the characteristics of an in-sewer … · observations from environmentally controlled annular ......

Towards a Balanced Methodology in

European Hydraulic Research 16-1

16. THE CHARACTERISTICS OF AN IN-SEWER EROSION EVENT – OBSERVATIONS FROM ENVIRONMENTALLY CONTROLLED ANNULAR

FLUME

S J Tait1, R M Ashley2, R Verhoeven3, F Clemens4,R Sakrabani5 A Schellart4, R Veldkamp4, L Aanen6

1 Pennine Water Group, University of Sheffield, Department of Civil and Structural

Engineering, Mappin Street, Sheffield, S1 3JD,UK E-Mail: [email protected]; 2 Pennine Water Group, University of Bradford, UK 3 Hydraulics Laboratory, University of Gent, Belgium 4 Department of Civil Engineering, Technical University of Delft, Delft, Netherlands 5 National Soil Resources Institute, University of Cranfield 6 WL Delft Hydraulics, Delft, Netherlands

ABSTRACT

This paper provides an overview of a collaborative project recently completed at WL Delft Hydraulics. It describes tests in an annular flume, in which sewer sediment deposits could be formed under carefully controlled and monitored environmental conditions and then subjected to a test series in which the bed shear stress was progressively increased. The characteristics of the eroded sediment were then closely monitored. The sediment deposits were formed using two different types of sewer sediment. The deposits in the reported experiments were formed using real in-sewer sediments, from a catchment in the UK (Dundee) and from a catchment in the Netherlands (Loenen). During the erosion tests, total and volatile suspended solids concentration, particle size distribution of the eroded sediment, and COD and DO levels were all recorded. In addition, the biodegradability of the eroded sediments were also analysed using a respirometric method. This provided the most detailed data set covering the physical and biochemical aspects of an erosion event associated with in-sewer sediments ever recorded. The bed surface topography was also measured so that the influence of the deposit formation condition on bedforms could also be examined.

16.1 INTRODUCTION

Most sewer systems in the EU are combined, that is they can carry both foul wastewater and storm water. In many EU countries there have been no significant changes in sewer design and operation for decades and the existing urban drainage systems reflect design approaches that did not consider the impact of sewer systems on the wider environment. It is now realised that existing systems can have a significant impact on the environment. In particular the management of the transportation of solids within a sewer network is starting to be considered an important aspect of sewer operation and design. Traditionally, most designers sized their sewers based on a notional design storm so that for the majority of the time, when there is no storm sewer conduits are oversized. Sediment deposition is common due to the wide range of solid types (density and particle size) and the typical low flow rates, velocities and shear stresses that occur during dry weather in combined sewers. During periods of high rainfall, sewer systems cannot transport the rate and volume of runoff from the urban catchment to the treatment works and are therefore designed to discharge excess storm flows via overflows to natural watercourses. Pollutants, including solids, within these

Tait et al Budapest, 22-23 May 2003

mailto:[email protected];



discharges can have a significant environmental impact on the receiving waters. High suspended sediment concentrations have been commonly observed in these overflows during the initial periods of storms and have been termed “first foul flushes”. One of the main physical causes of these suspended sediment peaks is the erosion of existing in-sewer deposits.

A number of field studies have reported observing a “first foul flush”, however there has been no consensus on the definition of such an event. Thornton and Saul (1986) loosely defined the first flush as the initial period in a storm event in which the observed concentration of suspended sediments and other pollutants was significantly higher than observed in the preceding dry weather flow or in the later stages of the storm. Geiger (1987) examined entire storm events and proposed a definition of the first foul flush based on plotting the percentage cumulative pollutant load against the cumulative flow. This method was adopted, by Gupta and Saul (1996), with minor changes to define the “first foul flush”. Stotz and Krauth (1984) and Geiger (1987) both observed that in larger catchments the distinct pattern of the flush would be less. This asceryion is supported by the observations of Bertrand-Krajewski et al., (1993) and Saget et al. (1995) who concluded that the first foul flush mechanism was more prevalent in smaller catchments. More recent studies Gupta and Saul, (1996) have attempted to develop empirical site-specific relationships between the properties of the flush and the characteristics of the catchment, the storm event and the resultant sewer flow and sediment behaviour, but with limited success.

In summary it is clear from a number of field studies that in many sewer networks rapid increases in suspended sediment concentrations have occurred during the initial parts of storms. No widely applicable relationship has been identified from the data collected in these field studies. All reported empirical relationships have been found to be site specific. This suggests that the methodology of empirically linking catchment characteristics to rapid variations in sewer sediment concentrations and transport ignores important physical mechanisms. Therefore in an attempt to better predict the suspended sediment behaviour of sewer systems a more deterministic approach has also been adopted (Mark et al, 1995). Large computational sewer network models with sediment transport rate modules have been developed, e.g. MOUSE (DHI) and INFOWORKS (Wallingford Software) to simulate the hydraulic and sewer flow quality performance of sewer systems. In general these models can simulate accurately the time varying hydraulic performance of these complex networks providing suitable and sufficient field data is used to initially calibrate the model (Clemens, 2001). However, the ability of these commercial models to predict water quality changes is much more limited. The sediment transport rate modules employed have generally utilised transport rate relationships originally developed in fluvial environments. In-sewer deposits have been represented as homogeneous and generally granular, whereas in reality they comprise a wide range of sediment types (organic/inorganic) and particle sizes and can exhibit significant cohesive-like properties. This oversimplification and the use of inappropriate sediment transport relationships has led to poor simulation of suspended sediment transport and as a consequence low end-user confidence (Jack et al., 1996; Bouteligier et al., 2002). It is clear that a better understanding is required of how in-sewer sediments form deposits and how these deposits can develop strength and subsequently respond to periods of increasing bed shear stress observed during storm events.

Previous work that has examined the transport of sediments in pipes has concentrated on the movement of non-cohesive, uniformly sized sediment (Ackers, 1991; May, 1993; Nalluri and Ab Ghani, 1993), although some very recent studies have started to examine the movement of non-cohesive, sediment mixtures in pipes (DeSutter et al., 2003 Ota and Nalluri, 2002). Field studies have indicated that significant proportions of in-sewer deposits are highly homogeneous in nature, with ranges of particle size and density present (Crabtree, 1989). This variability introduces new types of physical phenomena such as fine particle and organic related cohesion, and the sheltering and exposure experienced in mixed grain size




sediment deposits. Most transport relationship used in engineering practice assume single sized granular sediment deposits (Ackers et al, 1996) and great care must be taken in the application of these relationships to ensure a proper understanding of the potential errors that are likely (Arthur et al, 1999). Crabtree (1989) also noted the importance of the organic sediment fraction within sewer deposits, he postulated that this fraction could be responsible for the development of “cohesive-like” forces between the particles. The highly organic fractions within the sampled sewer deposits also had a higher pollutant potential in terms of oxygen demand. Later work by Vollertsen and Hvitved-Jacobsen (2000) indicated that that biochemical processes associated with the organic fractions may have some impact of the erosional strength of the deposit. A small number of laboratory tests, using real sewer sediment deposits in a small pipe, indicated rapid reductions in dissolved oxygen in the overlying sewage appeared to be combined with a weakening of the in-pipe deposit. Unfortunately no detailed strength measurements were made. Recent work by Chebbo et al., (2003) has indicated that the surface layer on an in-sewer deposit can be highly organic in nature and that this material in this surface layer can have an important role in determining the in-sewer flow quality.

It is clear that for the large sewer network models to significantly improve their predictive performance with regard to the behaviour of solids within sewer networks then better algorithms are required which can simulate the deposition, erosion and subsequent transport of all types of sewer sediments that can be found in combined sewers. Major difficulties have been encountered in trying to develop understanding of the deposition, erosion and transport phenomena related to sewer sediment processes. Two groups of researchers have emerged; one group has attempted to identify surrogate sediments, whose characteristics would accurately simulate the behaviour of real in-sewer sediments, and then conduct a series of tests systematically changing hydraulic conditions to observe the transport potential of the sediments (e.g. Skipworth et al, 1999, DeSutter, 2000, Rushforth, 2001). The second group has maintained that only by the study of real in-sewer sediments in their environment one can observe and understand the physical processes which control the movement of in-sewer sediments (e.g. Arthur and Ashley, 1998, Ahyerre et al., 2001, Chebbo et al, 2003). The first type of study has been unable to demonstrate clearly that the surrogate sediment selected will adequately describe the behaviour of in-sewer sediments. Tests in laboratories have also tended to ignore the importance of the environmental conditions in which in-pipe deposits form. These are believed to have a significant affect on the way in which organic sediments can transform and thus influence deposit strength (Camuffo, 2001, Tait et al., 2003). Any results produced by these studies are heavily dependent on the initial conditions (type of surrogate sediment, mode of deposit formation). It is therefore difficult to apply the results obtained and relationships derived in a more general context. The second type of study has relied on observations of the behaviour of real sewer sediment in-situ. Although this type of study uses real sewer sediments, typically the hydraulic control is poor and there is limited opportunity for examining the repeatability of observations. Field studies are expensive in terms of resources and the collection of data physically problematic. This has led to field data sets being limited in the scope of hydraulic conditions and measured determinants. The lack of control also means that the conditions in which deposits form cannot be controlled and are rarely measured. The outputs from this work tend to be very site specific and in the form of only locally applicable empirical relationships.

Very few studies have been carried out in which complex sediment mixes were used. There have been studies in Canada to examine the erosional behaviour of contaminated estuarine sediments, with an organic content under controlled depositional conditions (Lau and Droppo, 2000). However no work has ever been carried out with in-sewer sediments in which both the deposit formation and subsequent erosion processes were studied under consistently controlled environmental conditions only possible in the laboratory.




16.2 EXPERIMENTAL WORK

Given the above difficulties, carefully controlled experiments in which the antecedent depositional environment as well as the hydraulic conditions during the erosion test are controlled is extremely rare in sewer sediment studies. This paper will report on a series of tests in which the conditions during which sediment deposits are formed is carefully controlled. This first or “depositional” part of the test was to simulate depositional conditions within a sewer pipe during dry weather flow. The second “erosion” part of the tests were carried out to simulate the conditions at the start of a storm event in which the deposited sewer sediment bed is subjected to increasing boundary shear stress but using a high level of control so that a systematic data set on the erosional behaviour can be obtained. The tests were carried out in an annular flume at WL Delft Hydraulics in the Netherlands in which the hydraulic conditions and many of the significant environmental parameters within the flume (DO and temperature) could be controlled. The ability to control the environment within the flume means that this facility has significant advantages over many other laboratory facilities. The annular flume was also sufficiently large so that the Reynolds Numbers that are generated are of the same order of magnitude as observed in real sewers. The circular nature of the facility causes the moving granular bedload to be “re-circulated” allowing the bedforms that develop to attain equilibrium. In this way the annular flume will represents a “long” length of sewer pipe in which a stable “natural” morphology can develop. The temperature controlled environment also allows for the separation of the processes caused by physical consolidation of the sediment deposits and by the biological transformation of the fine-grained organic sediment which are present in many in-sewer deposits. The only limitation when compared with real conditions is the retention of the same sediments and fluid within the fluid, rather than the continual refreshment of sewage and solids due to conveyed inputs from upstream.

16.2.1 Experimental Equipment The tests were carried out in an annular flume that had an external diameter of 2.2 metres

and an internal diameter of 2.0 metres ( ). The flume was located in a room in which the air temperature and humidity can be controlled. During these tests the air temperature was set to either 14°C or 4°C. These temperatures were selected so that the higher would be similar to the temperatures found in northern European sewers during the summer, the lower temperature was selected so as to significantly inhibit any biochemical processes. Both the top and bottom plates of the flume rotated, during the erosional tests. This was done so that the lateral flow circulations within the flume could be minimised and so ensure that as uniform a shear stress pattern as possible was applied to the sediment bed. Preliminary tests with 320µm uniformly sized sand indicated that in order to minimise lateral flow circulation the ratio of the top plate to bottom plate rotational speed should be 4.2. The flume was equipped with a number of systems that allowed the measurement of the bed shear stress, suspended sediment and dissolved oxygen concentration and temperature at a frequency of 2Hz throughout each test. A web camera was installed on the lower plate and viewed a vertical section of the bed. Images from the web camera were recorded at one-minute intervals during the depositional and erosional phases of each test. This gave a visual recorded image of what happened to the bed, at one location, and examination of the images proved useful in examining the physical mechanisms of particle dislodgement and erosion and in estimating the celerity of bedforms if they were present. The flume was also equipped with an automatic sampling device that would allow the recovery of discrete samples of 250ml at any time. The draw off point for the suspended sediment concentration devices, the DO probe and the discrete samples was located approximately 30mm above the sediment bed.

Figure 16-1




Figure 16-1. Side view (left) and cross-section (right) of annular flume and measurements systems at WL | Delft Hydraulics

16.2.1.1 Measurement of Boundary Shear Stress The boundary shear stress on the sediment bed was estimated by measuring the

displacement of the stainless steel annular container which contained the sediment bed (labelled A in ). The annular container was suspended by six sets of vertical wire ropes enclosed in hollow steel columns (an example is labelled B in ) located at 60° intervals around the circular flume. This arrangement ensured that the annular container did not make contact with either the base or Perspex walls of the annular flume. All the vertical load of the annular container and the sediment it contained was therefore supported by 6 sets of wire ropes. The base of the annular container was attached to the base of the annular flume by three sets of pre-tensioned steel wires, whose load deflection behaviour had been previously measured. The radial displacement of the annular sediment container could therefore be linked with the horizontal force being applied to the surface of the sediment bed by the moving fluid. Three cameras were fixed to the bottom plate of the annular flume and were configured to register the displacement of three black and yellow targets which were attached to the inside wall of the annular sediment container. These cameras provide a voltage output signal that is sampled at 2Hz. Initially it was hoped that by transforming these voltage signals using the empirically measured stiffness of the pre-tensioned steel wires it would be possible to measure the applied hydraulic force directly. However initial analysis of

Figure 16-2Figure 16-2




the camera signals showed extreme outliers, systematic deviations and noise. By relating each camera output against its camera specific offset value the systematic deviation between the cameras can be removed. The offset value for each camera is determined as the intercept at zero top plate velocity of a regression line of camera voltage against top plate velocity. The amplitude of the signal noise was seen to increase with the increasing top and bottom plate velocities. Outliers were also noted, these were attributed to small particles settling between the outside Perspex walls of the flume and the yellow and black targets located on the inner wall of the annular sediment container. Initially the outliers and signal noise were not addressed but the range of calculated shear stress was too great. It was clear that methods were needed to process the data in order to successfully calculate the time-averaged bed shear stress. A common way to achieve noise reduction is to apply a numerical filter to the data. A straightforward method is to remove the outliers and use a linear filter based on a moving average approach (Chatfield, 1999). However this did not prove successful. A more sophisticated approach is to transfer the signal into the frequency domain and then remove the higher frequency noise. This was found, by studying power density spectrum, to be similar to the frequency of the top plate of the flume. Further details of the methods used are in Schellart, (2002); Schellart et al., (2003). This data and processing methods allowed the bed shear stress in each part of the erosion test to be estimated.

A

B

Figure 16-2. Photograph of annular flume at end of erosion section of test 8. Water in flume is a very dark colour due to significant erosion.

16.2.1.2 Measurement of Sediment Biodegradability Oxygen utilisation rate (OUR) measurements were carried out on three samples of eroded

sediment obtained from the annular flume during each erosion test. OUR measurement is a respirometric method where the eroded sediment samples were placed in a batch reactor and the oxygen uptake was measured with a dissolved oxygen (DO) probe. The OUR test method used is described in Vollertsen (1998). The OUR experiment provides information on the biochemical nature of the eroded sample which comprises of organic matter, particles and dissolved solids. The organic matter is basically the substrate that provides the food supply




for the proliferation of the microorganisms present in the eroded sample. Quantifying OUR can give a measurement of the impact in terms of oxygen demand that eroded sediment would have on a receiving watercourse if these sediments were released from the sewer system.

The organic matter present in the eroded samples was quantified in terms of its Chemical Oxygen Demand (COD). The COD of the organic matter is known as readily biodegradable COD and is fractioned into fast and slowly hydrolysible COD. This information is required as part of the conceptual model proposed by Vollertsen & Hvitved-Jacobsen (1998). The fast and slowly hydrolysible COD undergo hydrolysis to form readily biodegradable COD. Hydrolysis is a process where complex molecules are broken down into simpler molecules for easy ingestion by the bacteria. This process is driven by the enzymes secreted by the microorganisms. Biodegradability is defined as the sum of the readily biodegradable substrate and fast hydrolisible substrate following the work of Hvitved-Jacobsen et al., (1998a, 1998b).

16.2.2 Experimental Procedures

16.2.2.1 Test Procedure Each test comprised two parts: a deposit formation phase, and then a controlled erosion

test. During the first part, the deposit formation phase, the sediment was carefully placed in the inner annular sediment container, which was already filled with water and then gently scraped flat. In each test a sediment bed 5cm deep was placed in the inner annular container. The flume was then slowly filled with tap water and the upper plate replaced. In each test the upper plate was 270mm above the level of the original sediment bed. The temperature of the room had already been set at the required temperature. In each test the DO level was set at 70% of saturation. The flume was equipped with a system that allowed the DO level to be maintained at a proscribed level of the saturation level. A small tank of water, at the same temperature as the water in the flume, was located in the centre of the flume. It was continually aerated so that the DO levels in the tank were always close to saturation. If the DO level in the flume dropped below the set value of 70%, an automatic PC controlled pump would operate and introduce the highly aerated water from the central tank into the flume. This meant that after the initial reduction in dissolved oxygen the DO level in the flume could be maintained at the 70% saturation level throughout the deposit formation phase. During the deposit formation phase the upper plate rotated slowly to ensure adequate mixing throughout the water in the flume of the dissolved oxygen. The deposit was then left for a pre-determined amount of time to consolidate.

After the consolidation period was over the deposit was then subjected to a controlled erosion test. The speed of the lower and upper plates of the flume were increased in a step wise manner so that the bed was subjected to a series of shear stress steps in which the shear stress was steady but its value increased from step to step. During each step that lasted approximately 30 minutes the suspended sediment concentration and dissolved oxygen levels were measured continuously. One of the advantages in using an annular flume over field studies to investigate erosion behaviour is that any sediment found in the water column must have been removed from the bed. In the field situation it is difficult to isolate the influence of upstream sediment supply. Therefore sharp increases in total suspended sediment concentration with time indicate high erosion rates from the deposit. Discrete samples of the eroded material were also taken using the automated sampling system. Samples, 250ml in size, were recovered approximately every 300 seconds. These samples were used to determine total suspended solids, volatile suspended solids concentrations using standard filtration techniques and particle size distributions using a Malvern particle size analyser. Some of these samples were increased in size, to 2500ml, and then also used to assess the biodegradeability of the suspended sediments using a specialist technique aimed at




examining the oxygen utilisation rate first developed by Vollertsen (1998) and outlined in section 16.2.1.2. Oxygen utilisation rate measurements were made at three times during each erosion test, once when the deposit had just started to erode, an intermediate time when the erosion rate increased significantly and finally at the highest shear stress when the deposit was highly mobile.

At the end of the erosion test detailed topography measurements were made of the entire deposit surface so that the size, shape and the frequency of the bedforms that were created during the mobilisation of the deposit were quantified. This was done using a laser light sheet and video camera, so that the bed topography of the whole sediment surface could be measured with a streamwise, lateral and vertical resolution of approximately 1mm. The flume was then operated for a further 3 hours, and then 21 hours, at the highest shear stress level and the bed surface re-measured so that the stability of the bedforms could be examined. This data is being used to investigate whether the organic content of the deposits will have a significant effect on the height and shape of bedforms that will develop in sewers.

16.2.2.2 Test Programme The test programme contained 5 tests in which two sewer sediment mixtures were used.

These tests are numbered 3 to 8, the first two tests used artificial surrogate sediments and their results are not reported here. Each test had two parts the depositional and then the erosional section. During each test the temperature and dissolved oxygen levels were measured and continually adjusted, as described above, so that constant levels could be achieved during the deposit formation phase of each test. The levels of temperature and DO were selected to resemble those in a northern European sewer. The sediment deposits were left in these conditions for periods of approximately 18, 36, and 60 hours before being subjected to stepwise increases in boundary shear stress to assess the erosional stability of the deposit.

Table 1 Summary of experimental tests and deposit formation conditions

Deposit Formation Conditions Temperature DO Level Duration

Test No.

Sediment

°C (% of saturation) (hours) 3 Loenen 14 70 36 4 Dundee 14 70 36 5 Dundee 4 70 36 6 Loenen 14 70 18 7 Loenen 14 70 60

The two types of sediment mixture were sourced from different combined sewer systems,

Dundee in the UK (Ashley et al., 1992) and Loenen in the Netherlands (Schellart 2002). Typical physical characteristics of the sediment mixtures used are given in and Table 2.

Figure 16-3




0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

0.01 0.1 1 10

Sieve Size (mm)

Prop

ortio

n M

ass

Ret

aine

d

Dundee SedimentLoenen Sediment

Figure 16-3. Particle size distribution of the inorganic particles within each sewer sediment mixture.

Table 2 Average sediment characteristics of original sediment mixes used to form deposits Inorganic characteristics Proportion of organic

sediment by mass Sediment

D50 (mm)

σg (%)

Dundee 0.48 4.6 2.5 Loenen 0.265 1.8 1.7

where:

σg = (D84/D16)0.5 (eqn . 1)

and D84 = particle size for which 84% of particle size distribution is finer by mass, D16 = particle size for which 16% of particle size distribution is finer by mass, σg = sorting coefficient

16.3 RESULTS

This series of tests provides, for the first time, comprehensive data on the influence of various environmental parameters on the development of deposit strength for in-pipe deposits composed of real in-sewer sediments. The data has also provide high quality observations necessary to understand how real in-sewer sediments are released during a “first foul flush” event and how sediments in a sewer can transform between high flow events and that how the biochemical processes associated with released sediments can alter the in-sewer flow quality.

16.3.1 Erosion of Sewer Sediment Deposits For all the sewer sediment deposits, which were formed at 14°C, a characteristic pattern of

erosion was observed. Figure 16-4 indicates the erosional pattern observed in test 4. As the upper and lower plate speeds, and thus the bed shear stress, were increased the total suspended solids increased very slowly at first. After approximately 5000 seconds the total




suspended solids started to rise sharply, this continued for a further 500 seconds and then the TSS values remained fairly constant even though the shear stress being applied to the bed was being raised in a step –wise fashion. In the final stage of the test the bed started to erode very rapidly and as a consequence the TSS values rose sharply. It was at this stage of the test that significant quantities of inorganic sediments began to be released. A similar pattern was observed in Test 3, which used a bed of Loenen sewer sediment deposited under identical environmental conditions. In this test the TSS rose slowly for the first 12000 seconds before increasing very rapidly, for a period of 30 minutes. The TSS then dropped and remained rather stable before it started to rise rapidly again after 20000 seconds had elapsed, indicating the start of a second phase of bed erosion.

In both tests the percentage of organic material in suspension (100-60%) was much greater then the mass percentage of organic material in the bulk deposits (2.5-1.6%) indicating that the organic material was significantly preferentially entrained. However the two rapid phases of bed erosion were categorised by a rise in the proportion of inorganic sediment in the flow. It is clear that the rapid increases in TSS observed in both tests were related to sudden releases of inorganic sediment rather than any substantial increases in organic material. Generally after the first erosion phase the percentage of inorganic material increased for the rest of the duration of each experiment.

0

50

100

150

200

250

0 5000 10000 15000 20000 25000Time Elapsed (s)

Susp

ende

d So

lids

Con

cent

ratio

n (m

g/l)

Upp

er p

late

spe

ed (r

ev/m

in

10)

TSS (mg/l)

VSS (mg/l)

Upper plate speed

Figure 16-4. Total suspended soilds and volatile solids measurements from discrete sampling, along with upper plate rotational speed, for a sewer sediment deposit, formed

using sediment from Dundee, at 14°C with a DO level of 70% of saturation Comparison between tests 4 and test 5 indicates the influence of temperature on the

development of resistance to erosion by a sewer sediment deposit. In these tests, the sediments used were taken from the same location, and had very similar particle size distributions and organic content. They were consolidated for identical periods of time and the DO level in the overlying water was maintained at 70% saturation the only environmental difference was the temperature of the water. In test 4 this was set to 14°C and 4°C in test 5, this was selected so that any biochemical processes would be negligible. When exposed to a





similar pattern of shear stress steps it is clear that the overall mobility is lower, with TSS values approximately 50% less, and the pattern of erosion of the deposit formed at 4°C is significantly different. In test 5, after 10000 seconds the TSS values rise slowly and then more quickly, in a similar pattern to purely granular sediment deposits. No “two stage” erosion pattern as observed in tests 4 and 6 is observed. The critical shear stresses for erosion have been estimated at 0.45N/m2 for test 4 and 0.6 N/m2 for test 6. This value is in accordance with a predicted threshold of motion using a method such as Shields (1936). However once the deposit has been mobilised the proportion of inorganic sediment in the suspension rises slowly as the shear stress increases as in the previous tests.

0

20

40

60

80

100

120

140

160

0 5000 10000 15000 20000 25000

Time Elapsed (s)

Susp

ende

d So

lids

Con

cent

ratio

n (m

g/l)

Upp

er P

late

Spe

ed (r

ev/m

in

10)

TSS (mg/l)

VSS (mg/l)

Upper Plate Speed

Figure 16-5. Total suspended soilds and volatile solids from discrete sampling, along with upper plate rotational speed, for a sewer sediment deposit, formed using sediment from

Loenen, at 14 °C with a DO level of 70% of saturation The impact of different deposit duration was examined by comparing the erosional

response for the three deposits composed of sediment from Loenen but formed with different depositional durations. It can be seen that the longer the duration the weaker the bed, both in terms of the amount of TSS and also the top plate speed required to start erosion. The two stage erosional process is seen in the tests with the longer deposit durations (36 and 60 hours) however, this is not seen in the deposit with the shortest formation duration of 18 hours. The critical shear stress for erosion was estimated to be 0.65N/m2 for the deposit with an 18 hour consolidation period and 0.65 N/m2 for the 60 hour test. This suggests that the biochemical processes not only weaken the bed but that their influence also affects the pattern of erosion.



0

5 0

1 0 0

1 5 0

2 0 0

2 5 0

0 5 0 0 0 1 0 0 0 0 1 5 0 0 0 2 0 0 0 0 2 5 0 0 0

T im e E la p s e d ( s )

Susp

ende

d So

lids

(mg/

l)

Upp

er p

late

spe

ed (r

ew/m

in

10)

T S S ( m g / l )

V S S ( m g / l )

U p p e r P la t e S p e e d

Figure 16-6. Total suspended soilds and volatile solids from discrete sampling, along with upper plate rotational speed, for a sewer sediment deposit, formed using sediment from

Dundee, at 4 °C with a DO level of 70% of saturation

0

20

40

60

80

100

120

140

3 4 5 6 7 8 9 10 11

Top Plate Speed (rpm)

Ave

rage

Sus

pend

ed S

olid

s C

once

ntra

tion

(mg/

l)

18 hours36 hours60 hours

Figure 16-7. Total suspended solids and volatile solids from discrete sampling, along with plate rotational speeds for a sewer sediment deposit, formed using sediment from Loenen at

14°C with a DO level of 70% saturation for a range of consolidation times.

16.3.2 Biodegradability of eroded sediments The results obtained from the OUR values of the eroded sediments recovered from the

annular flume show that in the tests with the Dundee sediments which had the same duration of consolidation but with deposits formed at different temperatures, that the biodegradability




of the substrate was higher at 4°C than 14°C (Fig. 8). At lower temperatures, biodegradability of the substrate was high because of the low consumption of substrates by the microorganisms during the deposition (dry weather flow) stage. There also seems to be a net production of substrates at low temperatures. The growth rate of the microorganisms was also observed to be higher at 14°C than at 4°C.

In the tests with the Loenen sediments, which were conducted at the same temperature but different duration of consolidation, it was seen that the biodegradability was lower after longer durations of consolidation (60 hours), see . The growth rate of the microorganisms is lower after the longer duration of consolidation (60 hours) than when compared to the shorter durations of consolidation (18 hours and 36 hours). After many hours of consolidation, there appears to be a depletion in the supply of substrates which may cause a subsequent reduction in the growth rate of the microorganisms. In an actual sewer there is a constant supply of substrates from the upstream flow of sewage, however, in an annular flume, once the substrates are utilised there is no further supply to replenish the depleted substrate. The depletion of substrates also has a consequential effect on the microorganisms. The concentration of microorganisms is lower after longer duration of consolidation (60 hours) when compared with tests in which the “dry weather” consolidation period was only 18 hours and 36 hours. also shows that the biodegradability was the highest when the Dundee sediment was incubated at 4°C for 42 hours consolidation and the lowest when Lounen sediment was incubated at 14°C for 60 hours.

Figure 16-9

Figure 16-9

Figure 16-8

Figure 16-8. Variation of biodegradability for suspended sample at various conditions

0

20

40

60

80

100

120

140

160

180

200

0 5000 10000 15000 20000 25000 30000Time, sec

Bio

degr

adab

ility

, g

CO

D/m

3

Lounen 14 C 36 hrs Dundee 14 C 36 hrsDundee 4 C 36 hrs Lounen 14 C 18 hrsLounen 14 C 60 hrs

shows an example of the variation in OUR for Test 7. In general when comparing all these figures, the readily biodegradable substrate was the highest when the high shear stress is applied. This is probably because of the extra erosive strength exerted on the sediment bed releases the readily biodegradable substrate associated with the solids. However when the low and middle shear stress was used there was no significant difference in the biodegradability.




Figure 16-10 shows the data for the water quality parameters as well as the

biodegradability for Test 7. In general for all the conditions, there was a general increasing trend in the Total Suspended Solids (TSS), Volatile Suspended Solids (VSS) and COD with the increase in shear stress. The biodegradability was also increasing with the increase in shear stress. The eroded sediment was in an anaerobic condition (DO = 2 – 3%) towards the end of the experimental which may explain the increase in biodegradability.

0

1

2

3

4

5

0 10 20 30 40 50 60 70 80 90 100Time, hrs

OU

R, m

g/lh

0.21 N/m2 0.35 N/m2 0.70 N/m2

Figure 16-9. OUR profile for Loenen Sewer Sediments incubated at 14oC for 60 hours

0

50

100

150

200

250

300

350

400

0 5000 10000 15000 20000 25000 30000Time, s

TSS

, VS

S, C

OD

mg/

l

0123456789

Bio

degr

adab

ility,

g

CO

D/m

3

TSS VSS COD Biodegradability

Figure 16-10. Water Quality parameters and Biodegradability of Loenen sediments incubated at 14oC for 60 hours




16.4 DISCUSSION

It is clear that the biochemical processes have a measurable impact on the behaviour of a sewer sediment deposit subject to increasing bed shear stresses. The resistance to erosion that a sewer sediment deposit, both in terms of the threshold of initial motion and the total amount of sediment released can be doubled given similar sediment mixes and temperature and DO levels. In the tests, which were conducted at 14°C similar temporal “two stage” patterns of TSS increase were observed. This suggests that deposits are not random and homogeneous in nature once biochemical transformation processes have been active. It was believed that the “two stage” TSS/VSS response reflected the development of a surface layer on the deposits and that the surface layer developed during the consolidation period. As the shear stress increased during the erosion test particles were initially eroded from the surface layer. Later in the test as the TSS values started to increase rapidly, and the proportion of inorganic particles increased, it was believed that the surface layer no longer covered the whole surface but had been breached, releasing material from below it. The high proportion of organic material in the early stages of these tests indicated that the surface layer may be highly organic in nature.

The results showed that there was more biodegradability at 4oC than at 14oC. There was a net production of substrates at 4oC which also coincides with the reduction in the oxygen concentration in the annular flume. At low temperatures where the growth rates of the microorganisms were low, the substrates present are utilised very slowly which results in a net production of it. However for the tests with the Loenen sediments, there was very low biodegradability after very long durations of consolidation. After long hours of consolidation, there is a depletion in the supply of substrates. Consequently there will be a reduction in the growth rate of the microorganisms. It therefore appears that there are significant transformations occurring within deposits composed of sewer sediments over timescale in which they would be deposited within typical urban drainage systems.

In all of the experiments, there was an increase in the biodegradability coinciding with the increase in shear stress. At high shear stresses, more eroded materials were released from the sediment bed to the ‘sewage’ thus increasing the supply of substrates for the proliferation of the microbial population.

16.5 CONCLUSIONS

This paper provides preliminary results from a series of experiments in which deposits were formed in a flume using real sewer sediments. Measurements were taken to quantify the resistance of the formed deposits to erosion, and examined some of the biochemical processes associated with in-sewer deposits.

Sewer sediment deposits formed under different environmental conditions exhibit different resistance to subsequent erosion. In conditions in which the biochemical processes are inhibited (due to low temperatures) deposits behave as though they were composed of entirely inorganic sediment. When biochemical processes are significant deposits are generally weaker and appear to exhibit a two stage erosion process when subjected to steadily increasing levels of bed shear stress.

In situations in which biochemical processes are essentially inactive within a deposit the biodegradability of the released sediment can be high, and can therefore have an enhanced environmental impact if discharged into receiving waters. In situations where substrate is limited the biodegradability of the eroded sediments is reduced. The deposit environment (duration and temperature) can therefore significantly affect the biodegradability.




16.6 ACKNOWLEDGEMENTS

This paper reports on a project that involved researchers from a number of institutes, the authors are the principal scientists and representative of a large number of people who helped conduct the experiments, carry out the analysis and supply the sediment. In particular the following made a significant contribution to the project and without whose efforts the data presented in this paper would not be available, Mr Trevor McIlhatton, University of Liverpool, Mr Tony Schuit of TU Delft, Ir John Cornelisse, Mr John Coolegem and Mr Pierre Bosland of WL Delft. The project was funded by the EU, under its MRI programme, project no. HPRI-CT-199-0103, its support is gratefully acknowledged.

16.7 REFERENCES

Ashley R.M. Wotherspoon D.D.J. Goodison M.J. McGregor I. Coghlan B.P. (1992) the deposition and erosion

of sediments in sewers. Water Science and Technology 26(5-6) 1283-1293. Arthur S., Ashley R.M. Tait S., Nalluri C. (1999) Sediment transport in sewers – a step towards the design of

sewers to control sediment problems. Proc. Inst. Civil Engrs, Water Maritime and Energy 136, 9-19. Arthur S., Ashley R M. (1998). Near bed solids transport and first foul flush in combined sewers. Wat.Sci.Tech.

Vol.37, No.1, 1998 Ahyerre M. Chebbo G. Saad M. (2001) Nature and Dynamics of Water Sediment Interface in Combined

Sewers. ASCE Jour of Env.Eng. 127(3) 233-239. Bertrand Krajewski J-L, Briat P., Scrivener O. (1993) Sewer sediment production and transport modelling: a

literature review. Jour. of Hyd. Res. 31(4) 435-460. Bouteligier R Vaes G. Berlamont J. (2002) In sewer transport models 3rd Int Conf. On Sewer Network

Processes Paris. (is it in WST) Camuffo G. (2001) L’erosione dei depositi di dedimenti organici e implicationi nei sedimenti fognari. Testi di

Laurea, Universita degli Studi di Padova. Chatfield C. (1999) The analysis of time series-an introduction. Texts in Statistical Science, Chapman and Hall

CRC. Chebbo G. Ashley R.M. Gromarie M-C (2003) the nature and pollutant role of solids at the water sediment

interface in combined sewer networks. Water Science and technology 47(4) Clemens F H L R (2001) Hydrodynamic models in urban drainage, application nad calibration. PhD thesis Delft

University of Technology, Delft. De Sutter R (2000) Erosion and transport of cohesive sediment mixtures in unsteady flow. PhD Thesis

University of Gent. DeSutter R., Rushforth P.J. Tait S. Huygens M., Verhoeven R. (2003) validation of existing bedload transport

formulae using in-sewer sediment. Jour. of Hyd. Eng. 129(4) 325-333. Geiger W.F. (1987) Flushing effects in combined sewer systems. Proc. Of the 4th Int. Conf. On Storm

Drainage, Lausanne, 40-46. Gupta K. Saul A.J. (1996) Specific relationships for the first foul flush in combined sewers. Water Research

30(5) 1244-1252. Hvitved-Jacobsen T., Vollertsen J. & Nielsen P.H. (1998a), A process and model concept for microbial

wastewater transformations in gravity sewers, Wat. Sci. Tech., Vol. 37, No. 1, pp. 233-241 Hvitved-Jacobsen T., Vollertsen J. & Tanaka N. (1998b), Wastewater quality changes during transport in

sewers – an integrated aerobic anaerobic model concept for carbon and sulfur microbial transformations, Wat. Sci. Tech., Vol 38., No. 10, pp. 242-249

Jack A. Petrie M. Ashley R.M. (1996) Diversity of sewer sediments and the consequences for sewer flow quality modelling; Wat Sci Tech. 33(9) 207-214

Lau Y.L Droppo I.G. (2000) Influence of antecedent conditions on critical shear stress of bed sediments. Water Research 34(2) 663-667.

Mark O. Appelgren C. Larsen T. (1995) Principles and approaches for numerical modelling of sediment transport in sewers. Wat. Sci. Tech. 31(7) 107-115.

Ota J. Nalluri C. (2003) Urban storm sewer design; approach in consideration of sediments. Jour. of Hyd. Eng. 129(4) 291-297.





Rushforth P J (2001) Erosion and transport of cohesive sediment mixtures in unsteady flow. PhD thesis University of Sheffield.

Saget A. Chebbo G. Bertrand-Krajewski J-L (1995) The first foul flush in sewer systems. Proc. Int. Conf. on Sewer Soilds, Dundee, 58-65.

Schellart A., Clemens F., Veldkamp R. Tait S. (2003) A procedure for evaluating shear stress measurements in an annular flume using frequency domain filters. Accepted for IAHR Congress Thessaloniki, Greece, August 2003.

Schellart A. (2002) The use of turbidity measurements for sewer sediment monitoring. Masters Thesis Delft University of Technology.

Skipworth P J Tait S J Saul A J (1999) Erosion of sediment beds in sewers: Model development. ASCE Jour of Env. Eng. 125(6) 566-573.

Stotz G, Krauth K.L. (1984) factors affecting first flushes in combined sewers. Proc. Of 3rd Int. Conf. on Urban Storm Drainage, 869-878.

Tait S.J., Marion A. Camuffo G. (2003) Effect of environmental conditions on the erosional resistance of cohesive sediment deposits in sewers. Wat. Sci. Tech. 47(4) 27-34.

Thornton R.C. Saul A.J. (1986) Some quality characteristics of combined sewer overflows. Public Health Engineer 14, 35-38.

Vollertsen J. (1998), Solids in Combined Sewers – Characterisation and Transformation, PhD Dissertation, Environmental Engineering Laboratory, Aalborg University.

Vollersen J. Hvitved-Jacobsen T. (2000) Resuspension and oxygen uptake of sediments in combined sewers. Urban Water 2 21-27.

Vollertsen J. Hvitved-Jacobsen T. (1998), Aerobic Microbial Transformations of Resuspended Sediments in Combined Sewers – a Conceptual Model, Wat. Sci. Tech., Vol. 37, No. 1, pp. 69-76

the characteristics of an in-sewer … · observations from environmentally controlled annular ......

Documents