Extended CSO control storage: what could possibly go wrong?

The older areas of the London sewer system are combined sewers that carry foul flow and rainwater runoff. When it rains, the combined sewer system often becomes overloaded and excess sewage discharges from combined sewer overflows (CSO) into the rivers Lee and Thames. CSOs are required to reduce the risk of sewer flooding, limit damaging surcharging of the sewer system and overloading of the Becton and Crossness STWs. Currently, discharges to the tidal Thames occur about 50–60 times in a typical year at the most frequently overflowing CSOs. An estimated total of 40 million cubic meters of polluting discharges enter the river in a typical year from CSOs. These overflows are having an adverse effect on the environmental quality of the tidal Thames including ecological degradation and increased risk to human health.

The overall London Tideway Improvements programme includes capacity and treatment upgrades to five main STWs that discharge to the tidal Thames and the construction of the Lee tunnel (LT) from Abbey Mills to Becton STW and the Tideway Pumping Station that will drain the tunnel system. The Thames Tideway Tunnel (TTT) project completes the overall planned improvements and consists of a main tunnel from Acton storm tanks to a connection to the LT at Abbey Mills. Combined sewage controlled by the tunnel system will be stored and transferred to Becton STW for treatment. The capacity of the Beckton STW has been increased to capture a portion of the combined sewage generated during a storm and to empty the tunnel system within 48 hours.

An important consideration to the planning and design process, is the impact of storing combined sewage in a tunnel environment for time periods that influence the quality or chemical nature of the stored sewage. Odour generation is identified as a key process with issues associated with the transfer of H₂S from the surface into the tunnel, the dynamic chemical processes that occur during storage, the emptying regime for the tunnel and the adequacy of the ventilation system. However, it is recognised that the prediction and safe treatment of odour is a complex process. As a consequence, the TTT team conducted a series of studies that have enabled a better understanding of the processes associated with the formation and release of odour. These included the construction and testing of a purpose built pilot rig facility to mimic the operation of the prototype tunnel and the collection and analysis of a large number of CSO samples representative of the diversity of the sewerage catchment.

This paper focuses on the development of the pilot facility, which was constructed at one of the main London Tideway Tunnels (LTT) sites (Abbey Mills PS), aiming to mimic the operations of the LTT and to complete a series of tests to investigate the potential for sulphide generation during the filling, storage and emptying phases of the tunnel operation. This testing and research is unique by virtue of the size of the facility, the ability to sample actual combined sewage and, due to the detailed scope of the measurement system, continuous monitoring and sampling of the air and liquid phase of sewage within the rig.

A schematic diagram of the facility and the system of measurement and control is shown in Figure 1. The pilot facility consists of three parallel cylindrical pipes, each with a diameter of 1.2 m and a length of 16 m. During a CSO storm event, the pipes are filled with combined sewage from a representative CSO discharge adjacent to the pilot facility. To mimic a drop shaft operation, sewage was first pumped into the top of a 7 m tall vertical pipe and cascaded down the drop structure before it is distributed into the three pipes. The drop shaft was fitted with a series of internal shelves that dissipate energy but also aerate the flow, potentially stripping out gases. At the base of the drop pipe there was a weir to provide equal flow distribution to the three storage pipes. During storage of the sewage in the pipes, a number of key parameters including dissolved oxygen (DO) concentration, temperature and dissolved sulphide (DS) concentration was monitored continuously using sensors placed at access manholes. Sewage samples were collected every eight hours and analysed for biological oxygen demand (BOD₅).

A summary of the system filling and emptying operating conditions has been described in detail in previous studies undertaken by Thames Water Utilities Limited (TWUL). Similarly, a detailed description of the sampling strategy and analysis techniques applied are given in the ‘Thames Water report Odour and Septicity Pilot Plant Measurement System and Testing Protocol.

Except for the equipment shown in Figure 1, an additional ventilation pump for each pipe was installed, which fed into the 150 mm up-stand closest to the cascade tower. Ventilation pumps were used in tests with part full pipes and were designed to replicate one air change per day in the headspace. One air exchange per day is a requirement to be met during empty tunnel situations with the aim to maintain the air quality in the tunnel, prevent the built up of odours and slime growth on the tunnel walls.

Odour and H₂S measurements were completed at different locations to assess the magnitude and change in odour concentration prior to the cascade, at the cascade exhaust stack and in the storage pipe headspace. All measurements of odour concentrations were undertaken by Odournet, UK. Gas phase analysis was undertaken by transferring the liquid samples into a Nalophene bag, where H₂S in the headspace was monitored until the concentration was stabilised. Gas phase odour concentration was determined using olfactometry analysis to BSEN13752. To establish the constituents of the odour in the samples, Gas Chromatography-Mass Spectroscopy (GC-MS) was applied. The H₂S concentration of the samples was measured using a Jerome H₂S analyser. The sampling regime for odour, highlighting the duration of the test, sample location, sample test and number of samples over the duration of each test is given in Table 1.

Following the storage of combined sewage in the pipes, samples were collected for quantification of sulphate reducing bacteria in the combined sewage, biofilm and sediment. Three 50 ml samples of combined sewage were collected from each pipe using auto samplers immediately before the pipes were drained, after which triplicate samples of sediment were collected from the sediment traps (see Figure 2) in each pipe using sterilised spoons. Triplicate biofilm samples were collected in each pipe using specially designed biofilm samplers. After sampling, the combined sewage samples were filtered through 0.2 μm membrane filter, thereby retaining the micro-organisms in the samples on the filter. Microbial Deoxyribonucleic Acid (DNA), was extracted from each sample using a MoBio Ultraclean Soil DNA extraction kit. Subsequently, the dissimilatory sulphite reductance (dsr) genes were quantified in each sample using quantitative polymerase chain reaction (qPCR) (WRc 2001).

Following the methods outlined above, fourteen combined sewage events and one dry weather flow event were captured and analysed during the period February 2010 to November 2011. Six tests were completed with the pipes full, eight with the pipes partially full and one test with the system being emptied and refilled. Tests for the partially filled pipes were completed with and without ventilation. For the ventilation of the pilot plant a small fan was used. A minimum of one air exchange per day was achieved within the system. The primary findings highlight that the septicity rig was fit for purpose and that the results obtained yielded valuable insights for the prediction of H₂S and odour generated in a storage tunnel system.

The results of the 14 tests that were undertaken using the Septicity rig showed that:

A reduction in the levels of DO in the stored combined sewage was observed within short period of storage (6 hours from the start of each test) which indicates that there is potential for anaerobic conditions to develop within the LTT system. Other key parameters that were impacted by septic conditions during storage operation of the system include biochemical oxygen demand (BOD₅), suspended solids, oxidation-reduction potential and Total Nitrogen (Total N). The impact of temperature variations on the development of anaerobic conditions was also considered because an increase in temperature generally leads to an increase in microbial activity and hence the growth of sulphate reducing bacteria (Hvitved-Jacobsen 2002). Given that the sulphate bacteria are associated with the biofilms on the walls, even a short empty period would be sufficient for the bacteria to survive in the biofilms and therefore lead to a faster built up of sulphide within the system (Nielsen and Hvitved-Jacobsen, (1988)). The temperature of the combined sewage ranged from 11 to 19 °C at the start of the tests, with a slight reduction in temperature over the duration of each test. It is anticipated that the temperature of combined sewage in the LTT will be in a similar range hence the results from the test rig are likely to be representative of those in the prototype system.

The relatively fast reduction in DO had a considerable impact on the water chemistry, leading to a pH reduction from 7.3 to 6.2. The pH of the combined sewage is an important parameter as it determines the proportion of dissolved to total sulphide that is present in combined sewage and the potential levels of H₂S that are expected to be generated within the tunnel and released to the atmosphere. For sewage with a pH of 6.5, approximately 75% of the DS is in molecular form and it is in this state that odour problems are amplified (WRc 2011). At neutral pH (pH 7), approximately 50% of the total DS is in ionic form and 50% in molecular form. Above pH 7, the majority of sulphide is present in ionic form either as Hydrogen sulphide ion (HS⁻) or sulphide ion (S²⁻) (Gostelow & Parsons 2000). The values of pH observed in the septicity test rig confirm that molecular sulphide will be present, leading to a high potential for H₂S and odour generation within the tunnel during storage operations.

Generally, the results of the tests showed that DS was released as H₂S in the system after 10 to 24 hours from the start of the filling event. The rate of sulphide generation varied for each event and was significantly dependent on the temperature and the chemical and biological properties of the combined sewage (Gasperi et al., (2008); Kim et al., (2007)). The relationship between DS build-up and storage time showed a linear correlation up to a duration of 48 hours. After 48 hours a non- linear correlation between DS and storage time was observed. The impact of the key parameters including DO, temperature, pH and DS levels as a function of storage time are summarised in Figure 3.

Similar trends to those obtained for combined sewage were observed for Dry Weather Flow (DWF). The results of the DWF tests showed that sulphide build-up is highly dependent on the initial characteristics of the sewage and temperature. As expected, for tests undertaken under the same operating conditions, the build-up of DS in the dry weather flow sewage was greater than that in combined sewage.

A series of odour measurements were made for each of the 14 tests to establish the potential for odour generation at the cascade and within the head space of the pipes. The results of the odour measurements showed that the odour concentration of the air emitted from the cascade vent ranged from approximately 200 European Odour Units per cubic metre (OU_E/m³) to 15,000 OU_E/m³, with an average across all trials of around 2,000 OU_E/m³. Generally, the levels of odour concentration measured from the vent during summer months, were significantly higher than those measured during the winter months. It is expected that odour, in varying concentrations, would be released during the operation of the tunnel system if not treated. However, the concentration of these odour causing compounds is indicated to be relatively low and hence easily treatable by the air management plant designed for the project. An auxiliary output from the septicity rig studies was the development of a relationship between H₂S and odour units. Space limitations preclude the full explanation and presentation of the relationship, as part of this study, but further information could be provided to interested parties via contact with the primary author.

In order to determine the contribution of other odorants on the odour concentration, analytical measurements of volatile organic compounds were undertaken. The results showed that the relationship between H₂S and odour concentration varies and was dependent on the concentration of H₂S (Gostelow & Parsons 2000). As expected, at relatively low H₂S concentrations, Volatile Organic Compounds (VOCs) showed some contribution to odour. A number of potentially hazardous compounds were detected in the air samples collected from the cascade exhaust. However, the concentration of these compounds was generally well below the regulatory and risk criteria published by the Environment Agency (H4 Guidance-Odour management).

Sulphide reducing bacteria were identified within both the bed deposited sediments from the combined sewage and in the biofilm on the sides of the pipes. Analysis techniques were developed to successfully identify the presence of the bacteria but, due to the small number of tests, it was not possible to quantify their role in terms of H₂S generation over the duration of each test. At the end of each test the samples extracted from the rig contained sulphate reducing bacteria in the ratio 43% combined sewage, 56% biofilm and 1% sediment. These results highlight the importance of biofilms and it is recommended that these new measurement techniques are applied in further research within laboratory environs to examine the nature and role of these bacteria.

Capture of discharges from CSOs can reduce the risk to health of recreational users, improve water quality and protect the ecology of the tidal Thames. Capture of discharges in the transfer and storage tunnels of the LTT will lead to sulphide production when the combined sewage is stored. A pilot scale storage facility was been built by to investigate sulphide production during the operation of storage facilities with an aim to understand sulphide generation and to develop designs that mitigate against adverse impacts. Identified was a significant influence on sulphide build-up and H₂S generation from key parameters such as temperature and the initial wastewater characteristics (pH, DO, BOD₅ etc.). The microbial study of the sediments and biofilms showed that on average, 43% of the sulphur reducing bacteria were present in the combined sewage, 56% in the biofilm and 1% in the sediment.

The series of odour tests showed that there are likely to be several odorous compounds generated during tunnel operation. The concentration of these compounds was shown to be relatively low and the intensity of the odour was moderate to strong. However the ventilation and conventional odour treatment system at each release location are designed to safely treat the odours generated. Further research is currently undertaken as part of the TTT project to assess the impact of wastewater quality during extended storage conditions on the lining of the tunnel and on STW treatment processes.

The authors would like to thank the Thames Water Utilities Limited (TWUL) and Tideway for funding these studies and Odournet UK for the olfactory tests of the samples collected.