Monitoring accumulation sediment characteristics in full scale sewer physical model with urban wastewater

One of the major problems of combined sewer systems is the accumulation of sediments. Bed deposits inside conduits are produced especially during dry weather flow conditions caused by insufficient velocities or after heavy rain episodes. A consequence of accumulation in sewers is the loss of capacity to transport solids in conduits, because the height of bed deposits causes a decrease of the effective flow area and increases the hydraulic resistance. The organic content of sediment deposits also can produce gases (methane and hydrogen sulfide) and odours that are unhealthy for breathing (Ashley et al. 2004). Furthermore, the erosion and transport of sewer sediments caused by overflows during rainfalls is an important source of pollution leading to an impact in the environment, as combined sewer overflows or a surcharge at wastewater treatment plants (WWTPs) (Suárez & Puertas 2005). In order to prevent all previous problems it becomes necessary to clean sewer systems regularly, producing important costs of maintenance (Lange & Wichern 2013).

To approach the knowledge of sediment accumulation and transport processes in sewers, main investigations were focused in measurements in real sewers or studies of sewer hydraulics in full scale models. Uncertainties in flow rates and concentration of solids measurements are the main disadvantage of studying sediment transport in real sewers (Bertrand-Krajewski et al. 2003). To control this source of uncertainty, different laboratory experiments were performed on test flumes under controlled conditions. These full scale models often replace wastewater and sediments with non-cohesive granular materials (Ota & Nalluri 2003). Surrogate sediments with cohesive properties were also used to study erosion and sediment transport but without presence of organic matter, biofilms or gross solids (Tait et al. 1998; Banasiak et al. 2005). However, both non-cohesive granular particles and artificial sediment mixtures show unrepresentative behaviour of real sewer deposits. Few investigations were found that matched real wastewater under controlled laboratory conditions with the aim of calculating sediment accumulation and transport rates in sewers (Rushforth et al. 2003; Lange & Wichern 2013).

The aim of this study was to study the daily evolution of bed deposits in combined sewer pipelines under controlled conditions with real urban wastewater. For this purpose, a series of tests were carried out in a flume test facility located at A Coruña WWTP (600,000 inhabitants). As part of the experimental procedure, it was necessary to control sewer hydraulics and determine wastewater characteristics. At the same time, sediment heights were monitored with different techniques in order to calculate sedimentation growing rates.

The flume test facility was equipped with two different pipe geometries. A conventional circular PVC pipe with an inner diameter of 300 mm was compared with an egg-shaped cross section pipe. Both pipelines were made of plastic with a length of 8 m. The egg-shaped geometry was developed from a concatenation of arcs. For that, the top and bottom radius (R and r, respectively) and the total height (H) were defined. Both cross section pipes should keep an equivalent area so that results could be comparable. Therefore, egg-shaped cross section presented a total height of H = 385 mm, a top radius of R = 110 mm (220 mm width) and a bottom radius of r = 55 mm (Regueiro-Picallo et al. 2016).

A Nortek Vectrino© Acoustic Doppler Velocimeter was used to record 3D flow velocities. Centerline velocity profiles were measured at a distance of 4 m from the inlet chamber of each pipe, using one of the main apertures of the control section. Velocity data were measured with a vertical resolution of 5 mm and 2.5 mm for positions close to the pipe invert and with a sampling frequency of 25 Hz during 120 s, to ensure that the turbulence intensity was within 5% of its measured long-term average. All velocity data were de-spiked using the phase-space thresholding method of Goring & Nikora (2002) as suggested by Cea et al. (2007).

As part of the wastewater characterization, total suspended solids (TSS) and total dissolved solids were monitored and correlated from the records of turbidity (SOLITAX) and conductivity (LANGE 3798-S) probes, respectively. In addition, wastewater chemical oxygen demand (COD) was correlated with an organic matter (UVAS) probe (results not shown here). An automatic sampler placed at the head tank was utilized to collect a sample every 6 h in order to fit the best linear calibration of these probes. Manual grab samples at the head tank and the receiver tank were taken to complete the sediment and wastewater characterization.

A series of tests were conducted in order to study the evolution and characteristics of bed solids and the efficiency of circular and egg-shaped pipes. For this purpose, a constant discharge was established during several days with different slope and water depth conditions. Two slope values were set at 2‰ and 5‰. In addition, two different downstream boundary conditions were analyzed. On the one hand, tailgate was completely open so a free discharge was generated at the end of the pipes (critical depth). On the other hand, a water depth of 100 mm was established in the receiver tank with a fixed opening value of the tailgate. In two experiments (Nos. 4 and 7) with fixed downstream water depth condition, a small submergible pump was installed in the head tank to resuspend particles. This forced recirculation introduces largest particles into the pipes, avoiding their long-term accumulation in the head tank. All test conditions were summarized in Table 1.

For each test, sediment bed heights were daily measured. As part of the procedure, it was necessary emptying both pipes in a controlled way without affecting the sediment bed structure. Two different methods were set to record the deposit height. An imaging technique was utilized in order to measure several sediment profiles at the main apertures of both pipes and, at same time, the ultrasonic sensors recorded the sediment height along both pipes.

Sediment height was also monitored with ultrasonic sensors in the five opened windows while images were taken. The resolution of sensors was 0.13 mm and the deviation of the ultrasonic beam was 4.6 °. Therefore, measurements were dismissed when sediment heights were less than 2 mm in circular and 8 mm in egg-shaped pipe.

At the end of each test, two different types of solids were sampled from the pipes. A coarser mixture of granular and cohesive bed material was observed at the bottom while fine grained and organic slimes were stuck to the pipe walls. Particle size distribution of both types of sediments was obtained with a wet-sieved analysis following the ISO 2591-1:1988. In addition, it was calculated the amount of total solids (TS), the moisture content (W), the percentage of volatile (VS) and the density of each type of solid following APHA (1995) methods. In order to obtain the COD, it was followed McGregor et al. (1996) stirring method for preparing the sediment sample.

No bed deposits were recorded under critical depth downstream conditions for slopes of 2‰ and 5‰ (test Nos. 1 and 5) in both pipes. Bed deposits were only measured in circular pipe in experiments with fixed downstream water depth conditions. For the analysed hydraulic conditions, no sediment accumulation was recorded in the egg-shaped cross section pipe. The appearance of bed deposits was similar for all experiments; a granular-cohesive sediment mixture was formed because of the insufficient velocity of granular particles and the high presence of organic substances. Although fibrous materials may have allowed sands and organic solids to settle (Lange & Wichern 2013), a low content of fibers was observed due to the fact that sieving treatment removes a high amount from the wastewater used in the tests.

Particle size ranges associated with the circular pipe bed deposits and wall slimes are presented in Figure 5(c). Following Crabtree (1989), sewer sediment classification system, the bed deposits can be classified mainly as Type A sediment, composed of a mixture of fine-medium sand. The organic content of the samples ranged from 7% to 12%, indicating the presence of some cohesive sediment. The mean particle size of bed deposits matched with highest inorganic sediment fraction in Figure 4(b). Besides, small differences in the particle size (mean ± standard deviation) of bed deposits were found between experiments with the recirculation pump mixing in head tank (240 ± 28 μm) and without mixing (177 ± 32 μm). The pipe-wall biofilm are considered as Crabtree's Type D sediments. The biofilm is composed mainly of silt particles with some fine sand due to fat adhesion. The organic content of the samples is not very high, ranging from 11 to 18%, and the mean size is about 76 μm.

Table 2 summarizes the characteristics of the bed deposits and sediment biofilms in the circular pipe experiments. Bottom deposits were composed by particles with higher density and particle size, while wall slimes show higher moisture content. The sediment bulk density, moisture and organic content of the in-pipe deposits are in the range of Type A and Type C sediments of Crabtree's classification. The organic wall slime density and moisture are also in the range of the Crabtree's values. Smaller mean values were found for the organic content of biofilms and for the COD of bed deposits and biofilms in comparison with the reported values at the literature.

A mean growth rate of 3.8 mm/d was measured for test Nos. 4 and 7 with head tank mixing, while a rate of 1.4 mm/d for test Nos. 2, 3 and 6. The measured growth rates are similar to those found by Lange & Wichern (2013) for similar experiments with real wastewater conducted in 300 mm acrylic pipe with 1‰ slope. These authors reported a mean value of the daily change of bed deposits of 2.85 mm, with sporadic daily variations related with dry weather – wet weather flow variability. The higher sediment accumulation rate reported in tests with recirculation in the head tank are attributable to different particle size and composition, as it was mentioned in previous sections, and with higher sediment discharge. The mean sediment discharge was 1.2 and 0.9 g/s for experiments with and without head tank mixing, respectively.

Although circular pipe flow were accelerated as bed deposit height increased, the centerline Reynolds shear stress remained quasi-constant during the experiments, presenting a constant value of 0.170 N/m² near the bed in test No. 6 (Figure 8(c)). As the shear stress is not influenced by the sediment accumulation within the duration of the tests, a constant accumulation ratio can be expected. Higher mean velocity (0.39 vs 0.24 m/s) and centerline Reynolds shear stress (0.400 N/m²) values avoid the formation of bed deposits in the egg-shaped pipe.

A test flume facility was developed with the purpose of studying sewers hydraulics under controlled conditions with real wastewater from a WWTP. A circular plastic pipe with a diameter of 300 mm was compared with an equivalent area egg-shaped cross section. Bed deposits were only measured in the circular pipe for certain downstream conditions, while no-accumulation was recorded in the egg-shaped pipe. The deposited bottom solids showed higher density and particle size values, while wall biofilms introduced more moisture and organic content.

Bed deposits were daily measured to calculate their evolution. Depending on the particle size and the input sediment discharge, sediment growth rate range between 1.4 and 3.8 mm/d within the duration of the experiments (7–11 days). As the height of the deposit increased in circular pipe, slightly higher centerline velocity profiles were recorded but Reynolds shear stress remained nearly constant. Comparing the two different cross section pipes, egg-shaped pipe showed higher mean velocities and shear stress than circular conduit under the same fixed downstream water depth conditions. The experiments performed confirm that under low flow conditions egg-shaped cross section pipes present better self-cleaning conditions. As the sediment re-suspension and erosion of bed deposits in sewers during wet weather conditions is one of the main source of pollution associated with CSO events, the egg-shaped pipes can also be considered an efficient way of CSO mass reduction to receiving media.

This work was funded by the project ‘SEDUNIT: Análisis de los procesos de acumulación, erosión y transporte de sedimentos cohesivos en sistemas de saneamiento unitario’, Ref. CGL2015-69094-R (MINECO/FEDER, UE). The construction of the flume test was funded by CDTI through the FEDER-INNTERCONECTA Project Reference ITC 20133052, powered by companies ABN Pipe Systems S.L.U., EMALCSA and EDAR Bens SA. The research work of Mr. Juan Naves was financed by the Spanish Government (FPU14/01778). The authors also would like to thank to Montse Recarey for her assistance with the experimental work.