Biofouling in water distribution and drainage systems

Accepted Mauscript, Environmental Technology Reviews (16/06/2014), DOI: 10.1080/09593330.2014.923517

1

Biofilm development in water distribution and drainage systems:

2

dynamics and implications for hydraulic efficiency

3

M.W. Cowlea; A.O. Babatundea*; W.B. Rauenb; B.N. Bockelmann-Evansa; A.F. Bartonc

4 5 6 7 8 9 10

a

b

Hydro-environmental Research Centre, Cardiff School of Engineering, Cardiff University, The Parade, Cardiff, CF24 3AA, UK

Graduate Programme in Environmental Management, Universidade Positivo, Av Prof. Pedro Viriato Parigot de Souza, 5300, Curitiba/PR, 81280-330, Brazil c

School of Science, Information Technology and Engineering, Federation University Australia, Victoria 3353, Australia *

Corresponding Author: [email protected]

11 12 13 14

This is an accepted manuscript of an article published in Environmental Technology Reviews on 16/06/2014, available online: http://www.tandfonline/10.1080/09593330.2014.923517

15 16

Abstract

17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

Pipeline distribution systems account for the vast majority of the physical infrastructure in the water and wastewater industry. Their effective management represents the primary challenge to the industry, from both an operational and public health standpoint. Biofouling is ubiquitous within these systems, and it can significantly impede their efficiency, through increase in boundary shear and associated flow resistance caused by characteristic change in surface dynamics. Nonetheless, conventional pipeline design practices fail to take into account such effects, partly because research findings that could contribute to upgrade and optimise design practices appear scattered in the literature, and are often offering conflicting views as to its causes. This makes it difficult for the adoption of adequate predictive and preventative measures. The aim of this review is to update and contribute to a better understanding of the development and impact of biofilms and biofouling within water management pipelines, particularly within the academia and the general engineering community. The review has confirmed that the potential impact of biofouling on pipeline performance can be significant and that current design approaches are outdated for biofouled surfaces. Further research on this topic is therefore, essential, to ensure that both current and future systems are as effective as possible, both environmentally and financially. In particular, more advanced mathematical modelling frameworks which include the dynamic and case-specific nature of biofouling should be developed. Such a framework could give rise to a real time monitoring platform to assist the adoption of more cost-effective approaches to maintain and repair the system.

36 37

Keywords: Biofilm, biofouling, drainage system, drinking water distribution system, hydraulic efficiency

38

1


1

1. Introduction

2

1.1 Background

3

The effective management of pipeline distribution systems is arguably the single most

4

important challenge to the water and wastewater industries from both an operational and

5

public health standpoint. This challenge is exacerbated by the environmental complexities of

6

such systems which can have highly diverse and variable flow rates, contents, and

7

temperatures. Fouling mechanisms (individually and cumulatively, see Figure 1) both

8

contribute and are governed by these inherent complexities.

9

pipeline biofouling which refers to the natural, albeit sometimes undesirable process through

10

which a complex microbiological slime layer, composed of microbial cells and colonies

11

embedded within a highly hydrated, protective polymer matrix – referred to as a biofilm –

12

forms upon the surface of the pipeline. The term biofouling, can also include the

13

physicochemical interaction of the biofilm with the pipe surface and external environment

14

such as scaling and corrosion. Although, the focus of this paper is solely on biofouling, it

15

should be noted other mechanisms such as scaling and the accumulation of sediments,

16

loose deposits and FOGs (fats, oils and greases) (see Figure 1) can contribute to fouling

17

within pipelines and subsequently impair their ability to convey flow resulting in reduced

18

hydraulic efficiency.

19

2

Of particular concern, is


a) Rising/Force Main

b) Gravity Sewer Main

c) Drinking Water Main

Dry Biofilm High Flow Level Mean Flow Level

Wet/Dry Biofilm

Low Flow Level

Wet Biofilm

Wet Biofilm Sediment/Loose Deposits, including organic and inorganic material

Sediment/Loose Deposits, including organic and inorganic material (including fats, oils and greases)

1 2 3

Figure 1. Typically fouled water and wastewater pipes, including a) rising/force wastewater main, b) traditional gravity fed wastewater main, and c) a pressurised drinking water main.

4 5

Any pipe conveying a liquid is potentially susceptible to biofilm development and thus

6

biofouling to some degree, as microorganisms; namely bacteria, algae, fungi, mosses and

7

invertebrates seek to exploit the desirable growth conditions that their surface provides.

8

Such ecological advantages include: a constant source of nutrients, sufficient aeration and

9

waste removal [1, 2]. The resulting microbial system, typically dominated by bacterial

10

species – in particular the Proteobacteria phylum (namely Bataproteobacteria, and

11

Gammaproteobacteria within drinking water, wastewater and hydropower pipelines) is

12

generally classified in terms of their structure (on the macro-scale) as either low-form

13

gelatinous or filamentous (or both), with the former being more common within most pipeline

14

systems [3-5].

15

The presence of such microbial structures on the surface can significantly alter the

16

pipelines solid-liquid interface, typically resulting in increased boundary shear stresses and

17

associated flow resistance, thereby affecting the pipe’s hydraulic efficiency over time. For

18

example, the primary cause of energy losses and thus flow capacity reductions within

19

pipelines is due to friction along the solid-liquid interface which tends to increase with

20

increasing surface roughness and interface instabilities [6]. This is illustrated in Figure 2,

3


1

which shows the impact of biofouling on the performance of a range of concrete pipelines

2

with internal diameters from 400 mm to 1000 mm. This is evident through an increase in flow

3

resistance resulting from an increase in equivalent roughness (namely the Nikuradse-type

4

equivalent sandgrain roughness, ks), as estimated using the widely applied Colebrook-White

5

(C-W) equation [7, 8]. From Figure 2 it can be seen that an increase in ks from the “clean”

6

pipe value of 0.06 mm to the UK’s recommended ks range for fouled pipes, namely 0.6 mm
0.60

3

m/s), whereby sediments and other loose deposits remain suspended within the water

4

column or at least deposited in areas and amounts which can be re-suspended later

5

following the next flow event [32], an accurate wall roughness must be known. Otherwise,

6

the deposits will accumulate upon the surface, further impeding the flow and potentially

7

resulting in clogging, surcharge and ultimately flooding issues [34]. Moreover, an accurate

8

underlying wall roughness is also required in the modelling of effective flushing strategies [35,

9

36].

10

Furthermore, within DNs, biofilms also contribute to the production of unwelcome

11

gases, namely hydrogen sulfide and methane, which present their own problems for the

12

industry, ranging from odour and corrosion issues, to potentially endangering maintenance

13

crews [37, 38]. Hence, biofouling is likely to be more substantial and have a more significant

14

impact in DNs than in DWDSs. This is the general perception within the drinking water and

15

sanitation industries, whereby biofouling is perceived to have a greater impact upon DNs

16

than DWDSs. Nevertheless, greater emphasis within the industry and in literature is put on

17

the latter, primarily due to the greater pumping requirements of the application and biofilm

18

related water quality issues. Such water quality issues include impeded taste, odour and

19

colour; in addition to causing potential health problems to consumers, ranging from viral and

20

bacterial gastro-enteric diseases, to infections such as hepatitis A and giardiasis [5, 39-41].

21

Furthermore, with DWDSs, the biofouling impact on surface roughness is generally

22

considered to be of secondary importance to these water quality issues. This is because the

23

poor water quality will generally result in more customer complaints and it is generally

24

considered to be compromised by a very thin biofilm (i.e. > 30 μm). Therefore, it is the

25

general practice of pipe owners to make use of disinfectants and regular flushing events to

26

minimise biofouling within DWDSs. However, biofilms have been found to have a high

27

resilience to these control measures [5], and in any case, even a relatively thin biofilm (i.e. >

7


1

160 μm) can potentially cause a considerable increase in the pipe’s frictional resistance

2

(Schultz and Swain 1999), particularly in long pipe runs. The early observations within water

3

mains by Seifert and Kruger [23] and Sharp [24] and Minkus [22], also highlight the potential

4

impact that biofouling can have on DWDSs, notwithstanding the reported biofilm thickness’s

5

(i.e. the order of 1 mm to 9.4 mm) being unrepresentative of biofilm typically found within

6

modern, well maintained DWDSs (which seldom exceed 1 mm). Furthermore, the resultant

7

decreases in flow capacity within DWDSs as a result of biofouling, will also increase the

8

planktonic (free-floating) bacteria concentrations, through an increase in pipelines hydraulic

9

retention time (HRT) [42]. Consequently, the water quality is impaired and likelihood of

10

further fouling and fouling issues (i.e. public health problems) is increased. Therefore,

11

although, the magnitude of growth may differ significantly between the two applications (i.e.

12

DN > DWDS), the impact will nonetheless be considerable in both cases for different

13

reasons, i.e. in terms of operational performance/costs and public health.

14 15

3. Biofouling within Pipelines – Nature, Processes and Causes

16

3.1 The Process of Biofouling

17

The majority of all biofilm development occurs within a thin layer located near the solid-liquid

18

interface, and which coincides with the boundary layer defined in fluid mechanics [43].

19

Typically, biofilm development comprises of four stages, (as shown in Figure 3), namely: (i)

20

conditioning stage; (ii) initial cell attachment stage; (iii) main development stage; and (iv)

21

equilibrium stage. These stages are outlined below;

22

(i) Conditioning Stage – this stage is initiated within seconds of the biological matter

23

entering the pipeline, with the spontaneous adsorption and formation of a conditioning layer

24

or film. The conditioning film is formed mainly by organic molecules, however, films

25

consisting of inorganic materials such as metallic oxides or fine clays have also been

26

documented [44]. The duration of the conditioning stage is commonly referred to as the lag8


1

time, and it is a function of the operational conditions of the pipeline including fluid viscosity,

2

flow hydrodynamics and surface roughness (see Figure 6 and Table 1).

3

(ii) Initial Cell Attachment Stage - Initial microbiological adhesion occurs during the

4

initial cell attachment stage and this is predominately encouraged by the conditioning film,

5

owing to: i) neutralisation of the surface charge, ii) provision of nutrients and iii) polarisation

6

of the forces between the film and the microorganisms. Therefore, the conditioning film is

7

essentially the catalyst in the initial attraction and attachment of the discrete planktonic

8

bacterium, and is therefore, a vital component in the successful development of biofilms

9

within pipelines. Initially, the surface will only consist of a few randomly distributed cells (or

10

initial colonisers), adhered to the surface via weak, reversible forces known as Van-der-

11

Waals forces [45]. Cell division and EPS secretion then follows, along with the formation of

12

substantially stronger bonds, which anchor the now densely packed cell matrix to the pipe

13

surface [46].

14 15 16

Figure 3 - Idealised biofilm development for a high and low flow velocity scenarios. Initially the surface becomes colonised by nutrients and discrete microorganisms, which is then followed by rapid growth

9


1 2

and the formation of a dense film (log growth phase) and then ends with stabilisation (steady state) and the detachment of dead material from the surface (plateau and death phase) [19, 46].

3 4

(iii) Main Development Stage – This stage is characterised by further colonisation

5

and growth which takes place over time. This results in an increasingly thicker and denser

6

structure, which protrudes further into the flow. Within the boundary layer, viscous effects

7

cause the flow velocity to decrease steadily to zero at the wall. In the near vicinity of the wall,

8

the turbulent fluctuations of the flow are considerably reduced [43]. Therefore, as the biofilm

9

structure grows in the direction normal to the wall, different parts of the biofilm will be subject

10

to different conditions, which become gradually more hostile as the distance from the wall

11

increases. This stage of development continues until a point of equilibrium is reached

12

between the favourable and adverse growth conditions. Typically, under idealised conditions

13

(i.e. sufficient nutrient availability) and within many drainage networks, this will occur when

14

the biofilm has extended through the boundary layer and into the outer flow region. At this

15

point the biofilm’s internal cohesion is significantly impaired by the numerous adverse

16

conditions associated with the outer flow region, namely increased flow shear.

17

(iv) Equilibrium Stage, “Steady State” - provided that the environmental and

18

operational conditions (e.g. flow velocity, boundary layer structure, nutrient content etc.)

19

remain reasonably constant, the biofilm thus formed tends to reach a pseudo-steady state.

20

Depending on the subjected conditions this can be between 14 to 385 days [47, 48], with the

21

latter typically associated with low nutrient and DWDS conditions (i.e. Assimilable Organic

22

Carbon, AOC in the order of 5 µg/L).

23

Once a biofilm has formed on a section of a pipe under favourable conditions and

24

locations (i.e. at joints and bends), it may quickly spread through the entire pipeline system

25

and induce colonisation in other areas that were not initially favourable to growth. This

26

happens as cells and/or whole clusters are “sloughed off” the surface and are carried by the

27

flow as floating biofilms which then settle downstream [49, 50]. Further growth can be

10


1

promoted downstream by the waste generated by the upstream biofilms, which can be

2

utilised as either a conditioning material or as a direct source of nutrients. Therefore, the

3

attainment of a pseudo-steady state in a region of a pipeline is not necessarily indicative of

4

equilibrium in the whole system.

5

It should be noted that in a given pipeline, the properties of the mature biofilm such

6

as its overall structure; surface topography; thickness; morphology and microbial

7

composition [5, 15, 19] will change over time. This is due to competition between biofilm

8

species, the relatively short term survival of biofilm species (the life and death cycle), and the

9

varying (seasonal and daily) operational and environmental dynamics of the pipeline (Table

10

1).

11 12 13

Table 1. Key aspects and perceived impacts on i) substrate accumulation ii) biofilm structural composition and iii) biofilm dynamic behaviour due to flow interaction, within pipelines Factor

Impacts upon growth (i) (ii) (iii)

Residual disinfectant concentration* Flow hydrodynamics

 

Nutrient and biological content

Surface Material









Reference Cloete et al. [51], Hallam et al. [47], Tsvetanova [52], Zhou et al. [53] Pedersen [54], Stoodley et al. [30], Stoodley et al. [16], Cloete et al. [55], Lauchlan et al. [56], Tsvetanova [52], Lambert et al. [10] Costerton and Lewandowski [1], Melo and Bott [46], Stoodley et al. [30], Gjaltema et al. [57] Pedersen [54], Van der Kooij et al. [58], Niquette et al. [59], Hallam et al. [47], Manuel et al. [60], Zhou et al. [53]

Seasonal/Daily Perkins and Gardiner [61], Barton   Variations in [15] Conditions *Only applicable within certain drinking water distribution systems 14 15 16

11


1

3.2 Causes and Nature of Biofouling within Pipelines

2

A number of factors impact upon biofilm development within pipelines and the resultant

3

growth interacts with the system hydraulics. This section is focused on the review of the key

4

factors as listed in Table 1 with the exception of disinfectant concentration which has been

5

comprehensively reviewed over the years, most notably by Bridier et al. [62].

6

3.2.1 Flow Hydrodynamics and Nutrient Availability

7

There is an inherent link between flow hydrodynamics and nutrient availability on biofilm

8

development, owing to their influence on mass transfer and diffusion rates.

9

The mass transfer and diffusion rates of a system are predominantly governed by the

10

level of turbulence in the flow, which is usually estimated by the Reynolds number (Re). Re

11

represents a balance between the magnitude of inertial and viscous forces. Since high Re

12

values are associated with high velocity flows, it follows that viscous effects are not

13

important in establishing the flow condition in the turbulent flow regime. Low Re values

14

indicate that viscous effects significantly influence the flow condition under relatively low

15

speeds. The flow within most DWDSs and DNs is typically turbulent in nature. However,

16

laminar flow conditions have been observed especially in areas of low water consumption

17

(i.e. typical at night and rural areas) and/or towards the end of long branches and the

18

network periphery, where the flow can be very low or periodically stagnant.

19

Conceptually, the boundary layer of turbulent flows can be divided into two regions

20

namely; (i) the viscous or laminar sublayer and (ii) the logarithmic sublayer. The viscous or

21

laminar sublayer is closest to the pipe wall and has a thickness calculated as δ’ = ν/u*;

22

where ν is the kinematic viscosity of the fluid and u* is the shear velocity which is calculated

23

as u* = (τ0 /ρ)

24

the value of δ’ is a function of the type of fluid and the flow condition. For example, an

25

increase in flow rate leads to an increase in u* and a decrease in δ’ for the same fluid, i.e. it

26

causes a reduction of the boundary layer thickness. Beyond the logarithmic layer lies the

; where τ0 is wall shear stress and ρ is the specific mass of the fluid. Hence,

0.5

12


1

outer flow region, where the mean flow velocity is that of the free stream. As illustrated in

2

Figure 4, there are three types of boundary layers, namely hydraulically smooth (Figure 4a),

3

transitional (Figure 4b) and hydraulically rough (Figures 4c and 4d). The classification

4

depends upon the thickness of the absolute surface roughness height (k) relative to δ’. A

5

boundary layer is classed as hydraulically smooth for k < δ’ and it is classed as hydraulically

6

rough for k > δ’. For k  δ’, the boundary layer is classified as transitional. For each of these

7

classifications, the influence of the surface roughness on biofilm development is inherently

8

different, with the greatest impact occurring under hydraulically rough conditions and the

9

least impact under hydraulically smooth conditions.

10

11 12 13

Figure 4. Boundary layer classifications, including; a) hydraulically smooth, b) transitionally rough, and c) hydraulically rough [15]

14 15

Previous studies have shown that the boundary layer structure is altered by the

16

presence of a biofilm [18, 63]. Andrewartha and Sargison [63] found that biofilms altered

17

both the turbulent structure and thickness of the boundary layer. The altered boundary layer

18

then impacts upon further biofilm development, thereby establishing a dynamic two-way

19

(symbiotic) feedback relationship. This process has a subsequent effect on flow resistance

20

and flow rate. Such changes in the operating conditions affect δ’ further, as well as its

13


1

relationship to k, and so on until a point of equilibrium can be reached. This complex matrix

2

of interacting causes and effects is illustrated in Figure 5. If these conditional changes are

3

significant, then they can be considered as influential upon the resulting biofilm development

4

as the flow hydrodynamics.

5

6 7 8

Figure 5. Schematic representation of the dynamic feedback relationship that exists between the boundary layer hydrodynamics, biofilm development, operational and environmental conditions

9 10 11

The degree of influence that flow hydrodynamics can have upon

biofilm

development is highly dependent on the system’s flow classification [30, 64].

12

In laminar flow conditions there is a relatively thick boundary layer. The ample

13

boundary layer and the low near wall shear forces are in theory conducive to successful

14

biofilm development [30]. However, such a large boundary layer combined with the inherent

15

lack of mixing within laminar conditions is non-conducive to successful mass transfer, as it is

16

likely to retard the influx and diffusion of microorganisms, dissolved oxygen and nutrients to

17

the surface, thus potentially impairing overall biofilm growth rate. On the other hand, within

18

DWDSs which utilise disinfectants, the retarded diffusion rates are likely to reduce the

19

disinfectant’s effectiveness, to the benefit of the biofilm. Moreover, the low flow speeds in

14


1

laminar conditions will promote planktonic growth, through an increase in HRT, which will

2

subsequently increase the likelihood of accumulation and growth on the surface [42].

3

Ultimately, laminar conditions provide numerous benefits for successful and significant

4

biofilm growth, although, it’s overall growth rate will be impaired by the low diffusion rates.

5

Consequently, the resultant biofilm coverage, within laminar conditions is generally irregular,

6

isolated and sparsely located across the surface [30, 64, 65]. For example, both Gjaltema et

7

al. [57] and Stoodley et al. [30] noted that in low nutrient and flow (i.e. laminar) conditions,

8

the resultant biofilms were isolated and sparsely located. The overall effect of a biofilm on

9

frictional resistance under laminar conditions has been found to follow the traditional smooth

10

pipe friction law relationship [10]. Whereby, the overall pressure drop is primarily influenced

11

by skin friction and hence by the total surface area of the biofilm as opposed to the shape or

12

structure of the fouled surface [16]. Therefore, in laminar conditions, conventional design

13

guidelines [9, 25, 26] will apply to biofouled surfaces.

14

In fully turbulent flow conditions, the laminar sublayer reduces significantly in

15

thickness relative to the total boundary layer thickness. In such situations, the frictional

16

resistance of the biofouled surface has been observed to increase dramatically with Re [10,

17

27]. The overall pressure drop in turbulent conditions is influenced to a greater extent by

18

surface roughness, which produces form drag when sufficiently great (i.e. from transitional to

19

hydraulically rough). Therefore, the structure, shape and nature of a fouled surface also has

20

the ability to influence the pressure drop in fully turbulent flow [16]; and in turn, these

21

characteristics are significantly affected by turbulence. The considerably reduced laminar

22

sublayer and increased turbulent mixing in the near proximity of the wall (induced by the

23

presence of the roughness element within the logarithmic region) greatly increases the influx

24

and diffusion of microorganisms, dissolved oxygen and nutrients to the surface. Thus, the

25

resultant biofilm coverage is likely to be more dense and compact than in laminar conditions

26

[66-68]. The additional turbulence will also result in more efficient waste removal. Therefore,

27

the favourable mass transfer and diffusion rates will likely increase the overall fouling and

15


1

the inherent relationship between turbulent flow and surface roughness will significantly

2

accentuate its total impact [67]. Percival et al. [67], found more rapid and extensive biofilm

3

growth at relatively high Re (including Re = 1.90x104 and 3.50x104), which was followed by a

4

statistical steady-state. However, Lambert et al. [10] observed a significant decrease in the

5

biofilm thickness as a result of the increased turbulence in the vicinity of a pipe bend.

6

Therefore, flow shear is a key controlling factor on biofilm development within pipelines, and

7

its resultant equivalent roughness scale [29]. Moreover, the typically dense and compact

8

coverage inherent within turbulent conditions (with sufficient nutrient loading) may lead to

9

skimming flow, i.e. the relocation of the velocity profile to the top of the roughness element –

10

which is, in this case, the top of the biofilm layer [16]. Skimming flow has been documented

11

to cause significantly higher flow resistance, and can be triggered by as little as an 8.3%

12

surface coverage [69]. Other factors contributing to form drag, namely the biofilm’s shape

13

and thickness, are likely to have a greater impact upon the overall pressure drop under

14

turbulent flow conditions after the onset of skimming flow [16].

15

The favourable mass transfer and diffusion rates associated with turbulent flow

16

conditions will also amplify the overall impact of nutrient loading. For example, Melo and

17

Bott [46] reported a 400% increase in biofilm thickness when nutrient levels increased from

18

4.00 mg/l to 10.00 mg/l, within a system in which the average streamwise velocity remained

19

constant at 1.20 m/s. Alternatively, irrespective of the favourable mass transfer conditions, if

20

the nutrient loading is reduced or is relatively low to start with the opposite is likely to occur,

21

and the overall growth and development will tend to be more restricted and sparse, i.e.

22

similar to that in laminar conditions [30, 46, 57].

23

Another important hydrodynamic aspect is that of the formation of elongated cell

24

clusters in the downstream direction (known as streamers) which have been documented to

25

occur under high flow conditions [16, 30, 64, 67]. Howevr, such filamentous biofilms can also

26

develop irrespectively of the hydrodynamic conditions, provided that certain bacteria species,

27

such as Hyphomicrobium, Spharotilus and Beggiatoa are present. The resulting cell 16


1

formation will further aid cell adhesion by providing a greater attachment and shelter area, in

2

addition to providing the embedded microorganisms greater access to essential nutrients

3

and dissolved oxygen within the flow [67].

4

Consequently, systems and areas (i.e. contractions, expansions and bends) of high

5

turbulence are likely to foster substantial and dynamic biofilm growth, but the maximum

6

biofilm thickness is limited. Moreover, unlike within laminar conditions, current design

7

practices and theories cannot accurately evaluate the resultant growths frictional behaviour.

8

This, coupled with the complex growth patterns and dependences inherent in turbulent

9

conditions makes the task of designing an efficient pipeline challenging, if not impossible.

10

3.2.2 Pipe Material

11

Microorganisms have been found to adhere and thrive upon a wide variety of pipe materials,

12

ranging from concrete and metal, to plastic-based materials, such as high-density

13

polyethylene (HDPE) and polyvinyl chloride (PVC) [59, 70, 71]. The properties of these

14

materials that have been shown to have a significant impact upon microbial attachment and

15

subsequent biofilm development include:

16

(i) Surface roughness - typically, all microbial material found within pipelines are likely

17

to be significantly smaller than the gaps and crevices that make up the overall surface

18

roughness. Therefore, they will often find shelter and protection from turbulent flow and

19

shear forces within these roughness elements. This type of protection is only required within

20

the logarithmic region, and consequently surface roughness is only likely to affect microbial

21

accumulation and biofilm development when the boundary layer is classified as either

22

transitional (see Figure 4b) or hydraulically rough (Figure 4c). However, when the system is

23

classified as hydraulically smooth (see Figure 4a), the surface roughness is unlikely to have

24

any significant impact upon the degree of microbiological material attachment, other than

25

providing a greater surface area. This is because the relatively low velocities occurring in the

26

laminar sublayer are less likely to dislodge deposited materials. Moreover, in situations of

17


1

low resistance to attachment, biofilms may fixate upon the roughness peaks or high points to

2

gain an ecological advantage within the flow regime. In such situations, i.e. in laminar

3

conditions, traditional hydraulic theory applies.

4

Within transitional or hydraulically rough conditions the magnitude of the absolute

5

surface roughness will either promote or hinder microbial attachment and development by

6

providing (or not) sufficient attachment area and protection. This implies that microbial

7

adhesion is likely to be slower upon smooth pipe materials, compared to rough materials [15,

8

52, 56]. Moreover, smoother surfaces will generally induce higher near wall velocities and

9

provide less protection and attachment areas than rougher materials. In contrast, the

10

rougher the material, the greater the area of protective and attachment potential, both of

11

which favour greater microbial accumulation [1, 54, 57, 67, 72-74]. By favouring initial biofilm

12

development when internal cohesive forces are relatively weak, rough surfaces are likely to

13

accommodate more mature biofilms, as a biofilm can propagate out of the numerous

14

roughness crevices, following the formation of strong and permanent internal bonds (see

15

Figure 6). Naturally, the opposite is true for relatively smooth materials. This theory is also

16

supported by the findings of several authors who have found that plastic- and copper- based

17

materials (typically smooth), supported less biofilm growth over the short and long term [54,

18

59, 70, 74]. In particular, Niquette et al. [59] found that the density of fixed biomass on a

19

cement-based surface (typically rough) was 2.63 times greater than on PVC. However, this

20

may not always be the case, as shown by Momba and Makala [71] where it was reported

21

that concrete-based materials supported less fixed biomass than plastic-based materials.

22

18


1 2

Figure 6. Biofilm propagation over time on a rough surface and some unique growth phenomena

3 4

Interestingly, Barton [15] and Andrewartha [75] suggested that certain growth

5

practices could potentially smoothen the initial roughness of a pipe surface, and therefore

6

reduce its associated frictional resistance. Such growth practices are common with low level

7

fouling, which is typically within many DWDSs. Therefore, there is a high potential that within

8

DWDSs, biofilm development could in fact improve the hydraulic performance for an initially

9

rough surface. However, this may not always be the case, as shown by Barton [15] who

10

noted that under low fouling conditions biofilms formed upon the peaks of roughness

11

elements to gain an ecological advantage. This, combined with the typically isolated growth

12

patterns associated with low nutrient loading will likely have the opposite effect, whereby the

13

overall frictional resistance of the pipeline will be significantly impaired, via an accentuation

14

in the absolute roughness and induce wake interactions.

15

(ii) Chemical composition, resistance to corrosion and abrasion – In certain

16

situations, surface roughness can be considered to be a key aspect in microbial

17

accumulation and potential biofilm growth. Therefore, any factor that can potentially affect it

18

should be assessed. For example, cement-based materials, are particularly susceptible to

19

one of the most well-known and documented corrosive processes within DNs, namely

20

corrosion via microbial production of acids (in particular sulphuric acid) [76-78]. In such

21

situations, hydrogen sulphide, a commonly found compound within DNs, is oxidised by

22

sulphur-oxidising bacterium (such as Thiobacillus Thiooxidans) to form sulphuric acid, which 19


1

attacks the surface of the pipe, increasing its internal roughness and impairing structural

2

integrity. Moreover, iron and steel based DWDSs have also been found to be impaired by

3

microbial corrosion [59]. Although, many water authorities (particularly, within the UK) do not

4

recommend the use of such materials within future pipelines, thus limiting its potential impact

5

within DWDSs.

6

The additional roughness caused by corrosion can also promote further attachment

7

of microorganisms and more importantly nutrients [59]. This leads to further microbiological

8

corrosion in a positive feedback mechanism with negative consequences for the pipe

9

structure. It is generally considered that the increased magnitude of fouling within pipes

10

made of traditional materials, compared to plastic-based materials can ultimately be linked to

11

their susceptibility to corrosion [59]. This problem has been addressed by manufacturers

12

through the lining of susceptible pipes with corrosion-resistant materials, such as plastics

13

and resins. However, this may add to the cost and carbon footprint of the fabrication process.

14

The surface topography and roughness of a pipe may also be affected by abrasion

15

caused by debris impacting upon the surface during operation. Abrasion resistance refers to

16

the ability to withstand mechanical erosion. The extent of the problem depends on the type

17

of abrasive event, frequency, flow velocity and pipe material. Abrasion is particularly high in

18

areas of high turbulence. Any selected pipe material used in DNs requires a significantly

19

high abrasion resistance, as it is not uncommon in such applications for grit and other

20

suspended solids to be present.

21

3.2.3 Seasonal effects – Temperature

22

The internal temperature of a pipeline can also have a significant impact upon the resultant

23

biofilm development. As temperature is generally considered to be a significant controlling

24

factor in biological growth, it has the potential to offset other factors, especially if high

25

enough. This is highlighted by the use of high disinfectant levels within DWDSs in summer

26

months [79]. Moreover, it has been observed that large temperature deviations can

20


1

encourage filamentous- type growth, which otherwise would not have been expected, based

2

on other conditions [15].

3

3.2.4 Summary of Interacting Conditions – Controlling Factors within DWDSs and DNs

4

In reality, in natural systems the aforementioned factors will be continuously interacting with

5

each other. By assessing these interactions, the factor deemed most influential and

6

controlling can be determined. Such information is vital in the development of improved

7

design considerations.

8

There is compelling evidence in the literature to suggest that flow hydrodynamics and

9

residual disinfectants (if utilised) are the two most influential factors governing biofilm

10

development within pipelines, due to their potential to remove existing biofilm and/or

11

counteract further growth [47, 52, 53]. This is highlighted by their common use in pipeline

12

cleaning for biofilm control and maintenance strategies [5]. The influential impact of flow

13

shear has been reported e.g. by Percival et al. [67] and Perkins et al. [27], who inferred that

14

high freestream velocities (> 1.77 m/s) limited biofilm development in the pipelines

15

investigated. Perkins et al. [27] further reported that such a control measure tended to

16

reduce the overall pressure loss within a hydropower pipeline.

17

A comparison of the relative impact of flow hydrodynamics and residual disinfectant

18

treatments was made by Tsevtanova [52], who found that the latter, which is usually only

19

applicable in certain DWDSs, was the most influential of the two. However, nutrient content

20

and fluid temperature were not examined in such study, despite their inherent importance to

21

biological growth (see Section 4.1 and 4.4). Thus, flow hydrodynamics (particularly in terms

22

of boundary layer diffusion and flow shear) and residual disinfectants (if used) are likely to

23

have an impact upon biofilm growth, although ultimately, it is the nutrient content (or lack of it)

24

that is the underlying limiting factor in DWDSs. On the other hand, within DNs where it is

25

more likely that sufficient nutrients are available, flow hydrodynamics will be the primary

26

controlling factor. However, if nutrient levels are high enough, they can potentially offset any

21


1

hydrodynamic effects [30]. Although, the nutrient levels documented within the study were

2

particularly high even for DNs.

3

The impact of material properties, as reported in the literature is seemingly conflicting.

4

A number of studies have suggested that different pipe materials can have a considerable

5

impact upon biofilm formation [13, 53, 54, 59, 71, 80, 81], while some other studies suggest

6

that any potential impact from the pipe material is offset by other factors, namely flow

7

hydrodynamics, nutrient availability and disinfectant levels [46, 47, 55, 56, 82, 83]. Therefore,

8

materials which are relatively smooth, such as plastics and metals are not necessarily less

9

rough when fouled than materials with a higher natural roughness, such as concrete. For

10

example, in an investigation involving different wastewater pipeline materials and flow

11

velocities, Lauchlan et al. [56] found that the magnitude of the material effect was

12

indistinguishable from the standard deviation of the whole data set obtained under similar

13

flow conditions. On the other hand, flow velocity was found to have a significant impact on

14

the resultant roughness scale value. A relationship between the two variables was obtained

15

however, the data scatter was large and in some cases, of several orders of magnitude. In

16

agreement with the findings of Lauchlan et al. [56], Cloete et al. [55] reported notable

17

changes in the biofilm growth rate as a result of varying flow velocities, and no significant

18

difference in growth between the pipe materials assessed. The types of materials assessed

19

included asbestos-cement, coated cast iron, galvanized steel, and PVC, all typically used

20

materials both in DNs and DWDSs. The average streamwise velocities used, however, were

21

particularly high, ranging from between 3 m/s and 4 m/s, and therefore unrepresentative of

22

most field conditions (particularly within the UK). For example, in the UK, DWDSs are usually

23

operated at average freestream velocities of between 0.04 m/s to 2.00 m/s, with most

24

tending towards the lower end of the spectrum, with 0.06 m/s being the average [41].

25

Appraising the aforementioned studies critically, it may be concluded that the

26

material properties undoubtedly have an impact upon initial attachment, both in terms of

27

protection and induced turbulent mixing. Whether this promotion of initial attachment can be 22


1

linked to future growth will likely depend on the favourability of other factors. In the situations

2

where material properties were considered more influential, factors such as flow velocity and

3

disinfectant residuals, were seemingly low (i.e. 0.00 m/s to 0.30 m/s and < 0.10 mg Cl/l -

4

0.20 mg Cl/l, respectively) [54, 59]. Such conditions can occur in many DWDSs, particularly

5

at the end of long runs and branches, and are prevalent in DNs, as disinfectants are not

6

used and flow rates are naturally low.

7

For the sake of conclusion, in reality, there is no absolute controlling factor to biofilm

8

development (in Table 1) and, thus, to the effective roughness of a biofouled surface. The

9

answer to this depends upon the specific condition of the pipeline in question, as appraised

10

in this Section. Such an understanding is not yet reflected in current pipeline design

11

practices [9, 25, 26].

12

3.3 Extracellular Polymer Substances (EPS)

13

The Extracellular Polymer Substances (EPS) are vital to a biofilm and thus contribute

14

significantly to biofilm- and biofouling- related problems within pipelines. These highly

15

hydrated, “slime-like” substances, in which the biofilm cells are embedded, are mostly

16

produced and secreted by the cells themselves [84, 85]. They are essentially the “glue”

17

which holds the microorganisms and biological matter together, and to the surface. The

18

stronger these bonds, the stronger the structural integrity of the biofilm, both internally and

19

externally [86].

20

phosphate) and inorganic (e.g. iron and manganese) materials, and can protect the

21

individual components of the biofilm from the negative influence of their surrounding

22

environment, such as flow shear and residual disinfectants [48, 62, 85, 87-90]. For example,

23

it has been documented that even a relatively high shear stress (≥ 3 N/m2) caused by a

24

flushing event, did not completely remove a biofilm from the pipe wall [5].

The EPS can also trap free-floating organic (e.g. carbon, nitrogen and

25

The EPS also contributes to many of the commonly associated properties and

26

characteristics of biofilms and biofouled surfaces, such as their: i) “slimy” or gelatinous

23


1

appearance, ii) visco-elastic and filamentous nature, and iii) adsorptive nature [15, 17].

2

Therefore, EPS contributes significantly to the additional energy dissipation mechanisms

3

associated with biofouled surfaces and the inherent difficulties of their frictional evaluation.

4

The higher the EPS fraction of the biofilm biomass, the more visco-elastic the biofilm layer

5

becomes and therefore the more energy it can potentially remove from the flow, leading to a

6

higher effective roughness [17].

7

The EPS matrix can account for over 90% of the dry biomass [85] and commonly

8

consists of a variety biopolymers, including polysaccharides, proteins, nucleic acids, lipids

9

and extracellular DNA [85, 91, 92]. However, the exact proportions of each are highly

10

variable both in space and time, as they are influenced significantly by environmental

11

conditions. The type of the microbial community present also influences the overall EPS

12

composition ratios. Environmental factors, such as flow shear stress and nutrient content, in

13

addition to the biofilm age and growth rates have been shown to influence the EPS

14

production rate and exact composition, both directly and indirectly through changing

15

microbial community structure [5, 93-95]. It has also been reported that the slower the biofilm

16

growth rate, the more energy is available for EPS production [96]. Furthermore, as biofilm

17

growth rate is greatly influenced by nutrient availability, so also is EPS content and

18

production rate [97]. For example, the introduction of additional phosphates (commonly used

19

to prevent corrosion within DWDSs) of the order of 3 µg/l to 300 µg/l was found to promote

20

biofilm growth (both in terms of thickness and coverage) and inhibit EPS production [98].

21

Similarly, it has been documented that in wastewater treatment processes

22

phosphate levels caused increased carbohydrate levels in the EPS [99]. Phosphorus

23

released from plastics such as polyethylene (PE) was also documented to promote growth,

24

although EPS production was not monitored [79].

reduced

25

Flow shear has also been documented to impact upon the structure and composition

26

of the EPS matrix and thus the biofilm itself. Generally, high shear and turbulent conditions

27

favour the production of more dense and compact biofilms, as such conditions encourage 24


1

EPS production and higher cohesiveness [100]. Consequently, under low shear conditions

2

the resultant biofilm is likely to be less stable. Biofilm growth under favourable EPS

3

production conditions will have a high resistance to external detachment forces [60].

4

Consequently, when EPS is inhibited the resultant biofilm (although, maybe thicker) is likely

5

to be less stable and more susceptible to flow shear and residual disinfectant. The flow

6

conditions in DWDs and DNs are likely to vary, and therefore, the cohesive forces of the

7

EPS will very too.

8

Polysaccharides and in some cases proteins, are generally reported as the

9

predominant component of the EPS matrix, representing over 50% of the overall EPS

10

fraction [85, 91]. The mechanical stability of the EPS matrix comes from a multitude of

11

relatively weak physicochemical forces, the majority of which are supplied by the

12

polysaccharides due to their filamentous nature [101]. Polysaccharide concentration can

13

therefore be considered a useful measure of overall stability and resilience. However, as

14

EPS composition also depends on microbial community structure, and thus differing among

15

discrete biofilms [102], a more robust EPS quantification should incorporate multiple EPS

16

aspects, in addition to polysaccharide concentration.

17 18

4. Quantifying Pipeline Hydraulic Efficiency

19

4.1 Traditional Approach

20

Historically, the most widely employed relationships for assessing the frictional effects of

21

Newtonian liquids flowing within a pipeline are that of the Darcy-Weisbach equation [103], C-

22

W equation [7] and the Moody Diagram [104]. These relationships are based on empirical

23

information and geometric similarity considerations for the hydraulically smooth, transitional

24

and fully rough turbulent flow regimes. Surface roughness characteristics such as height,

25

orientation, geometric arrangement and spacing are defined globally within the

26

aforementioned relationships by one-dimensional roughness scales, namely ks and n. It is 25


1

also common practice to employ the Hazen-Williams frictional relationship and roughness

2

coefficient, despite its documented limitations, particularly notable when simulating large-

3

diameter pipelines [105, 106]. The C-W equation can better represent a wider range of pipe

4

diameters and flow conditions. The C-W equation is given by:

1 √𝜆

= -2 log (

𝑘𝑠 2.5 + ) 3.7𝐷 𝑅𝑒√𝜆

(1)

5 6

Where λ is the friction factor and D is the internal pipe diameter. The Darcy-Weisbach

7

equation is given by:

𝐻𝑓 =

̅ 𝜆𝐿𝑈 2𝑔𝐷

(2)

8 9

Where Ū is the average freestream flow velocity, L is the pipe length and g is the

10

acceleration due to gravity.

11

4.2 Accounting for Biofouling

12

The traditional approach of adopting the C-W equation to simulate pipeline hydraulics has

13

been proven to be inadequate in evaluating the frictional resistance of biofouled pipelines

14

[15, 18, 21, 29]. However, under certain situations this is proven not to be the case, namely

15

at the polar extremes of the Moody diagram (i.e. very low and very high flow), and traditional

16

approaches are valid irrespective of the presence of a biofilm. For example, Lambert et al.

17

(2009) documented that a 25 mm diameter biofouled pipe followed a smooth pipe law

18

frictional relationship at Re < 5000 (equating to a Ū ≈ 0.2m/s). This is attributable to the

19

larger boundary layer associated with such conditions and thus the onset of hydraulically

20

smooth flow. Similarly at the other extreme, high detachment inducing shear forces are likely

21

to limit the extent of biofilm growth. However, such situations are generally uncommon within

26


1

most DNs and DWDSs. The application, therefore, of traditional practices in most cases can

2

lead to under- or over- estimated pipeline flow capacities, which will result in unforeseen

3

efficiency issues. For example, if the flow capacity of a pipeline is underestimated, it may fail

4

to achieve the design velocity required for self-cleaning and the likelihood of future fouling

5

and fouling issues will increase. Furthermore, in such an instance, the pipeline would have

6

been oversized and this could add unnecessarily to the cost and environmental impact (i.e.

7

in terms of the carbon footprint) of the project, due to additional pipe material and ground

8

excavations required.

9

Lambert et al. [29] used experimental observations to obtain a modified C-W

10

equation, which is aimed at addressing the inadequacy of the original equation for relating

11

frictional resistance and equivalent roughness for biofouled pipes. The modified C-W

12

equation is given by:

1 √𝜆

=−

1 √8𝜅

In (

𝑘𝑠 2.5 + ) 0.85𝐷 𝑅𝑒√𝜆

(3)

13 14

where κ is a modified von Kármán constant applicable to biofouled pipes, which was found

15

to be lower than the conventional value of 0.421 ± 0.002 [107] and a function of Re such as:

𝜅 = 1.00 × 10−6 𝑅𝑒 + 0.26

(4)

16 17 18

A more practical form of equation 3 is equation 5, which combines the modified C-W equation of Lambert et al. [29] with the Darcy-Weisbach equation (equation 2), such as:

̅ = 𝑈

𝜅√𝑔𝐷𝑆𝑓 2

In (

𝑘𝑠 2.5𝜐 + ) 0.85𝐷 𝐷√2𝑔𝐷𝑆𝑓

19

27

(5)


1

Figure 7 illustrates the detrimental impact of applying the traditional C-W equation as

2

opposed to the modified C-W equation proposed by Lambert et al. [29] for biofouled pipes, in

3

terms of flow rate estimations. For example, the flow rate estimated for a 400 mm internal

4

diameter pipe with ks = 0.6 mm, using equation 1 and 3 are 0.19 m3/s and 0.09 m3/s,

5

respectively. There is considerable disparity between these two estimates, 56% in this

6

instance. This highlights the potential error that could arise through the application of the

7

traditional C-W equation for biofouled pipes.

8 9 10 11 12

Figure 7. Percentage reduction in flow rate, Q from the application of the original Colebrook-White (CW) equation (equation 1) to the modified C-W equation (equation 3) for a range of pipe diameters from 400 mm to 1000 mm, each flowing full and with a pipe invert slope of 1:150. The flow within them was assumed to be uniform and hence Sf = invert slope.

13 14

Equation (5) is recommended in principle for use in simulating pipelines at the

15

pseudo-equilibrium biofouling stage, in that it presupposes the use of a constant ks value. It

16

was recently shown that the modified equation presented by Lambert et al. [29] had good

17

correlation with experimental results obtained below the critical shearing velocity of 1.77 m/s

18

[27]. It should be noted, however, that Lambert et al. [29] assessed a relatively small range

19

of environmental and hydrodynamic conditions. For example, only three different mean

28


1

velocity values, namely 1.15 m/s, 0.89 m/s, and 0.22 m/s were assessed. Furthermore, only

2

two very unique water sources were employed (Myponga Reservoir and River Murray water,

3

South Australia), which again limits the broader application of the equations, particularly the

4

derived κ relationships. Since the roughness characteristics of biofilms are highly dependent

5

upon the conditions they are subjected to, further experimentation is required to attest to the

6

validity or obtain a refined equation for use under a range of environmental conditions and

7

flow regimes [29]. Other predictive equations have been found in the literature for specific

8

situations [32] or based upon results with considerable scatter [56]. As such, they should

9

also be used with caution.

10

4.3 Gaps in the Quantification of Unsteady Effects

11

Under relatively constant operational conditions in a pipeline with a mature biofilm, i.e. one

12

that has reached the pseudo-equilibrium stage of development, the use of a constant

13

roughness scale value in equation (5) may represent well the actual conditions, provided that

14

the correct ks value is used. For example, Andrewartha [75] found that the frictional

15

behaviour of a hydropower channel covered by freshwater low-form gelatinous biofilms

16

supported the rigid wall similarity hypothesis normally used in pipeline modelling studies.

17

However, errors can arise from applying a generic global ks value for biofouled pipes, such

18

as ks = 0.6 mm or 1.5 mm, as also recommended in practical guidelines [9], without further

19

verification of the actual conditions. In particularly, these guidelines were derived from a

20

seemingly limited data sets and based on work carried out between 1966 and 1979.

21

Considerable advances have since been made within the industry, especially with regards to

22

the use of different pipe materials. All such advances will be considerably weakened by the

23

use of these out-dated design guidelines.

24

The traditional approach of using a constant roughness scale value in a one-

25

dimensional hydraulics model has been found to fail under highly unsteady conditions [10,

26

15, 18, 21, 32]. This is typically due to the biofilm’s vibrating and oscillating behaviour [10, 12,

29


1

15, 16, 20, 29, 30, 75]. For example, the effective roughness scale of a thin low-form

2

gelatinous biofilm has been found to be up to three times higher under normal flow

3

conditions than its dimensions would initially suggest, due to vibration-induced drag and

4

pressure loss [15]. In the case of filamentous biofilms (or streamers), the resultant increase

5

in drag is a function of the resonant/oscillating frequency of the streamer and is, therefore,

6

governed by its length and diameter, as well as the flow velocity [12, 16]. The effect of

7

filamentous formation on effective roughness can be significant.

8

Furthermore, flow induced biofilm oscillation and vibration behaviour will result in

9

temporal fluctuations onto the structure of the boundary layer and cause a phenomenon

10

known as vortex shedding

11

phenomena allows for a time-averaged analysis approach of the net effects of processes on

12

the energy dissipation in a biofouling pipeline – although such an approach has not been

13

reported in the literature. Typical assessments of flow-related impacts on pipeline biofouling

14

found in the literature involved the mean flow speed and turbulence intensity, as inferred

15

upon based on Re [10, 15, 27, 29, 108].

[12, 16]. In theory, the periodical nature of such fluctuating

16

Biofilms have also been documented to compress themselves under pressure, which

17

tends to confer an increased ability to resist the effects of flow shear [5, 67]. Such effects

18

have been shown to occur even when the surface is classified as hydraulically smooth [16].

19

In addition, significantly increased turbulent parameters (namely turbulence intensities and

20

Reynolds stresses) within the outer region of the boundary layer have also been associated

21

with the occurrence of biofouling, further indicating the potential to cause large-scale motion

22

and pressure losses [13, 14, 75]. It follows that even if a given ks value is representative of

23

space averaged conditions, it cannot include the dynamic temporal effects of biofouling

24

which can occur due to biofilm growth and/or varying operational and environmental

25

configurations [12, 15, 56]. As such, factors, either individually or cumulatively, are likely to

26

cause significant changes both in space and time to all aspects of the biofouled surface,

27

including its frictional characteristics (i.e. over the length and operational life of the pipeline 30


1

and from pipeline to pipeline). This was observed within operational pipelines by Lauchlan et

2

al. [56], who reported varying ks values for different pipelines operating under similar

3

conditions and which were in some cases, different by over an order of magnitude.

4

4.4 Dynamic ks Formulations – A Way Forward?

5

The fundamental lack of comprehensive information and data on the topic of biofouling

6

within DNs and DWDSs (i.e. over a full range of operational and environmental conditions),

7

means the task of improving design practice to incorporate biofouling effects at present, is a

8

challenging if not an impossible one. However, suggestion and recommendation on how to

9

move forward can be made based around the topics reviewed within this paper.

10

If an equation such as equation (5) is used to predict the hydraulic performance of a

11

pipeline operated under highly unsteady conditions, then the variation of ks with time should

12

be taken into account by using a separate formulation, namely a dynamic ks approach. One

13

way to achieve this would be describing the variation of such a parameter with respect to

14

multiple environmental and operational factors. Due to the considerable complexity of this

15

task, an indirect approach may be preferred, in which the unknowns are the steady state ks

16

value and the trend of variation of ks with respect to time. This has been achieved in

17

sediment transport applications involving bedform development [109, 110]. If the impact of

18

biofilm development on flow resistance variation can be represented in such a way, then a

19

predictive tool might become available to aid the design and operation control of pipelines,

20

which could be used, for example, to determine the frequency of cleaning interventions.

21

Such time-varying biofilm development models are presented within the literature [46, 60,

22

111], although only Sharp and Model [111] related the growth to a time-varying equivalent

23

roughness-scale [112]. The proposed model uses a growth rate constant expressed in

24

mm/year to represent the combined impact of biofilm development, internal corrosions and

25

tuberculation (in terms Hazen-Williams coefficient) over time.

31


1

Further studies on this topic centred on DWDSs and DNs could lead to the

2

refinement of existing formulae and/or the development of new ones. Particularly, as the

3

current prevailing understanding of the biofilm dynamics and flow interaction is

4

predominantly based on observation within hydropower [12, 15, 27, 108, 113, 114] and

5

marine [18, 21] applications, which have inherently different conditions to those typical within

6

DWDSs and DNs.

7

Nonetheless, design practices could benefit from the inclusion of a calibrated

8

dynamic roughness computation routine in the estimation of pipeline carrying capacity over

9

its lifecycle or between cleaning operations. By doing this, Equation (5) would be used to

10

calculate a time series of mean flow velocity and, thus, discharge values for the pipeline

11

during the period considered. This could lead to more realistic design and operation planning

12

measures being developed and contrFloibute to optimise operation.

13 14

5. Conclusion

15

This paper has provided an overview of the current understanding of biofilms and biofouling

16

in pipelines used within DWDSs and DNs. The review has shown that at present the

17

literature is fundamentally sparse and lacking in assessment of the key interacting processes

18

under a wide range of conditions. Moreover, the observed gaps in scientific literature are

19

reflected within the industry, whereby biofouling is not independently acknowledged in its

20

own right within current design practices. In addition, current design practice to deal with the

21

biofouling and its impact on pipeline hydraulics has little theoretical basis and is geared more

22

towards immediate cost savings, rather than longer term improvements in efficiency and the

23

benefits that would result. Therefore, in order to comprehensively address this problem, the

24

following recommendations are made with respect to further research and investigation:

25

1) How to treat biofilms and biofouled surface as an engineering roughness. The

26

diverse nature of biofilms means that generic solutions may not always be possible 32


1

and a case by case or site specific technical solution might be necessary. Hence, it

2

would appear that the most suitable approach would be to employ a dynamic

3

roughness scale that would be capable of qualifying both the space and time

4

averaged conditions of the fouled surface over a range of environment and

5

operational scenarios.

6

2) How to encourage the general engineering community to better consider the problem

7

of biofilms and biofouling in their day to day work, particularly within the water and

8

wastewater industries.

9

Further research is therefore, essential to better understand and evaluate the true

10

nature of biofouling, and its inevitable impact on pipeline flow resistance and capacity. More

11

advanced mathematical modelling frameworks that can be used to predict critical efficiency

12

losses would include adequate representations of the dynamic and case-specific nature of

13

biofouling. Such a framework could give rise to a real time monitoring platform to assist the

14

adoption of more cost-effective approaches to maintenance and repairs. In addition to more

15

accurately predicting the occurrence of biofouling, the environmental, construction and

16

operational conditions deemed to significantly prevent biofouling should be fostered, where

17

possible. For example, the use of pipes with smooth internal surfaces and operational flow

18

velocities of 0.9 m/s or higher.

19 20

Acknowledgments

21

The authors would like to acknowledge the support of the UK Engineering and Physical

22

Sciences Research Council (EPSRC) and Asset International Limited, particularly Dr

23

Vasilios Samaras for their support. Professor Binliang Lin of Tsinghua University, China is

24

also gratefully acknowledged for his contributions.

25

33


1 2

References [1]

3 4

Microbiol, 1995. 49: p. 711-45. [2]

5 6

Costerton, J.W. and Z. Lewandowski, Microbial biofilms. Annual Reviews

Batté, M., et al., Biofilms in drinking water distribution systems. Reviews in Environmental Science and Biotechnology, 2003. 2(2): p. 147-168.

[3]

LeChevallier, M.W., T.M. Babcock, and R.G. Lee, Examination and

7

characterization of distribution system biofilms. Applied and environmental

8

microbiology, 1987. 53(12): p. 2714-2724.

9

[4]

10 11

Santo Domingo, J.W., et al., Molecular survey of concrete sewer biofilm microbial communities. Biofouling, 2011. 27(9): p. 993-1001.

[5]

Douterelo, I., R.L. Sharpe, and J.B. Boxall, Influence of hydraulic regimes on

12

bacterial community structure and composition in an experimental drinking

13

water distribution system. Water Research, 2013. 47(2): p. 503-516.

14

[6]

15 16

Shockling, M.A., J.J. Allen, and A.J. Smits, Roughness effects in turbulent pipe flow. Journal of Fluid Mechanics, 2006. 564: p. 267-285.

[7]

Colebrook, C.F. and C.M. White, Experiments with fluid friction in roughened

17

pipes. Proceedings of the Royal Society of London. Series A, Mathematical

18

and Physical Sciences, 1937. 161(906): p. 367-381.

19

[8]

20 21

1933. 1292(1): p. 1-62. [9]

22 23

Nikuradse, J., Laws of flow in rough pipes. Mechanical Engineering Papers

Wallingford, H.R. and D.I.H. Barr, Tables for the hydraulic design of pipes, sewers and channels. 1994: Thomas Telford.

[10]

Lambert, M.F., et al., Understanding the impact of biofilm growth on pipe

24

roughness, in World Environmental and Water Resources Congress 2008.

25

2008, ASCE: Honolulu, Hawaii. p. 1-10.

34


1

[11]

Barton, A.F., et al., Hydraulic roughness of biofouled pipes, biofilm character,

2

and measured improvements from cleaning. Journal of Hydraulic Engineering,

3

2008. 134: p. 852.

4

[12]

Andrewartha, J.M., J.E. Sargison, and K.J. Perkins, The influence of

5

freshwater biofilms on drag in hydroelectric power schemes. WSEAS

6

Transactions on Fluid Mechanics, 2008. 3(3): p. 201-206.

7

[13]

Schwartz, T., S. Hoffmann, and U. Obst, Formation of natural biofilms during

8

chlorine dioxide and u.v. disinfection in a public drinking water distribution

9

system. Journal of applied microbiology, 2003. 95(3): p. 591-601.

10

[14]

Schwartz, T., S. Hoffmann, and U. Obst, Formation and bacterial composition

11

of young, natural biofilms obtained from public bank-filtered drinking water

12

systems. Water Research, 1998. 32(9): p. 2787-2797.

13

[15]

Barton, A.F., Friction, roughness and boundary layer characteristics of

14

freshwater biofilms in hydraulic conduits in School of Engineering. Hobart,

15

University of Tasmania. 2006, University of Tasmania.

16

[16]

Stoodley, P., et al., Oscillation characteristics of biofilm streamers in turbulent

17

flowing water as related to drag and pressure drop. Biotechnology and

18

Bioengineering, 1998. 57(5): p. 536-544.

19

[17]

20 21

Picologlou, B.F., N. Zelver, and W.G. Charcklis, Biofilm growth and hydraulic performance Journal of Hydraulics Division, 1980. 106(5): p. 733-746

[18]

Schultz, M.P., Turbulent boundary layers on surfaces covered with

22

filamentous algae. Transactions-american society of mechanical engineers

23

journal of fluids engineering, 2000. 122(2): p. 357-363.

24

[19]

Characklis, W.G., Bioengineering report: Fouling biofilm development: A

25

process analysis. Biotechnology and Bioengineering, 1981. 23(9): p. 1923-

26

1960.

27 28

[20]

Characklis, W.G., Attached microbial growths—II. Frictional resistance due to microbial slimes. Water Research, 1973. 7(9): p. 1249-1258. 35


1

[21]

2 3

layers. Journal of Fluids engineering, 1999. 121: p. 44. [22]

4 5

Minkus, A.J., Deterioration of the hydraulic capacity of pipelines. New England Water Works Association, 1954. 68(1): p. 1-10.

[23]

6 7

Schultz, M.P. and G.W. Swain, The effect of biofilms on turbulent boundary

Seifert, L. and W. Kruger, Unusually high friction factor in a long, water supply line. VDI Z, 1950. 92: p. 189-191.

[24]

Sharp, B.B., Examination of the friction in the Morgan-Whyalla pipeline.

8

Prepared for Engineering & Water Supply Department, Adelaide, Australia,

9

1954.

10

[25]

11 12

outside buildings. 2008: London: BSI. [26]

13 14

British Standards Institution, BS EN 752:2008, in Drain and sewer systems

Water Research Centre (Great Britain), Sewers for adoption. 6th ed. 2006, Swindon: WRc plc.

[27]

Perkins, S., et al., The influence of bacteria based biofouling on the wall

15

friction and velocity distribution of hydropower pipes, in 18th Australasian

16

Fluid Mechanics Conference 2012: Launceston, Tasmania.

17

[28]

Barton, A.F., et al. Deterioration of conduit efficiency due to biofouling. in 8th

18

National Conference on Hydraulics in Water Engineering. 2004. Queensland,

19

Australia.

20

[29]

Lambert, M.F., et al., The impact of biofilm development on pipe roughness

21

and velocity profile, in World Environmental and Water Resources Congress

22

2009. 2009, ASCE: Kansas City, Missouri.

23

[30]

24 25 26

Stoodley, P., et al., Influence of hydrodynamics and nutrients on biofilm structure. Journal of applied microbiology, 1998. 85(S1): p. 19S-28S.

[31]

Manning, R., On the Flow of Water in Open Channels and Pipes. Instn of Civil Engineers of Ireland, 1890.

36


1

[32]

Guzman, K., et al., Effect of biofilm formation on roughness coefficient and

2

solids

3

Environmental Engineering, 2007. 133(4): p. 364-371.

4

[33]

5 6

deposition

in

small-diameter

PVC

sewer

pipes.

Journal

of

Vignaga, E., The effect of biofilm colonization on the stability of non-cohesive sediments. 2012, University of Glasgow.

[34]

Chughtai, F. and T. Zayed, Infrastructure condition prediction models for

7

sustainable sewer pipelines. Journal of Performance of Constructed Facilities,

8

2008. 22(5): p. 333-341.

9

[35]

Shirazi, R.H.S.M., et al., Modelling the erosive effects of sewer flushing using

10

different sediment transport formulas, in 7th International Conference on

11

Sewer Processes and Networks, SPN7. 2013: Sheffield, UK.

12

[36]

Creaco, E. and J.-L. Bertrand-Krajewski, Numerical simulation of flushing

13

effect on sewer sediments and comparison of four sediment transport

14

formulas. Journal of Hydraulic Research, 2009. 47(2): p. 195-202.

15

[37]

16 17

odor and corrosion in sewers. Water Research, 2008. 42(15): p. 4206-4214. [38]

18 19

Nielsen, A.H., et al., Influence of pipe material and surfaces on sulfide related

Guisasola, A., et al., Methane formation in sewer systems. Water Research, 2008. 42(6): p. 1421-1430.

[39]

Momba, M.N.B., et al., An overview of biofilm formation in distribution

20

systems and its impact on the deterioration of water quality. Water SA 2000.

21

26: p. 59-66.

22

[40]

Prévost, M., et al., Suspended bacterial biomass and activity in full-scale

23

drinking water distribution systems: impact of water treatment. Water

24

Research, 1998. 32(5): p. 1393-1406.

25

[41]

Husband, P.S., J.B. Boxall, and A.J. Saul, Laboratory studies investigating

26

the processes leading to discolouration in water distribution networks. Water

27

Research, 2008. 42(16): p. 4309-4318.

37


1

[42]

2 3

pilot-scale distribution systems. 2003. [43]

4 5

Schlichting, H., Boundary Layer Theory New York. 1979, New York.: McGraw-Hill.

[44]

6 7

Eisnor, J. and G. Gagnon, A framework for the implementation and design of

Chamberlain, A.H.L., The role of adsorbed layers in bacterial adhesion. Biofilmsscience and technology. Dordrecht: Kluwer, 1992: p. 59-67.

[45]

Vigeant, M.A.S., et al., Reversible and irreversible adhesion of motile

8

Escherichia coli cells analyzed by total internal reflection aqueous

9

fluorescence microscopy. Applied and environmental microbiology, 2002.

10 11

68(6): p. 2794-2801. [46]

12 13

and Fluid Science, 1997. 14(4): p. 375-381. [47]

14 15

[48]

[49]

[50]

Stewart, P.S., Mini-review: convection around biofilms. Biofouling, 2012. 28(2): p. 187-198.

[51]

22 23

Kjelleberg, S. and M. Givskov, The biofilm mode of life: mechanisms and adaptations. 2007: The Publisher.

20 21

Boe-Hansen, R., et al., Bulk water phase and biofilm growth in drinking water at low nutrient conditions. Water Research, 2002. 36(18): p. 4477-4486.

18 19

Hallam, N.B., et al., The potential for biofilm growth in water distribution systems. Water Research, 2001. 35(17): p. 4063-4071.

16 17

Melo, L.F. and T.R. Bott, Biofouling in water systems. Experimental Thermal

Cloete, T.E., L. Jacobs, and V.S. Brözel, The chemical control of biofouling in industrial water systems. Biodegradation, 1998. 9(1): p. 23-37.

[52]

Tsvetanova, Z., Study of biofilm formation on different pipe materials in a

24

model of drinking water distribution system and its impact on microbiological

25

water quality. Chemicals as Intentional and Accidental Global Environmental

26

Threats, 2006: p. 463-468.

38


1

[53]

Zhou, L., Y. Zhang, and G. Li, Effect of pipe material and low level

2

disinfectants on biofilm development in a simulated drinking water distribution

3

system. Journal of Zhejiang University-Science A, 2009. 10(5): p. 725-731.

4

[54]

5 6

drinking water. Water Research, 1990. 24(2): p. 239-243. [55]

7 8

Cloete, T.E., D. Westaard, and S.J. van Vuuren, Dynamic response of biofilm to pipe surface and fluid velocity. Water Sci Technol, 2003. 47(5): p. 57-9.

[56]

9 10

Pedersen, K., Biofilm development on stainless steel and PVC surfaces in

Lauchlan, C., J. Forty, and R. May, Flow resistance of wastewater pumping mains. Water management, 2005. 158(2): p. 81-88.

[57]

Gjaltema, A., et al., Heterogeneity of biofilms in rotating annular reactors:

11

occurrence, structure, and consequences. Biotechnology and Bioengineering,

12

2004. 44(2): p. 194-204.

13

[58]

Van Der Kooij, D., H.S. Vrouwenvelder, and H.R. Veenendaal, Density and

14

composition of biofilms in drinking water distribution systems in the

15

Netherlands, in AWWA Water Quality Technology Conference 1995: New

16

Orleans, LA. p. 1055-1062.

17

[59]

Niquette, P., P. Servais, and R. Savoir, Impacts of pipe materials on densities

18

of fixed bacterial biomass in a drinking water distribution system. Water

19

Research, 2000. 34(6): p. 1952-1956.

20

[60]

21 22

in flow/non-flow conditions. Water Research, 2007. 41(3): p. 551-562. [61]

23 24

Perkins, J.A. and I.M. Gardiner, The hydraulic roughness of slimed sewers 1985. 79: p. 87-104.

[62]

25 26

Manuel, C.M., O.C. Nunes, and L.F. Melo, Dynamics of drinking water biofilm

Bridier, A., et al., Resistance of bacterial biofilms to disinfectants: a review. Biofouling, 2011. 27(9): p. 1017-1032.

[63]

Andrewartha, J.M. and J. Sargison, Turbulence and mean-velocity structure of

27

flows over filamentous biofilms, in 34th IAHR World Congress. 2011: Brisbane,

28

Australia. p. 2225-2232. 39


1

[64]

Lewandowski, Z. and P. Stoodley, Flow induced vibrations, drag force, and

2

pressure drop in conduits covered with biofilm. Water science and technology,

3

1995. 32(8): p. 19-26.

4

[65]

De Beer, D., et al., Effects of biofilm structures on oxygen distribution and

5

mass transport. Biotechnology and Bioengineering, 1994. 43(11): p. 1131-

6

1138.

7

[66]

Santos, R., M.E. Callow, and T.R. Bott, The structure of Pseudomonas

8

fluorescens biofilms in contact with flowing systems. Biofouling, 1991. 4(4): p.

9

319-336.

10

[67]

Percival, S.L., et al., The effect of turbulent flow and surface roughness on

11

biofilm formation in drinking water. Journal of industrial microbiology &

12

biotechnology, 1999. 22(3): p. 152-159.

13

[68]

14 15

1995. 38(4): p. 39. [69]

16 17

Dumbleton, B., A question of scale and slime. Water and Waste Treatment,

Nowell, A.R.M. and M. Church, Turbulent flow in a depth-limited boundary layer. Journal of Geophysical research, 1979. 84(C8): p. 4816-4824.

[70]

Kerr, C.J., et al., The relationship between pipe material and biofilm formation

18

in a laboratory model system. The Society for Applied Microbiology, 1999. 85:

19

p. 29S-38S.

20

[71]

Momba, M.N.B. and N. Makala, Comparing the effect of various pipe materials

21

on biofilm formation in chlorinated and combined chlorine-chloraminated water

22

systems. Water SA, 2004. 30(2): p. 175-182.

23

[72]

24 25

Kurth, J.C., Mitigating biofilm growth through the modification of concrete design and practice. 2008.

[73]

Verran, J., G. Lees, and A.P. Shakespeare, The effect of surface roughness

26

on the adhesion of Candida albicans to acrylic. Biofouling, 1991. 3(3): p. 183-

27

191.

40


1

[74]

2 3

Yu, J., D. Kim, and T. Lee, Microbial diversity in biofilms on water distribution pipes of different materials. Water science and technology, 2010. 11(1): p. 163.

[75]

Andrewartha, J.M., The effect of freshwater biofilms on turbulent boundary

4

layers and the implications for hydropower canals. 2010, University of

5

Tasmania.

6

[76]

Sand,

W.

and

E.

Bock,

Biodeterioration

of

mineral

materials

by

7

microorganisms—biogenic sulfuric and nitric acid corrosion of concrete and

8

natural stone. Geomicrobiology Journal, 1991. 9(2-3): p. 129-138.

9

[77]

10 11

Gaylarde, C.C. and L.H.G. Morton, Deteriogenic biofilms on buildings and their control: A review. Biofouling, 1999. 14(1): p. 59-74.

[78]

Nielsen, A.H., et al., Simulation of sulfide buildup in wastewater and

12

atmosphere of sewer networks. Water science and technology, 2005: p. 201-

13

208.

14

[79]

Lehtola, M.J., et al., Microbiology, chemistry and biofilm development in a pilot

15

drinking water distriction system with copper and plastic pipes Water

16

Research, 2004. 38(17): p. 3769-3779.

17

[80]

18 19

Percival, S.L., et al., Biofilm development on stainless steel in mains water. Water Research, 1998. 32(1): p. 243-253.

[81]

Abdel-Monim, Y.K., S.A. Ead, and S.A. Shabayek, Effect of time on pipe

20

roughness, in 17th Chandian Hydrotechnical Conference 2005: Edmonton,

21

Alberta, Canada.

22

[82]

Bland, C.E.G., R.W. Bayley, and E.V. Thomas, Some observations on the

23

accumulation of slime in drainage pipes the effect of these accumulations on

24

resistance to flow Public Health Engineering, 1975. 13.

25

[83]

Bott, T.R. and P.C. Miller, Mechanisms of biofilm formation on aluminium

26

tubes. Journal of Chemical Technology and Biotechnology. Biotechnology,

27

1983. 33(3): p. 177-184.

41


1

[84]

Flemming, H.C., Biofouling in water systems - cases, causes and

2

countermeasures Applied Microbiology and Biotechnology 2002. 59(6): p.

3

629-640.

4

[85]

5 6

Flemming, H.-C. and J. Wingender, The biofilm matrix. Nat Rev Micro, 2010. 8(9): p. 623-633.

[86]

Kalmokoff, M.L., et al., Adsorption, attachment and biofilm formation among

7

isolates of Listeria monocytogenes using model conditions. Journal of applied

8

microbiology, 2001. 91(4): p. 725-734.

9

[87]

10 11

Srinivasan, R., et al., Biofilm parameters influencing biocide efficacy. Biotechnology and Bioengineering, 1995. 46(6): p. 553-560.

[88]

LeChevallier, M.W., C.D. Cawthon, and R.G. Lee, Factors promoting survival

12

of bacteria in chlorinated water supplies. Applied and environmental

13

microbiology, 1988. 54(3): p. 649.

14

[89]

Cochran, W.L., G.A. McFeters, and P.S. Stewart, Reduced susceptibility of

15

thin

16

monochloramine. Journal of applied microbiology, 2000. 88(1): p. 22-30.

17

[90]

18 19

Pseudomonas

aeruginosa

biofilms

to

hydrogen

peroxide

and

Simes, L.C., et al., Drinking water biofilm assessment of total and culturable bacteria under different operating conditions. Biofouling, 2006. 22(2): p. 91-99.

[91]

Horan, N.J. and C.R. Eccles, Purification and characterization of extracellular

20

polysaccharide from activated sludges. Water Research, 1986. 20(11): p.

21

1427-1432.

22

[92]

Goodwin, J.A.S. and C.F. Forster, A further examination into the composition

23

of activated sludge surfaces in relation to their settlement characteristics.

24

Water Research, 1985. 19(4): p. 527-533.

25

[93]

Qi, X., et al., Biofilm formation of the pathogens of fatal bacterial granuloma

26

after trauma: potential mechanism underlying the failure of traditional antibiotic

27

treatments. Scandinavian journal of infectious diseases, 2008. 40(3): p. 221-

28

228. 42


1

[94]

2 3

Prakash, B., B.M. Veeregowda, and G. Krishnappa, Biofilms: A survival strategy of bacteria. Current science, 2003. 85(9): p. 1299-1307.

[95]

Donlan, R.M. and J.W. Costerton, Biofilms: survival mechanisms of clinically

4

relevant microorganisms. Clinical microbiology reviews, 2002. 15(2): p. 167-

5

193.

6

[96]

Kreft, J.U. and J.W.T. Wimpenny, Effect of EPS on biofilm structure and

7

function as revealed by an individual-based model of biofilm growth. Water

8

science and technology, 2001. 43(6): p. 135-142.

9

[97]

10 11

Sutherland, I.W., Biofilm exopolysaccharides: a strong and sticky framework. Microbiology, 2001. 147(1): p. 3-9.

[98]

Fang, W., J.Y. Hu, and S.L. Ong, Influence of phosphorus on biofilm formation

12

in model drinking water distribution systems. Journal of applied microbiology,

13

2009. 106(4): p. 1328-1335.

14

[99]

Hoa, P.T., L. Nair, and C. Visvanathan, The effect of nutrients on extracellular

15

polymeric substance production and its influence on sludge properties. Water

16

SA, 2004. 29(4): p. 437-442.

17

[100]

Stoodley, P., et al., Biofilm material properties as related to shear-induced

18

deformation and detachment phenomena. Journal of Industrial Microbiology

19

and Biotechnology, 2002. 29(6): p. 361-367.

20

[101]

Wloka, M., et al., Rheological properties of viscoelastic biofilm extracellular

21

polymeric substances and comparison to the behavior of calcium alginate gels.

22

Colloid and Polymer Science, 2004. 282(10): p. 1067-1076.

23

[102]

24 25

quantitative biofilm studies. Annals of Microbiology, 2011: p. 1-7. [103]

26 27 28

Molobela, I.P. and F.M. Ilunga, Impact of bacterial biofilms: the importance of

Darcy, H., Recherches expérimentales relatives au mouvement de l'eau dans les tuyaux. Vol. 1. 1857: Mallet-Bachelier.

[104]

Moody, L.F., Friction factors for pipe flow. Trans ASME 1944. 66(8): p. 671684. 43


1

[105]

Christensen, B.A., F.A. Locher, and P.K. Swamee, Limitations and Proper Use

2

of the Hazen-Williams Equation. Journal of Hydraulic Engineering, 2000.

3

126(2): p. 167-170.

4

[106]

Bennett, D. and R. Glaser. Common Pitfalls in Hydraulic Design of Large

5

Diameter Pipelines: Case Studies and Good Design Practice. in ASCE

6

Pipelines Conference 2011. 2011. Seattle, Washington

7

[107]

8 9

fully developed pipe flow Journal of Fluid Mechanics, 2004. 501: p. 135-147. [108]

10 11

McKeon, B.J., et al., Further observations on the mean velocity distribution in

Barton, A.F., et al., Characterizing the roughness of freshwater biofilms using a photogrammetric methodology. Biofouling, 2010. 26(4): p. 439-448.

[109]

Rauen, W.B., B. Lin, and R.A. Falconer, Modelling dynamic bed roughness

12

associated with bed form development, in 16th IAHR-APD Congress and 3rd

13

IAHR-ISHS Symposium. 2008: Nanjing, China.

14

[110]

Rauen, W.B., B. Lin, and R.A. Falconer, Modelling ripple development under

15

non-uniform flow and sediment supply-limited conditions in a laboratory flume.

16

Estuarine, Coastal and Shelf Science, 2009. 82(3): p. 452-460.

17

[111]

18 19

Sharp, W.W. and T.M. Walski, Predicting internal roughness in water mains. Journal (American Water Works Association), 1988: p. 34-40.

[112]

Filion, Y.R., H.L. MacLean, and B.W. Karney, Life-cycle energy analysis of a

20

water distribution system. Journal of infrastructure systems, 2004. 10(3): p.

21

120-130.

22

[113]

Andrewartha, J.M., J.E. Sargison, and K.J. Perkins, The Effect of

23

Gomphonema and Filamentous Algae Streamers on Hydroelectric Canal

24

Capacity and Turbulent Boundary Layer Structure, in 16th Australasian Fluid

25

Mechanics Conference (AFMC). 2007, School of Engineering, The University

26

of Queensland. p. 241-246.

27 28

[114]

Barton, A.F., et al., A Force balance to measure the total drag of biofilms on test plates in 16th Australasian Fluid Mechanics Conference (AFMC). 2007, 44


1

School of Engineering, The University of Queensland Gold Coast, Queensland,

2

Australia p. 819-824.

3

45