2, the fluid has two vortices, one in the input chamber and other in the output chamber.
Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro- To obtain quantitative results, the Matlab files were analyzed.A velocity matrix is taken every 0.1 s.After several seconds, all matrixes are combined by software in order to represent the velocity map of the fluid from experiments.These results are analyzed by a matlab code and incoherent vectors are deleted.Then, the punctual vorticity parameter,  , is obtained in each matrix, along with the average vorticity, W, for each test.The velocity map obtained in the experiments is plotted in Figure 2.
The vorticity or rotor vector of velocity is the same concept as the angular velocity of a fluid.
The vorticity was considered as the best parameter to represent the rotational behaviour of the fluid in the SCPT.The average vorticity value in the input chamber and in the output chamber is different (see Table 2).Improved performance in the SCPT is given by a considerable vorticity decrease between both chambers.If the average vorticity is high, as in the input chamber, the solid particles are rotating inside the fluid.The few particles that go to the output chamber precipitate because the average vorticity in this volume is low.The fluid that goes out of the system is clean and without solid particles, and in this way, the runoff water that is filtered to the terrain is less contaminated.In order to compare the laboratory test and the simulation results the average vorticity evolution is obtained.Desired performance of the SCPT indicates a negative trend which means that the vorticity between the input and the output chamber is decreasing.

Numerical results
On the one hand, the results obtained in the numerical simulations by FVM are shown in Figure 3.
25m. high.The flat panel has been studied in several positions, moving it in vertical and horizontal directions.The scale prototype and the different positions studied are showed in Figure 1.Experimental scale tests have been developed with this prototype using Particle Image Velocimetry (PIV).Furthermore, numerical simulations have been carried out using the Finite Volume Method (FVM) and the best flat panel configuration has been selected by means of an optimization using the Design of Experiments technique (DOE).Finally, experimental scale tests have been developed using the Particle Image Velocimetry (PIV) technique.Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro-Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE)IR.1943-4774.0000749 Figure 1.SCPT scale prototype studied: (a) dimensions and panel positions; (b) general assembly of the laboratory tests.The aim of this research paper is to complete the previous investigations about the SCPT, considering three advanced techniques: numerical simulation by FVM, optimization using DOE and experimental measures by PIV.
have been carried out, one for each flat panel position inside the scale model of the SCPT.For horizontal displacements the distances between the inlet wall and the panel have been 120, 145 and 170 mm, wich are H1, H2 and H3 respectively.For vertical movements the distances between the end of the panel and the bottom of the SCPT have been 70, 90 and 110 mm, wich are V1, V2 and V3 respectively.The first horizontal position, called H1, cannot be tested due to its proximity to the inlet fluid.Taking into account previous research works focused on the flow simulation and sediment transport over channels(Penko and Calantoni 2013), as well as previous full-scale experiments over the SCPT(Fernandez Barrera 2009, Fernandez Barrera et al. 2010, del Coz et al. 2011)     the vorticity was considered an important result.The rotational fluid movement inside the device and the conclussions derived from Penko and Calantoni provide the importance of the vorticity values in this specific case.


Geometrical model: the geometrical model which has been used in the SCPT numerical simulations reproduces the fluid volume inside the system.The dimensions of the geometrical model used in the numerical simulation and the experimental SCPT are the same.Finite volume types: the meshing of the geometrical model by finite volumes has been carried out using the "Automatic Method" in ANSYS.This method uses a patch conforming algorithm for tetrahedrons method control by a Delaunay tetra mesher.If possible, the mesher builds tetrahedrons volumes and tries to create a smooth size variation based on the specified growth factor.In this case, the volume size has been established between 2•10 -3 and 6•10 -3 m.The result has been a mixed mesh with mainly tetrahedral and hexahedral volumes, although others types of finite volumes have been included in the interface areas.The geometrical model consists of 389,265 elements and 110,066 nodes.The majority of the element quality values are about 0.75, what means a good mesh quality (Zienkiewicz et al. 2005; Madenci and Guven 2007).Boundary conditions and loads: the fluid flows by the input chamber, then it goes through the output chamber and finally it goes out the system by the outlet.The results of the numerical simulations have been considerably influenced by the configuration of Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro-Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE)IR.1943-4774.0000749   the transition area.The interface areas have been located between the inlet and input chamber, between the input chamber and output chamber and between the output chamber and outlet.
Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE)IR.1943-4774.0000749volumes in each camera.Furthermore, the dimensions of the input and output chambers are changed according to the position of the flat planel as has been studied in the DOE analysis.The parameters used in the analysis by DOE are the following:  Input parameters: o Horizontal position: from 70 to 170 mm, with an initial value of 120 mm.o Vertical position: from 70 to 120 mm, with an initial value of 95 mm  Output parameters: o Average vorticity in the input chamber o Average vorticity in the output chamber o Vorticity reduction in the SCPT The technique which has been used in DOE is Central Composite Design (CCD).The limited values used in the DOE technique are determined by the SCPT dimensions.The maximum and the minimum positions of the flat panel are taken as boundary values.The initial conditions are intermediate values.The total CPU time for each numerical model has been 7.955•10 3 seconds (2 hours, 12 minutes and 35.281 seconds).
Vorticity reduction inside the SCPT ic W : Average vorticity in the input chamber [rad/s] Author's post-print: M. Alonso-Martínez, A. Navarro-Manso, D. Castro-Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE)IR.1943-4774.0000749oc W : Average vorticity in the output chamber [rad/s] The vorticity reduction represents the efficiency of the system and indicates a decrease of the fluid rate inside it.So, the best efficiency is obtained with maximum vorticity reduction.The best position of the flat panel to obtain a high level of efficiency is shown in the response surfaces obtained by DOE.

Figure 3 .
Figure 3. Numerical simulation results: velocity vectors map in the SCPT for a flat panel position of H = 120 mm. and V = 90 mm.

Table 1
Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE)IR.1943-4774.0000749Withrespect to the PIV equipment configuration, the parameters of the camera and the laser must be carefully chosen to obtain accurate captures of the fluid movement.These parameters must be adapted to the fluid velocities inside the scale SCPT.The maximum velocity of the fluid in the SCPT is at the inlet.This value has been calculated based on the maximum inlet volume and the minimum inlet surface.The maximum velocity obtained was below 0.2 m/s.The calibration in this work has been 3.5 mm/pixel.The fastest particles, therefore, cannot travel more than ¼ of the sub-window size between images.So, the values of the pulse separation, Δt, and the pulse width, δt, have been estimated accordingly In addition to the pulse separation and pulse width there are other parameters of configurations in the PIV camera that depend on the boundary conditions of each test.The values of these additional parameters are provided in Table1.Values of the configuration parameters for the PIV camera.M. Alonso-Martínez, A. Navarro-Manso, D. Castro-Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE)IR.1943-4774.0000749 Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".
del Coz Diaz et al. 2011.However, in thispaper a new numerical model has been developed in order to study the behaviour of the SCPT from the Author's post-print: Fresno, F.P. Álvarez-Rabanal, J. J. del Coz Díaz."Improvement of a system for catchment, pre-treatment and treatment of runoff water using PIV tests and numerical simulation".Journal of Irrigation and Drainage Engineering.Volume 140, Issue 8 (August 2014) 10.1061/(ASCE)IR.1943-4774.0000749