CFD simulation of pathogen intrusion during hydraulic transients in water distribution networks

International Environmental Modelling and Software Society (iEMSs) 7th Intl Congress on Env Modelling and Software San Diego, CA, USA Daniel P Ames, Nigel W T Quinn and Andrea A Rizzoli (Eds ) http www iemss orgsocietyindex phpiemss 2014 proceedings CFD Simulation of pathogen intrusion during hydraulic transients in water distribution networks Mora Rodríguez, Jesús1; Delgado Galván, Xitlali1; Ortiz Medel, Josefina1; Ramos, Helena M 2; Fuertes Miquel, Vicente3; López Jiménez, P Amparo3 1 Dep.

Daniel P Ames, Nigel W.T Quinn and Andrea A Rizzoli (Eds.) http://www.iemss.org/society/index.php/iemss-2014-proceedings

CFD Simulation of pathogen intrusion during hydraulic transients in water distribution networks

Mora-Rodríguez, Jesús 1 ; Delgado-Galván, Xitlali 1 ; Ortiz-Medel, Josefina 1 ; Ramos, Helena M 2 ; Fuertes Miquel, Vicente 3 ; López-Jiménez, P Amparo 3

1 Departamento de Ingeniería Geomática e Hidráulica División de Ingenierías de la Universidad de Guanajuato Av Juárez No 77, Centro, 36000, Guanajuato, Mexico jesusmora@ugto.mx ,

2 Civil Engineering Department and CEHIDRO, Instituto Superior Técnico, Technical University of

Lisbon, Av Rovisco Pais, 1049-001, Lisbon, Portugal hr@civil.ist.utl.pt ,

3 Hydraulic and Environmental Engineering Department Universitat Politècnica de València Camino

de Vera, s/n 46022, Valencia, Spain vfuertes@upv.es , palopez@upv.es

Abstract: This work shows a CFD (Computational Fluid Dynamics) model for simulation of intrusion

flow in water distribution pipes based on experimental and theoretical models The intrusion flow in water distribution networks is known as pathogen intrusion, and it affects the water quality within the network Their consequences are potential epidemic diseases due to the consumption of polluted water In the experimental models the intrusion flow occurring during this phenomenon though failures was measured in pipes in steady and transient state The first experimental model simulates the intrusion flows during a water hammer, and the second one, simulates the steady-state intrusion flow with external porous media in order to simulate conditions of the intrusion phenomenon on buried water distribution pipes Finally, diverse water samples were taken around water mains to identify the pathogens and its concentration The experimental results were the base to validate two numerical models, a vapour–cavity MOC model, and a CFD model, both to simulate the transient event and the soil around the pipe to determine the intrusion volume and the concentration of contaminants that can enter in these specific events in water distribution networks

Keywords: pathogens in water, CFD, intrusion, water distribution networks, physical models, pipe

failures, vapour-cavity MOC model

1 BACKGROUND

In 1996, the International Community through the Organization for Economic Co-operation and Development (OECD) recognized the need for a greater understanding of the relationship between

drinking water and the transmission of epidemic diseases (Hunter et al., 2002) Since the 1990

decades, the pathogen intrusion is a phenomenon that has been studied with special focus, essentially in countries like United States and Canada, where it has been confirmed that one of the potential sources of epidemic diseases is related to the consumption of water contaminated (Kirmeyer

et al 2001, Karim et al., 2003; LeChevallier et al., 2003; Friedman et al., 2004) The risk of water pollution is related to several factors, Kirmeyer et al classified pathogen entry routes in 2001 focusing

on the level of risk resulting from the intrusion The meant routes with high risk were water treatment breakthrough, transitory contamination, cross connection, and water main repair/break Pathogen intrusion occurs in some cases when negative pressure conditions are achieved in the systems (Figure 1), allowing the entrance of external water around a leak and diminish the water quality

In the last years, the epidemic diseases related to the consumption of water contaminated have been increased in cases where the source of contamination is in the distribution network Furthermore, the cases related to the pollution where the source of contamination was in the water treatment have decreased The studies carried out in relation to the pathogen intrusion presented an approach from the point of view of water quality in which makes a microbiological risk assessment along the distribution of drinking water (Pedley et al., 2004; Zloczower et al., 2009)

Figure 1 Scheme to generate pathogen intrusion

This paper shows an analysis to obtain a quantification of pathogen intrusion during transient events considering the porous media around the pipe and the concentration of the superficial sources of polluted water near to the water distribution installations These three elements were taken to simulate numerical models to characterize the pathogen intrusion event based in two methods: Method of Characteristics (MOC) and Computational Fluid Dynamics (CFD)

2 PHYSICAL MODELS AND PATHOGENS DATA

There were used two different models to obtain the intrusion flow The first one is the model to obtain the flow in a steady-state considering the soil around the pipe and the second one is the model to obtain the flow in a transient event in a pipe In this way, it have been considered the factors to characterize the event of intrusion that can occur during transient pressure taking into account the soil around the pipe

2.1 Model scenario of intrusion flow with external soil

In the first model, a circular orifice of 1mm diameter simulated the failure of the pipe The pathogen intrusion was generated by a negative pressure on the pipe Polluted water is represented by a source of constant level of water over the pipe These conditions represent a typical leakage that is situated on a saturated environment with a hydraulic level of water (Figure 2), according to the theory

of small orifices discharge [1]

[1]

Where: Q 0 = discharge flow; C d = Discharge

coefficient; H0 = head difference

In order to simulate a more realistic scenario

of pathogen intrusion was modeled the pipe

buried by sand In this way, it was consider the

porous soil around the pipe With these

simulations the analysis allow to conclude how

the relationship is affected with the porous

media The granular material was sand with a

Coefficient of Uniformity of 2.4 and a

Coefficient of Curvature of 0.8, the values

indicate that the sand is well graded Figure 2 Model of intrusion in steady-state flow

The pipe is 32mm nominal diameter and 2.4mm thick The exterior tank had a constant water level of 0.37 m The pressure inside the pipe varied between -12.4 to -51.7kPa Five scenarios of negative pressure were simulated, these five pressure scenarios were chosen randomly The flow intruded on the pipe varied from 2.14×10-6 to 3.09×10-6m3/s The results of the relation of pressure drop with the

Ground Saturated soil

Pipe with defect Negative pressure level during transient

0 0

flow entering to the pipe are shown in figure 3 The fitting of the data is the relation between the pressure and the flow that is the base to calibrate the porous soil in the CFD model

Figure 3 Intruded flow rate and pressure through round hole 1mm with soil around the pipe

2.2 Model scenario of intrusion flow during a water hammer

The model begins with a hydroneumatic air

Wessel tank that maintains a pressure

required in the scenario The pipe is of

polyethylene 200m length and 50mm

diameter, downstream has a free discharge It

was installed a globe valve at the beginning of

the pipe in order to generate a water hammer

upstream Downstream the valve it was made

an orifice of 2mm to represent the failure and

where the flow is going to enter during the

transient (Figure 4) The circular orifice was

4mm diameter

Figure 4 Model of intrusion during a water hammer

The data were measured with a transducer pressure nearly to the orifice and the flow was measured

by a V-notch triangular Weir The initial data were: Pressure =1.47×105Pa and flow rate = 0.00254m3/s The time to close the valve to initiate the water hammer was 0.17s and the oscillations time was 11s The volume flow entering through the orifice during the transient was measured assisting with a high speed video with 500 fps, the results are shown on the figure 5; with those results where validated the numerical models

Figure 5 Experimental results of hydraulic transient and flow intruded during a negative oscillation

2.3 Pathogens around the water distribution networks

The relation of the pathogens with the water distribution networks are reported on the cases described of water born disease outbreaks Hunter (1997) reported that 15 of the 57 outbreaks in public water supplies in the United Kingdom between 1911 and 1995 were associated with contamination within the distribution In United States, 18% of 619 outbreaks reported in public water systems from 1971 to 1998 were caused by chemical or microbial contaminants entering to the distribution system or water that was corrosive to plumbing systems within premises (Craun & Calderon, 2001)

Several studies have demonstrated that the soil surrounding buried pipe can be contaminated with faecal-indicator microorganisms and pathogens (Kirmeyer 2001; Besner et al., 2011) In addition to contaminated soil, runoff from streets and agricultural land can contain high concentrations of microbiological and chemical contaminant Failures in physical and hydraulic integrity can lead to the influx of contaminants coming from groundwater, polluted runoff, wastewater, or some other form of contamination The problem of pathogen intrusion is aggravated in developing countries where the pollution sources come into contact with water distribution systems and intermittent water supplies are prevalent (Vairavamoorthy et al., 2007)

Karim et al (2003) collected 66 samples of soil and water around the pipes at eight sites in six states

in the United States The samples were taken to determine the presence of pathogen organisms About half of the water and soil samples showed total and feacal coliform bacteria indicating the presence of contamination of microorganisms Viruses were detected in 56% of the samples Although, in this study the samples were not analysed for bacteria, but the leaks of sewer networks provide a highly environment favourable to its growth (Pedley et al., 2004) In order to identify the pathogens and to quantify its concentration, there were collected samples of puddle water on streets and water from irrigation channels near to water distribution networks (Figure 6)

Figure 6 Sites of water sample to identify pathogens near water distribution networks

Table 1 shows the count of colonies after 24 hours of incubation The presence of the pathogen has been remarkable from the samples at it has a similar magnitude order than they obtained by Karim in

2003

Table 1 Results of pathogen on the samples around water networks

Puddle water Salmonella E Coli 4.3×104.7×1055 Irrigation channel water Salmonella E Coli 4.9×101.4×1065

3 NUMERICAL MODELS

3.1 MOC model

The experimental transient pressures were modelled by the water hammer equations, through the simplified continuity and momentum equations to obtain the boundary conditions for the CFD model Those equations are a set of two hyperbolic partial differential equations (Chaudhry, 1987; Wylie & Streeter, 1993; Ramos, 1995) For the solution of one-dimensional hydraulic transients, the Method of Characteristics (MOC) to the typical “reservoir-pipe-valve” gives the following equations

:

0 :





























I Q Q gA

c H

H

C

I Q Q gA

c H

H

C

B P B

P

A P A

P

[2]

Where: C = characteristic lines; H = head, c = wave speed; A = area; Q = flow; I = headloss term To

obtain numerically the dissipative effects of the installation and to represent the phenomena of vapour–cavity during the transient, a particular section of the pipeline has a pressure below vaporization pressure, it is calculated as an internal boundary condition The effect of air release is neglected and discrete vapour cavities can open at all pipe sections Hence macro-cavitation (large cavities) can be characterized by the existence of a vapour cavity volume [3]

[3]

Where: V = volume of vapour at the section i for the time j; Q R and Q L= discharge values at the right and left of the cavity; t = numerical time increment Ramos et al (2004) has presented an approach

of the damping effect considering the damping of the pressure peaks throughout time This dynamic effect can be influenced, on the one hand, by the non-elastic behaviour of the pipes, and on the other hand, by the friction effect The objective of the proposed technique is to allow the characterization of the energy dissipation, through the variation of the maximum and minimum pressure head observed

in a transient regime [4]

Where: h = head; K visc and K elas = decay coefficients for plastic and elastic effects, respectively; h 0 = dimensionless head at time 0 h0 = relative head losses; = t/(2L/c) The coefficients implemented

by Ramos (2004) to represent the physical behaviour are two parameters [5], the first one, the reduction in the head variation when induced by a discharge variation by non-elastic fluid and pipe deformation and the second one, the reduction in the discharge value caused by a head variation, due to a non-elastic response in the recuperation phase of the occurred deformation [6]

 , , 1 , , 1 2

1

,

,j j  Q Ri jQ Ri j Q Li jQ Li j t

visc elas h

K visc

elas

K

K e

h

K

K

h

visc

1 1

0 0

0























  

Experimental Pressure

The fitting of the numerical MOC model of the physical hydraulic transient is shown on Figure 7

Where: Q = discharge variation; KQ = reduction coefficient; H = head variation; J = headloss term;

c = wave speed; S = pipe cross-section

3.2 CFD model

The physical events of intrusion considering the soil around the pipe and the hydraulic transient were simulated together on a CFD model to quantify the potential intrusion volume during the water hammer The CFD model is used for solving numerically the governing laws of fluid dynamics Equations based on conservation fluid mechanics, considering dissipative and turbulent effects [6] and [7], are solved in a mesh domain, given suitable initial and boundary conditions

[6]

[7]

Where:  = density of the fluid; U i = mean velocity; u’ = fluctuating component of the velocity; p =

pressure; F ext = exterior source of mass The geometry of the pipe where the intrusion is represented includes the orifice and the exterior zone The mesh is designed with polyhedral elements (Figure 8)

Figure 8 Geometry and mesh for the CFD model Turbulence effects are simulated with one of the variations of the  model The model "Realizable k-ε" is a variation of the computational program STAR-CCM (2008), this model was developed by Shih et al (1995) It contains a variation in the transport equation for the rate of dissipation turbulent ε

In addition, it is proposed that the coefficient Cμ depends on the flow and of the properties of the

turbulence, it allows the model to satisfy certain mathematics constraints in the normal stresses consistent with the physics of turbulence The expressions for the transport in the model "Realizable k-ε" are [8] and [9]

[8]

¶

¶t redV+ re n( -ng)× da

A

V

se

æ è

ø

÷Ñe× da

A

k(Ce1Ce3G b -Ce2re)

é ëê

ù

ûúdV

V

ò

[9]

)

/(gS

c

J

H

KQ

Q  

gS

c KH











t

x j

i j

i

x j

j i j

i j

i

i

x

U x

x

p g x

u u x

U U

t

U



















































¶

¶t rkdV+ rk(n-ng)× da

A

ò

V

sk

æ è

ø

÷Ñk× da

A

ò + éëG k + G b-r e( + UM)ùûdV

V

ò

Where: k = turbulent kinetic energy;  = ratio of dissipation; G k = generation of turbulent kinetic energy

G b = kinetic energy generated by boundary push; Y M = contribution of the pulsatile expansion

associated to the compressible turbulence; C, C, C= constants for the realizable k model; k

and = Prandtl numbers for k and e respectively; Se = global variation in the time of parameters

The soil around the pipe is simulated on the CFD model through the relationship called "Porous Source" (STAR-CCM, 2008) where the properties of the porosity are specified from the equation [10]

v

P

F P   where P  Pv Piv [10]

Where: F P = Porous media; P, P v and P i = porous resistant tensor, linear resistant tensor due to viscosity and inertial resistant tensor, respectively To begin the simulation, the first scenario is considered with the initial conditions Constant pressure on the outlet boundary and known velocity on the inlet obtained from the flow data the stationary-time event is modelled The CFD code is then adapted to simulate the transient regimes for the scenario, where the solution on the time scale is unsteady implicitly and the numerical results for the velocity and the pressure at the inlet and outlet pipe section, obtained by the MOC vapour–cavity model as the boundary conditions of the pipe element to be analysed The time step is 0.4167s, running 37 steps, and the solution to each step is until 5,000 iterations Figure 9 represents the pressure in different time steps, when the velocity comes in and out of the pipe

Figure 9 Passive Scalar configuration representing the intruded flow during the water hammer with exterior soil

3.3 Simulation results

After numerical simulations of the CFD model, the volumes of intrusion and leakage during the transient event are obtained in order to quantify the potentiality of the pathogen contamination that can occur in drinking water pipe systems, the numerical results were compared considering the soil around the pipe and without soil (Figure 10)

The volume of intrusion without soil at the end of the water hammer is 3.5×10-4m3, considering the duration of the transient the intrusion flow is 1.1% related to the flow through the pipe before of the transient event Considering the soil around the pipe the volume of intrusion reduces 60% and the volume is 1.4×10-4m3 This volume of intrusion could be representative depending on the polluted source If it is consider the Salmonella concentration of 1.4×105UFC/100ml, the presence of pathogens caused by the intrusion during the transient event considering soil around the pipe with a 4mm circular would be 1.9×105UFC of Salmonella In fact, no presence of salmonella should be measured in water for good conditions

Two scenarios were defied and modelled for pathogen intrusion, in order to obtain the numerical models to represent the phenomenon during hydraulic transients considering the soil around the pipe

In this case, the results confirmed the fitting of numerical and experimental models A vapour–cavity MOC model was used to generate boundary conditions for the CFD model When soil has been considered around the pipe, the volume of intrusion and the pollutant concentration are considerably reduced The methodology to quantify the potential of the intrusion is proposed to improve the

knowledge and reliability in this type of event in the operation of water supply systems

Besner, M C., Prévost, M., & Regli, S (2011) Assessing the public health risk of microbial intrusion events in distribution systems: Conceptual model, available data, and challenges Water research, 45(3), 961-979

Chaudhry, M H Applied Hydraulic Transients, Litton Educational Publishing Inc.Van Nostrand Reinhold Co 1987

Craun, G F., & Calderon, R L (2001) Waterborne Disease Outbreaks Caused by Distribution System Deficiencies Journal-American Water Works Association, 93(9), 64-75

Friedman, M., Radder, L., Harrison, S., Howie, D., Britton, M., Boyd, G., Wang H., Gullick, R., LeChevallier M., Wood, D., & Funk, J (2004) Verification and control of pressure transients and intrusion in distribution systems AWWA Research Foundation and US Environmental Protection Agency

Hunter, P (1997) Waterborne disease: epidemiology and ecology John Wiley & Sons

Hunter, P R., Waite, M., & Ronchi, E (Eds.) (2002) Drinking water and infectious disease: establishing the links CRC Press

Karim, M R., Abbaszadegan, M., & LeChevallier, M (2003) Potential for pathogen intrusion during pressure transients Journal-American Water Works Association, 95(5) 134-146

Kirmeyer, G J., & Martel, K (Eds.) (2001) Pathogen intrusion into the distribution system American Water Works Association

LeChevallier, M W., Gullick, R., Karim, M., Friedman, M., & Funk, J (2003) The potential for health

risks from intrusion of contaminants into the distribution system from pressure transients J Water

Health, 1, 3-14

Pedley, S., & Bartram, J (Eds.) (2004) Pathogenic mycobacteria in water: A guide to public health consequences, monitoring and management IWA Publishing

Ramos, H Simulation and Control of Hydrotransients at Small Hydroelectric Power Plants PhD thesis, Technical University of Lisbon, Portugal (in Portuguese) 1995

Ramos, H., Covas, D., Borga, A., & Loureiro, D (2004) Surge damping analysis in pipe systems:

modelling and experiments Journal of hydraulic Research, 42(4), 413-425

Shih, T H., Liou, W W., Shabbir, A., Yang, Z., & Zhu, J (1995) A new k-ϵ eddy viscosity model for high Reynolds number turbulent flows Computers & Fluids, 24(3), 227-238

STAR-CCM+ 3.04.009, User’s Guide, CD-Adapco, USA, 2008

Vairavamoorthy, K.; Gorantiwar, S D & Mohan, S (2007) Intermittent Water Supply under Water Scarcity Situations Water Int 32(1), 121-132

Wylie, E B & Streeter, V L Fluid Transients in Systems, Prentice Hall 1993

Zloczower, N & Charuv, K K., Control of transient induced contaminant leakage and infiltration by implementation or air valves Water Environment Federation Collection Systems 2009