Combined Sewer Overflow Pollution

Chapter 1: Introduction

Background Information

The stormwater overflows, the flood risks, and the combined sewer overflows constitute one of the major areas of focus in the European countries. According to the study by García et al. (2017), a significant knowledge of the combined sewer overflow pollution becomes pertinent in the course of establishing measures which aid the reduction of emissions. Most of the European directives currently aim at quantifying and reducing the CSO for the purposes of protecting the environment as well as human health. Quality water would be needed for most of the leisure activities as well as achieving the agricultural need. According to the findings established by García et al. (2017), there is lack of proper knowledge associated to the quality as well as quantity of the overflows as noted through the policies. Commonly, the risk based approaches would primarily take into account the collection systems as well as the intensity of the episodes of rain that would lead to floods. It is worth noting that most of the sewers are considering the ultimate use of continuous water monitoring sensors, which are increasingly replacing the turbidimeters and traditional sampling. All these are done to estimate the suspended solids as well as the chemical oxygen demand in most of the sewers. Further relationship across the runoff and the pollutant load of TSS are closely attached to identification of the pollutants in the CSOs. The pollutographs and the hydrographs analysis would be applied in relating pea of concentration of flow of the pollutant and the mass as well as flow of the cumulative volumes. However, Yang et al. (2018) dismissed the fact that most of the traditional approaches towards flood and sewer analysis would find space in era full of technological developments. The authors noted that there has been unpredictable global climate change over the recent times, extreme weather events, accelerating urbanization, flood hazards and extreme precipitation. This is believed to have attracted both engineering and the non-engineering approaches meant to deal with the flood hazards and the heavy rainfall. The recently used floods modelling as well as simulations have been faced with a chain of challenges especially when it comes to handling the urban areas. There has been a chain of reasons towards the emerging challenges with some of them including the complex simulation process associated with a complex drainage system and high requirements and standards attached to the input data. Despite the challenges, Yang et al. (2018) still insisted that the quality and the quantity of data covering the storm sewer system remains indispensable when it comes to flood simulation. With the limits associated to data availability as well as the current computing power, it is evident that low precision or more simplified data would commonly be used for the relevant urban hydrologic modelling (Lennon et al. 2014). In some cases, however, the low precision data is likely to compromise the needful accuracy linked to flood simulation. A wide coverage of literature indicates that there has been increased use of the multi-scale data which is commonly utilized in examining the effect of data resolution and data precision (Thorne et al. 2018). This has been associated to sensitivity and accuracy of flood simulation and modelling. Some of the researchers have gone ahead exploring the impact of topographic data on simulation outcomes linked to the hydrologic models. For instance, the digital evaluation model has been evaluated for the output uncertainty associated to the Soil and Water Assessment Tool in a range of scenarios. Notably, DEM resolution would have an impact on the sub-basin classification, the stream network and watershed delineation. In addition, the Hydrologic Modelling System gave room for exploring the size and the number of the sub-watersheds, which are closely linked to runoff hydrographs.

It is worth noting despite the efforts in establishing the highlighted systems; most of the studies have never established the driving mechanisms related to the complexity of the urban storm sewer. This is based on the shortcomings of the outcomes of the hydrologic models. The missing details regarding the appropriate means of handling the sewers compelled Wang (2014) to provide insights associated to the combined sewer overflow, which is regarded as a mixture of the municipal wastewater and the urban runoff which is either discharged in canals or rivers. This would commonly happen during the heavy rainfalls thereby releasing the dissolved contaminants, bacteria and viruses, matter loads and other pollutants that would tamper with the aquatic life. According to the findings established by Wang (2014), CSO would ruin the quality of water either temporarily or spatially. Such components such as polychlorinated biphenyls have a history of polluting Harlem and Hudson Rivers. Due to the dangers brought by CSO, there is increased need for determining levels of pollutants and nutrients with which the finds would help in reducing water contamination while focusing on enhancing the quality of water in rivers.

Research Problem

The studies above have highlighted the issues with CSO and the problems that have been attached to it. Some of the studies even expressed worries of managing the sewers. The escalated levels of pollution and failure to manage the runoffs or the stormwater imply the inefficiency of the approaches that have been deployed before to contain the problems. Some of the problems have been deemed as being traditional and lack the capacity of containing the increased urban populations, and advanced sewers and drainage systems. Due to these traditional approaches, CSOs are becoming a threat to the environment as well as the human health. Areas of weakness further include the policies which may not be up to date or are not attuned to the dynamics. Therefore, the gap areas that need to be addressed include designing a system that would do planning, analysis and even produce a framework that would facilitate the stormwater runoff, drainage systems and the sanitary sewers. This can still be aligned to the design of appropriate control strategies that would help in managing the CSOs.

Research aim and objectives

The aim of this research is to investigate, analyse and design significant control strategies that would minimize CSO in sewer system using the Storm Water Management Model (SWMM) as a software tool. The supporting objectives include

To explore the damaging effect of Combined Sewer Overflows

To investigate control strategies that minimizes CSO in the sewer systems

To study the constituents of Storm Water Management Model

To examine the use of SWMM in regulating the CSO

To use SWMM to analyse and design control strategies that would limit CSOs

Structure of the Dissertation

The research is divided in five chapters. The first chapter provides the introduction to the research. In this case, the chapter touches on the background information, highlights the gap areas, denotes the research problem and reinstates the research aim and objectives. The second chapter provides the literature review in which a range of materials are reviewed for the purposes of extracting relevant findings linked to the research. The third chapter provides the methods that would help in designing the control strategies. With the help of the SWMM tool, it is possible to define the cycles for strategy development or tool development. The fourth chapter shares the findings of the research on the basis of the research process. The fifth chapter provides a discussion of the findings in chapter four. This would include the analyses of data that would lead to a justifiable conclusion.

Chapter 2: Literature Review

Chapter introduction

The need to design robust solutions for the urban flood risks as well as the CSO significantly attracts the attention of research in reviewing some of the articles related to the topic. Significantly, this taps into the theoretical and methodological contributions made by other materials towards the research. Areas of interest include the damaging effect of CSO, strategies to minimize CSO, the constituents of SWMM and the role of SWMM in designing and analysing the control strategies.

Combined Sewer System and Overflows

Sakson et al. (2018) noted that the research about sewer systems and the CSO is important as far as the idea of sewer treatment is put into consideration. The sewer systems which convey both the storm water and sanitary sewage are commonly referred to as the combined sewer systems. During dry season and moderate rainfall, the CSS is said to have the capacity of conveying all the flows to the appropriate wastewater treatment facility. However, heavy rainfall can increase the volume of the discharge, which may exceed the capacity of CSS thereby leading to the untreated sewage to get back to the basements and overflow from the manholes to the street surfaces. Thames21 (2020) highlighted the fact that due to overflows, Thames Water has incurred a total cost of £715 million while upgrading the 6 sewage treatment works for the last 3 years. More attention is given to interception of rainwater which would end in up in the sewer pipes if no control measures are put in place. Thames21 noted that for the purposes of boosting the capacity of CSS, it is more recommendable to stop more water from reaching the sewers. In Oregon, around 35% of the rainwater was impeded from reaching the sewers through the natural systems. Notably, green drainage would therefore be used in London for the purposes of boosting durability of the upgrades. In another case study, around 60% of the New York City is said to have a combined sewer system. In this combined sewer system, it is said to have a single pipe which carries sewage from the buildings and stormwater. The mix of sewage and stormwater is essentially sent to the wastewater treatment plant. It is noted that New York also experiences rainstorms which makes the combined sewers to receive higher volumes than the normal flows. Most of the treatment plants in New York would end up having a lesser capacity when it comes to excessive discharge. When this happens, the city would experience a mix of the untreated sewage and stormwater, which is directed to the waterways. The events would lead to what is referred to as the combined sewer overflows as noted by Meyer et al. (2013). The study has it that in the 1970s, there were more than nine billion gallons of the combined sewage which is said to have reached Lake Erie. However, in 2000, the discharge was cut by almost half with the help of the treatment plant improvements and a subsequent combination of new construction. A further Project Clean Lake introduced in 2010 focused on the installation of the smart green infrastructure and huge storage tunnels, which managed the increase in flow volumes as well as the overflows. The reason as to why CSOs are gaining attention over the recent times is due to the flooding problems in most of the receiving water bodies. Around 90% of the CSOs are said to contain the untreated commercial, domestic and industrial wastes. Meyer et al. (2013) also noted high chances of finding contaminants from such sources like oil and grease, biochemical oxygen demand, pathogenic microorganisms, toxics, suspended solids, nutrients and floatables among others. Due to this scale of pollution, the water quality has declined thereby positing threat to the aquatic species, public health and the general aquatic habitat.

Risks of a Combined Sewer System and overflows

Vazquez-Prokopec et al. (2010) noted that exposure to the polluted water from the combined sewer systems and combined sewer overflows would probably lead to waterborne infections. Such infections include gastroenteritis, hepatitis, and respiratory, skin, ear and even wound infections. While most of the waterborne diseases emanate from ingestion of contaminated water, they may equally be contracted via the inhalation of the water vapours. Vazquez-Prokopec et al. (2010) further highlighted that swimming or eating shellfish and contaminated fish would lead to symptomatic diarrhoea and nausea. Imminent risks have been linked to the probable presence of slicks and floating debris found in the receiving waterways. At the same time, varying amounts of the toxic materials are believed to lie at the bed of rivers, streams and even lakes. Over the recent times, contaminated sediment is said to have emerged as one of the main human health and ecological issue across the United States and Europe at large. It is evident that contaminated sediment carries with it both the chronic and acute toxic effects especially on the aquatic life. The sediment has been deemed as a consistent source of the bioaccumulative toxic chemicals, which can still concentrate along the fatty tissues of fish among other organisms. On the other hand, Bi et al. (2015) further insisted on characterization of the risky pollutants that might be introduced by the CSOs. Bi et al. (2015) pointed out that the CSO discharges might present such pollutants like TSS, toxics, microbial pathogens, oxygen depleting substances which are measured with the help of BOD5, floatables and nutrients among others. The authors also noted that it is to weigh the risks through determination of the ranges of concentrations and median pollutant concentrations that can be found in the untreated wastewater, treated wastewater, dry weather SSOs, wet weather SSOs and the urban stormwater. Perhaps, chances of experiencing the actual impact largely depends on the pollutant present, the duration of CSO event, and the status of the discharged water, exposure and the location of the discharge. The first class of risks associated to the pollutants present in the CSOs are the microbial pathogens. Angerville et al. (2013) described the microbial pathogens as microorganism which can cause diseases in the aquatic biota. The pathogens can fall in three significant categories which include the parasites, viruses and bacteria. Bacteria can either be indicators or pathogenic. The presence of the indicator bacteria points at chances of having the disease causing organism. The principal bacteria that falls under indicators include the enterococcus, E. coli and the fecal coliform. The pathogenic bacteria can cause disease and they are not limited to Yersina, Vibrio cholerae, Campylobacter, Shigella and Salmonella among others. Bi et al. (2015) indicated that viruses can be another risk of the CSOs. Based on research, there are over 120 enteric viruses which can be found in sewage. Notably, viral concentration would vary greatly and would essentially depend on availability of infections across the entire population which is served by the sewer system, the season and viral count enumerated. Some of the viruses that have been linked to CSOs include the Coxsackie virus, the infectious hepatitis virus and the poliovirus. Besides, some of the parasites that would be found in the CSOs include the helminths and the parasitic protozoa. The latter covers the Entamoeba, Giardia and Cryptosporidium. Helminths cover whipworms, tapeworms, roundworms and hookworms among others.

Apart from the Microbial pathogens, BOD5 has also been dubbed as another risk. It is associated to the measure of the levels of the oxygen demanding organic matter either in wastewater or just water. When amounts of the BOD5 are released to a water body, chances are that the dissolved oxygen can easily be depleted. With such low levels of oxygen, it makes it hard for fish to survive in the aquatic habitat. Apart from BOD5, TSS is also highlighted as a measure of small particles linked to the solid pollutants which float on the either the surface of water or wastewater (Rizzo et al. 2018). A range of materials in TSS are not limited to silt, industrial waste, animal matter and decaying plant. High levels of TSS can lead to such problems for the aquatic life and the stream health. TSS can possibly clog the fish gills, impair larval development, impair reproduction, reduce the growth rates and reduce resistance to diseases. Apparently, solid deposits can damage most of the habitats by simply filling spaces in the shelters. At the same time, TSS is likely to accumulate in the areas of the CSO thereby paving way for turbid conditions which would eventually smother the eggs and the aquatic insects (Vezzaro and Grum 2012). Nevertheless, CSOs introduces toxics, which constitutes the chemicals as well as the chemical mixtures thereby leading to the human health risk. Commonly, toxics would cover the synthetic organic chemicals, metals and even hydrocarbons. Concentration of these chemicals in the wastewater is one of the greatest concerns in most of the industrialized areas. Environmental problems that emanate from toxicity of the chemicals can either be long term or chronic. The chronic effects are normally subtle and would be difficult to be identified. The effects can be felt in terms of the reduced biological diversity, biomass and lower productivity among others. Other risks have been linked to nutrients and floatables. Nutrients are commonly linked to phosphorus and nitrogen. When there excessive amounts of phosphorus and nitrogen, then this would lead to rapid growth of the nuisance plants and algae (Masi et al. 2017). This can also lead to eutrophic conditions, which have been associated to oxygen depletion. Floatables, on the other hand, describes the debris, trash and any other visible materials which would be discharged at the time when there is an overflow of the sewers. Floatables in the CSOs would lead to accumulation of the detritus and litter in the streets. Floatables have the most adverse effect on the wildlife as far as ingestion and entanglement are put into consideration. Apart from highlighting areas of risk, Bi et al. (2015) went further discussing factors that introduce more risks due to CSOs. The first factor includes the elapsed time from the time when the wet weather started. It is evident that high pollutant concentrations would be expected in the primary stages of the CSO event. The second factor includes the period between recent episode of wet weather and the current one. Lengthier dry periods would attract high pollutant concentrations. Lastly, the duration and intensity of wet weather event also matters.

The damaging effect of Combined Sewer Overflows

When CSOs are left unmanaged, they are likely to have adverse effects on the human health and environment as well. Wang (2014) noted that CSOs would lead to strong oxygen depletion in most of the rivers thereby tampering with the water quality and the aquatic life. The contaminants brought in by CSO would tilt the oxygen demand, affect nutrients, and introduce toxic substances like ammonia, microcontaminants and heavy metals. Some of the physical parameters which are likely to be affected include the pH, redox, temperature, the flow and suspended solids among others. In case the CSO surpass the available capacity of the Waste Water Treatment, then there would be high chances of contaminated discharge received at the river. Wang (2014) established that domestic sewage, agriculture and industries form three major sources of the highly problematic phosphorus pollution especially to the aquatic environment. In his findings, Wang (2014) cited high nutrient levels at the downstream areas of most urban rivers which are close to the WWTP facility. It is even worse that some of the combined sewage systems are known for ferrying both the sanitary wastewater and the stormwater. If this happens, treatment can only be achieved during dry weather. During seasons of wet weather, the urban areas would experience heavy rainfall with water volumes even exceeding the pipe capacity. This means most of the waste would be discharged without being treated. Further observations regarding the damaging effect of CSOs were made by Tavakol-Davani et al. (2016). The authors noted that when the sewer network is overwhelmed, it is possible that any excess water would flow to the adjacent bodies for the purposes of relieving the system as well as preventing chances of flooding the manholes. This means that CSO would taint the water bodies with largely untreated waste water which comes from the industrial units, commercial as well as residential units. Such incidents would ruin the water quality while raisin an alarm over the public health concerns. The main negative effects brought by CSO might be frequent when there are increased incidents of sewer overloading as well as flooding across the urban areas. This is due to the fact that CSOs are commonly determined by significant changes in terms of the duration, frequency and intensity of rainfall. According to the arguments raised by Tavakol-Davani et al. (2016), some of the studies have noted the use of Global Circulation Models as part of regulating the adverse effects of CSO. As a result of the forecasted changes in terms of cycles of precipitation, there is needful design of the CSO control plans. Some of the centralized techniques target the lumped as well as large scale facilities while insisting on the need of enhancing the conveyance capacity of the entire sewer system. With the significant focus on the City of Toledo, the study noted the essence of controlling the effects of CSO by making use of RWH, which takes note of the CSO characteristics which need to be evaluated. Further studies conducted by Even et al. (2007) noted that Paris and the suburbs needed immediate attention as a result of the impact the CSO had on River Seine. The study noted that the sewer system and the overflow control became part of the critical concerns which needed to be factored in the management plans. The fate of urban effluents would essentially be characterised via the in situ samplings, modelling studies and laboratory experiments. Before the late 1990s, almost 50km stretch of Seine was permanently affected as a result of the high oxygen consumptions, which accounted for at least 112% of the total flux stream. Further analyses noted that 20% of the demand emanated from the CSO. Therefore, maintaining such a demanding system can be costly especially when permanent solutions are not developed.

The control strategies that minimizes CSO in the sewer systems

Over the years, system developers have been working on control strategies, which would reduce the CSO across the sewer systems. Garofalo et al. (2017) stated that one such strategy is felt with distributed real-time control (DRTC) system. The DRTC takes advantage of the multi-agent paradigm as well as the gossip based algorithm. In this case, the urban drainage system (UDS) constitutes the movable gates as well as water level sensors introduced across the network. Garofalo et al. (2017) further insisted that the UDS need to be modelled with the help of Storm Water Management Model, also denoted as the SWMM simulation software. It is worth noting that the SWMM is simply an open source computer model used by the popularly known hydraulic engineering community. The UDS aims at collection as well as delivering a combined wastewater from the defined urban catchment to the wastewater treatment plant (WWTP). The urban drainage system would have by its side the storage units, junctions, weirs and conduits. All the conduits are directed to the collector pipe before delivering the discharge to the WWTP. Garofalo et al. (2017) believed that the UDS approach only attends to the reduction of the flooding phenomenon. However, the approach can still be modified to solve the CSO issue. This is possible through a control or regulation of the water flow especially on the outfall node. When water levels on the side of the outfall goes up as a result of the flow control, then the rest of the gates would be triggered for them to store more water while helping the outfall to reduce to manageable levels. Therefore, increasing the number of gates in the UDS system can lead to the significant reduction of CSO. Rathnayake and Anwar (2019) asserted that while non-structural measures might be utilized in solving the CSO issues, there are still numerous interactions in a range of the subsystems like the receiving water bodies, wastewater treatment plant, the sewer systems and the catchments among others. The dynamic nature of the wastewater quality in most of the sewer systems makes the entire scenario more complicated. Rathnayake and Anwar (2019) still indicated that most of the researchers believe that green infrastructure is a long-term strategy that would help in reducing the CSOs. However, it should be noted that the green infrastructure is simply a significant approach used in balancing the natural water cycle with the help of both non-engineered as well as engineered methods of water management. At the same time, storage tanks within the combined sewer systems have been effective in the control of the CSOs. Rathnayake and Anwar (2019 also supported the fact that the Real-Time Control (RTC) plays a focal role in the sewer network control. It is worth noting that the control algorithms would continuously receive necessary feedback from across the sewer system while triggering important adjustments. While RTC might look more convenient, they only gain significance in the sewer systems and not the combined sewer systems. Complicated systems such as the CSS would call for the use of the dynamic control algorithm, which is based on the multi-objective optimization meant to minimize pollution. Significant parameters in the dynamic control strategy include the catchment inflow, water level in sewer chamber, the spill levels in the sewer chambers, the through flow in the interceptor sewer and the combined sewer overflow discharge. In the dynamic control, the multi-objective optimization module and the SWMM can be integrated to attain the stated parameters. The resulting model is believed to have the capacity of conducting the analysis and simulation of the sewer network and the stormwater networks meant to satisfy the hydraulics, the hydrological as well as the water quality requirements. Rathnayake and Anwar (2019) noted that the model has been used in Liverpool, UK where testing was done. Further modifications were introduced to the intercepting systems for the purposes of achieving the needful accuracy.

The constituents of Storm Water Management Model

SWMM is a commonly discussed tool when it comes to urban drainage design as well as planning. Niazi et al. (2017) noted that SWMM has commonly been used in the water quality models and watershed hydrology around the world. Based on the findings established by Niazi et al. (2017), the SWMM is built based on the assumptions and the functionality of six major components. The first component is the external forcing data. In this case, both the long term and single event precipitation time series is believed to be used in the SWMM. Notably, the single events largely represent the design storm applied in the simulation of the pollutant loads and the runoffs. Based on this component, SWMM is believed to have the capacity of approximating daily temperature fluctuations with the help of the maximums and the minimums. The Hargreaves approach is dominantly used in getting the time series daily data, single constant value and the monthly average values. The second component is the land-surface component. In this component, attention is first given to the surface runoff where the SWMM divides the entire area to be analysed into sub-catchments. A water balance can be reached by taking each of the sub-catchment as a significant nonlinear reservoir (Gironás et al., 2010). Secondly, attention is given to infiltration where SWMM is engaged in computing the infiltration with the help of the Horton’s method. The method is known for expressing the capacity of infiltration as a significant decreasing function of time that is exponentially empirical. Other approaches that can be utilized include the Larson and Mein formulations. The third component of the SWMM process is the subsurface component, in which the model is required to account for the infiltration leading to rise in terms of the water table. This property gives room for the program to essentially reproduce what is referred to as long recession periods linked to the runoff hydrographs. The subsurface flow would eventually be modelled via assignment of two reservoirs to every sub-catchment. One of the reservoirs represents the deeper groundwater zone while the remaining one stands for the vadose zone. While handling the model, it is important to perform the water balance on the reservoirs especially when dealing with the dynamics of flow and storage (Niazi et al. 2017). The fourth component is the conveyance system component in which the dynamic wave as well as kinematic wave routing is involved. Notably, kinematic wave is attached to the continuity of an equation. The approach rarely accounts for the backwater effects, flow reversals and the pressurized flow. The intervention of the dynamic model bridges the assurance of handing the Saint Venant equations. Two sets of the equations mentioned under the third component are important in generating the hydraulic solution. The fifth component is the contaminant build up, transport and fate. In this component, a user is given room to select a more functional association that governs the pollutant build-up and the characteristics of the pollutant. The build-up can be modelled with the help of the Langmuir saturation function. At some point, the wash-off can easily be modelled with the help of the rating curve where mobilization is regarded as a function of flow, constant concentration and independent of the build-up (Niazi et al. 2017). The fate as well as the transportation of the pollutants can be determined using the advection and the mass balance. The final component constitutes the LID controls, which can be modelled based on the interconnected and mixed layers known for representing the pavement, drainage, overflow control, under-drain portions, infiltrations, pavement, and soil. SWMM is known for effectively modelling the infiltration trenches, rain barrels, bioretention cells, green roofs, street planters, and vegetated swales. Infiltration rates can be determined through assumption of the exponential relationship of the moisture content and hydraulic conductivity. Rossman (2010) also looked at the visual objects of SWMM and the way they can be used to represent the stormwater drainage system. First, the rain gages would supply the precipitation data for either one or even more sub-catchment areas defined within the study region. Properties of rain gages are not limited to the name of the data source, source of the rainfall data, the rainfall data type and the recording time interval. Another visual object includes the sub-catchments, which are also the hydrologic units associated to the land in which the drainage system elements are involved in directing the runoff to the discharge point. Other objects include the junction nodes, outfall nodes, flow diver nodes, conduits, the storage units, and the pumps. According to Zhao et al. (2017), SWMM plays a significant role in regulating, analysing, and designing control strategies that limit CSOs.

Chapter 3: Methodology

Chapter introduction

The project intended to establish the strategies that would be used alongside the SWMM tool for the purposes of taking care of the CSOs as highlighted at the start of the research. While the project is associated to integration of modules in SWMM 5.0, the adoption of the waterfall SDLC model is felt suit for development of the process. The model offers the linear sequential phases in which every phase is said to depend on the significant deliverables tapped from the prior phase. The waterfall model provides a guide through the phase for requirements, modeling and design, implementation, integration and testing, deployment and lastly, maintenance.

Requirements

This is the first phase of the entire process and covers the needs, purpose and the design with a significant outline of the specifications. The project carries with it three important modules that would be established on the SWMM platform. The first module, or strategy, covers the optimal load control which points at reduction of the pollution load (Miller 2016). The second module is the multi-objective optimization which provides the simulation model for dynamic rainfall and water quality routing and finally, the Low Impact Development (LID) control which evaluates chances of using the infrastructural support in managing the runoffs and limiting the impact of the CSOs. The pollutant parameters under observation include the flow, in terms of the flow rate and volume, total suspended solids, indicator bacteria, pH, settleable solids, biochemical oxygen demand, dissolved oxygen, nutrients and other toxic pollutants present in the stormwater (Rathnayake and Tanyimboh 2015). At the centre of all the parameters intend to point at the appropriate model to be used, the CSS characteristics to be considered, data that would be needed, the sampling criteria to be used, data management as well as analytical procedures to be used.

Modelling and design

For the purposes of achieving the optimal control and the LID control, the modeling and design phase is important. The phase has a number of stages in it. The first stage is problem formulation. Some of the pollution load evaluation parameters cover the effluent quality index (EQI) which is needed in the optimal control while determining the pollution load. Water quality parameters under check include the chemical oxygen demand (COD), total suspended solids (TSS), total Kjeldahl nitrogen (TKN), nitrates and nitrites (NOX) and the Biochemical oxygen demand (BOD). Based on these parameters, there needs to be configurations for the intercepting sewer with an objective function purposed to reduce the pollution load for the water making way to CSOs. The function can be expressed as

Minimize K1 = ∑ni=1LI

In this case, n and Li gives the interceptor nodes or the chamber points for the CSO as well as the pollution load. The continuity equations can be expressed as shown below.

Qi + qi-1 – qi = 0

Ac∆hc/∆t = Ii – Qi ; hc < hs

Ac∆hc/∆t = Ii – Qi – Oi; hc >hs

0 qi qmax,I

Where Ac is regarded as the surface area associated to the CSO chamber while qmax represents the maximum flow rate detected at ith conduit.

hs = spill level linked to the sewer chamber

hc = the water level for the sewer chamber

qi = the through flow in the interceptor sewer noted at node i

Ii = the catchment inflow to the node i

Qi = the flow from ith sewer chamber to the interceptor node i

Oi = the combined sewer overflow detected at node i

The module is made possible with the help of the NSGA II which is coupled with C programming language. However, multi-objective optimization can still be attained with the help of computational methods in one dimension (1D) determined through Saint Venant equations. The latter is used in modeling the conduits and pipes in either a steady or quasi steady state.

∂Q/∂t + ∂/∂x[BQ2/A] + gA∂h/∂x + gA.cf QQ/RhA2 = 0

∂Q/∂x + B (h) ∂h/∂t = 0

In which

Q = is the discharge

t = time

x = Displacement on x axis

A = cross-sectional area

B = cross section water width

h = water level

Rh = hydraulic radio

Cf = resistant coefficient

β = Boussinesq Coefficient

g = gravitational acceleration

The two models shown above only give a picture of the parameters and the relationship that prompts the control. As part of the design process, it was more appropriate to come up with a model that would represent the working design as far as different modules different are put into consideration.

The processes function is an important module in the system because it carries with it different function of the system. One of the functions is the optimal control, which takes advantage of the input while assessing different parameters.

Implementation

The implementation phase demands that smaller units of the system be implemented first before they can be integrated in the next phase. Every unit is therefore developed as well as tested for the functionality. For the purposes of developing the system, it was appropriate for the project to sample rainfall spatial heterogeneity for a hypothetical case. The hypothetical sample has storage tanks with around 10 combined sewer overflows. The flow routing as well as transport across the entire network can be reflected in the time delays which can range from around 5 minutes to 20 minutes.

The SWMM implementation takes into consideration the EFD rules and the RTC approaches. Based on the Equal Filling Degree Control (EFD), the system compares the CSO volumes with that of the BC scenario. Besides, the Model Predictive control would first check on the CSO volume while keeping in mind the fact that the MPC discharged would be less compared to the CSO volumes at the creek and to the river (Miller 2016). The percentage differences would be checked for the simulated CSO volumes and the BC as well as the EFD scenarios. It is also important to analyze the CSO events in the implementation phase. Some of the CSO events would even take place during the day or night. However, most of these events can easily be described based on the physical characteristics and the control algorithms. More attention is given to the behavior of nodes when there are changes in the flow rate within the system. A simple three step procedure is followed in this phase and these include the rainfall-runoff calibration, which is meant for estimating the sub-catchment parameters (Rathnayake 2015). Secondly, there is need to check on the base case calibration for the purposes of configuring the estimated parameters, and lastly, the EFD verification ensures the detailed model attains same output while applying the control strategies.

Integration and Testing

In this context, the SWMM US EPA is an open source product which needs to be integrated in the Low Impact Development module and NSGA II module. A significant check on the interacting processes is essential for the purposes of establishing the interception. The most sensitive side of the integration process is how the code works and whether it agrees with the code attached to the next module. While the LID module might delay for long to check on the properties, the optimal control and the multi-objective optimization should have reasoned functions that work in conjunction. Sampled code that needs a timely analysis can be shown below.

Return module line

The above code is part of the initialization code, which need to be integrated to the following codes that open or close nodes for water quality flow. The metaheuristic approach can be used when the EPA-SWMM5 is in place. The following code would attract further system integration and testing.

While the code would control node 1, it can still be tested along the code for node_flood_wgt_dict and target_depth_dict to see if the system can respond when data prompts certain signals.

Target_depth_dict = {“Node kt1”: {“target”: 4.0, “weight”: 2},

“Node kt2”: {“target”: 3.5, “weight”: 1}}

If the module for the simulation start time and date and the hotstart file are saved, then a function “run_ea” would initiate EA while noting the control functions. It is worth noting that integration may not be attained in one cycle. This means that tests have to be conducted from time to time to make sure that the entire system is consistent and would produce same results over time. In the decision flowchart, time t+1 imply that the process is an iteration which the system needs to attain (Sadler et al. 2019). However, attention need to be given to changes made to the input, which would change the kind of an output expected in each of the sub-systems. Therefore, additional variables need to be introduced to the starting module for it to adjust the changes.

Deployment

For deployment reasons, the project would consider two sub-catchment areas S1 and S2 within London for the purposes of demonstrating the modules integrated in the SWMM tool (Miller 2016). The design storm can be attained with the help of the SCS Type II distribution. For spatial variations reasons, the rain assigned to sub-catchment S2 would start 18 minutes after the rain event for the S1 sub-catchment. Both S1 and S2 can drain into parallel storage units which are generic. Two orifices, denoted as R1 and R2 would regulate the flow rate out of the storage units. However, R1 and R2 would meet at Junction J1 before the discharge would leave the system. The rules or conditions-based scenario would take 5 seconds, the routing time step in the MPC would take at least 10 seconds, and the passive scenarios would equally take 10 seconds (Sadler et al. 2019). The sampled constrains and objectives would appear in three significant cases during the deployment phase. For Case A, the case is simple and would only involve minimizing floods. For case B, tidal boundary conditions especially at the outfall would be necessary in maintaining the target water or load. Case C is multi-objective and checks on conditions that would allow the system to open the online tanks when the discharge cannot be managed. Significant properties under check during deployment include

Sub-catchment properties (slope, width, name, impervious and area )

Storage unit properties (Bottom elevation, initial depth, maximum depth and the surface area)

Pipe properties (Roughness, diameter, length)

Node properties (bottom elevation, max depth)

Orifice properties (height, discharge coefficient)

Maintenance and Control

Three important cases, reflecting the three modules need attention. For Case A, the scenario is passive and the outlets are all open with an inactive response system. For case B, the scenario is dubbed the rule based control where the logical rules can be used in controlling the orifice openings and discharge in the storage units. The heuristic rules have the capacity of changing a dynamic and actuated system. The MPC control appears in Case C with the capacity of adjusting the actuators on the basis of the forecast conditions. Most of the actuators in the system aim at regulating the conditions back to normal with attractive any negative outcomes. The actuators would perhaps open online tanks, which is a simulation of an alternative sewer meant for controlling the overflow.

Chapter 4: Results and Discussion

The project regards minimization of urban flood risk as well as combined sewer overflows. The concerns are directed to the pollution load and the stormwater gauge, which are both parameters of interest when it comes to designing a convenient system. In an effort to curb the risks highlighted in chapter 3, the project sampled a hypothetical case of figure 2 in chapter 3 as a platform to integrate three significant strategies that shall be incorporated into the same design. The three strategies include the optimal control, LID control and multi-objective control. The project made use of simulations with the help of SWMM version 5.1 and EPANET 2 sub-module in place of the NSGA module. The first concern was to map an area where the strategies would be applied. Mapping the area allows one to work on the junctions, nodes and conduits which direct to the online tanks. The code below paved way for mapping the following area.

For optimal control as the first strategy, it is more convenient to establish the parameters under study. Pollution load and stormwater level are more outstanding. The pollution load can however constitute more elements such as BOD, COD and nitrates among others depending on the load characteristics. It was appropriate to create the rain file and the pollutant load file while testing the code.

For the pollutant load

The SWMM allowed inputting the optimum conditions for the system with readings being reflected in the files while storing them in the memory. Some of the inputs and input curves in SWMM are as shown below.

The core purpose of having an SA was to maximize the CC and NSE while reviewing the overall hydrograph. Around 20 runs could be performed in the tool for the purposes of specifying the range with the help of the initial values. In the SWMM, it was important to vary the parameters while corresponding to CC and NSE for every run. For optimal control, the stormwater needs to be consistent for the system to have a history or a record of a possible runoff. Conditions for a runoff can be set. The level of R13, which records a rainfall of 90 mm can be regarded as a cutoff for a possible occurrence of a runoff with runways becoming incapacitated. In the pollutant editor, the optimal load can be determined by concentration levels. High concentrations of nitrates or phosphoric compounds would indicate high pollutant loads which needs attention. At the same time, low oxygen concentrations in a waste indicate the presence of many bacteria and other microorganisms, which can be dangerous to human health. Optimum control plays the critical function of data collection and notifying the system of the load conditions. The control can be determined through curves or the hydrographs thereby tracking the conditions for several hours. While optimal control may sound as a stage in the entire design, it is a strategy on its own because in most of the manual system, the step is necessary in just giving a history of the pollutant loads and stormwater.

The second strategy is multi-objective optimization, which picks the conditions from optimal load and designs the nodal, pump, junction and link characteristics. The strategy presents conditions that would make the system to react to particular conditions. A node would sense the conditions in the conduits depending on the prevailing conditions. Moreover, a number of rules in place determine whether the load conditions would activate the online tanks. Online tanks is a representation of an alternative sewer that stores extra load, which can only be treated after the first phase is offloaded from the system. Some of the nodal characteristics can be introduced as shown.

For the nodal outflow,

The system establishes a number of rules at the node as shown below

As such, nodal characteristics determine the operation of online tanks, pumps and the links, which communicate with the junctions. The links open or close depending on the nodal characteristics and load characteristics. The links connect the treatment plant to the sewer system, the sewer system to the pumps and the pumps to the online tanks, acting as an alternative sewer system, for the purposes of storing extra load before being discharged to the treatment plant. In essence, they form the most critical component in the system while regulating the overflows as a result of the stormwater, or excessive industrial discharge. The multi-objective optimization ensures that conditions of the system are at their best without creating problems. Data collected during the determination of the optimal load would be necessary in regulating the conditions at this stage. The code below was deemed necessary.

If the system is exposed to excessive discharge, the conduits would sense an overload. This is the same thing that happens when there is an overload in the pollutants, which includes very low concentration of oxygen and high concentration of the chemicals. While running the simulation, the following output is possible.

From node23.40 to node 21.00, the system is red pointing at an alert. Based on the preset conditions, there are a number of conditions that would lead to an alert. First, excess stormwater would lead to such an alert. Strained conduits, orifices, and valves would imply that the system cannot handle the discharge if the amount of the load persists. At this point, chances are that excessive stormwater is making its way to the treatment plant, which may ruin the quality of water coming out of the plant. Secondly, the red alert would indicate excess concentration of pollutants detected in the system. This implies that there are chemicals and microorganisms that are making their way to the runways without any treatment. Lastly, the red alert is indication of faulty conditions, which might be a rare case in the automation process. What matters is what the system does if there is a red alert. The green section indicates that the system is responding to the abnormalities by opening pump 14.45 and 1.30 while activating the nodal characteristics at node 1.96 and 3.20. The pumps would help in pumping the extra water or industrial waste into online tanks at 0.37. The action would help the permissible amount to enter the treatment plant and get discharged via node 0.29 and pump 0.32. There are a number of presumptions in the working of the response system. First, it is presumed that the system is fast enough in treating and releasing the waste before another phase enters the system. Secondly, the nodes and links are accurate enough for the purposes of relaying accurate information to the control system. Lastly, the system boosts the water quality to acceptable levels deemed harmless to the immediate environment. After releasing the necessary discharge to the treatment plant and sending the extra discharge to the online storage tanks, the systems looks as shown below.

After regulating the conditions, the system parameters go back to their default status as the red alert is cleared off the conduits. The active parts of the system include the online tanks and the treatment plant until the discharge or the overflow is fully treated and released at normal flow rate. The nodal characteristic are consistently checked and compared to the threshold standards. If the waste is absolutely discharged from the online tanks, the system or the simulation looks as shown below.

The online tanks remain empty while leaving manageable volumes at the treatment plant. The system is now ready to receive another discharge for treatment and further discharge to the runways or the green infrastructure that balances the cycle of infiltration. If the system clears the waste, every node and pump stands at 0.0, which means that the system is not working because it has not received the waste. This situation can be shown below.

While the process seems to work in phases, the flow was seamless and transitional in terms of regulating the conditions and responding to an overflow for the purposes of retaining optimum conditions. The situation in Figure 11 is only possible during dry season while Figure 10 represents sporadic episodes of stormwater and industrial discharge. The last bit of the strategies includes the LID control which gives a picture of the impact of the optimal control and the multi-objective optimization. This strategy narrows down scaling the low impact of CSOs as a result of the strategies that have been laid down. At discharge point, the LID control works in hand with the online tanks while enhancing the seepage of the overflows and cutting down on the runoffs.

The LID module only does comparison of the prior runoffs with the current runoffs after putting necessary measures in place. Based on the curves generated, it is evident that there is a reducing impact of the overflows because of the highly regulated discharge with a treatment plant put in place. The following code ran to produce the necessary report for the LID.

For the purposes of producing consistent results, the codes for the three modules are integrated with one function calling the other or returning to 0 at the end of the process. Based on the simulations, the first 2 runs could not work properly but after doing adjustments, the code worked effectively with functions responding to conditions in a more accurate way.

Chapter 5: Conclusion

The aim of the project was to design a sewer system that minimizes the risk of urban floods and the combined sewer overflows. This is one of the gap areas in research due to the consistent problem of untreated discharges that emanate from the overflows. Based on this argument, the research collected some of the findings from a range of studies before building the background of the study. The background details denoted the relationship between the pollution load and the runoffs while extracting the hydrograph analysis. The imminent use of the hydrologic models is said to be less satisfying for the CSOs that are likely to be discharged in rivers and canals as noted in the case of Harlem and Hudson Rivers. Due to such shortcomings, it was necessary to investigate, analyse as well as design some of the control strategies which would reduce the CSOs with the help of the SWMM tool. In a preview of a range of studies, notable areas that needed attention included a deep analysis of the effect of the CSOs. Based on a range of findings, CSOs are believed to have harmful effects on people and the environment. It leads to depletion of oxygen in rivers thereby affecting the aquatic life. Some of the control strategies highlighted in the course of reducing CSOs include making use of some traditional models, modern approaches and analytical tools. All of these approaches had shortcomings, which prompted the idea of developing a new approach. The third chapter focused on the methodology based on the waterfall SLDC. This approach demands that each functional unit of the project need to be tested separately before integrating the system. Based on the SLDC model used, the project highlighted the necessary requirements, modeling and design, implementation, deployment and testing as well as maintenance. As part of the project, three modules were necessary and they included the optimal control, multi-objective optimization and LID control. In a collective system, the optimal control focused on detecting the pollutant load or the stormwater load before relaying the information to the multi-objective optimization. The nodal characteristics, link, pump, orifice and storage tanks found the better side of the multi-objective strategy. Most of the responses are established in the second phase of the project while looking at conditions that would activate the node or link. The LID control came last as far as evaluation of the impact of reducing runoffs is put into consideration.

References

Angerville, R., Perrodin, Y., Bazin, C. and Emmanuel, E., 2013. Evaluation of ecotoxicological risks related to the discharge of combined sewer overflows (CSOs) in a periurban river. International journal of environmental research and public health, 10(7), pp.2670-2687.

Bi, E.G., Monette, F., Gasperi, J. and Perrodin, Y., 2015. Assessment of the ecotoxicological risk of combined sewer overflows for an aquatic system using a coupled “substance and bioassay” approach. Environmental Science and Pollution Research, 22(6), pp.4460- 4474.

Even, S., Mouchel, J.M., Servais, P., Flipo, N., Poulin, M., Blanc, S., Chabanel, M. and Paffoni, C., 2007. Modelling the impacts of Combined Sewer Overflows on the river Seine water quality. Science of the Total Environment, 375(1-3), pp.140-151.

García, J.T., Espín-Leal, P., Vigueras-Rodriguez, A., Castillo, L.G., Carrillo, J.M., Martínez- Solano, P.D. and Nevado-Santos, S., 2017. Urban runoff characteristics in combined sewer overflows (CSOs): analysis of storm events in southeastern spain. Water, 9(5), p.303.

Garofalo, G., Giordano, A., Piro, P., Spezzano, G. and Vinci, A., 2017. A distributed real-time approach for mitigating CSO and flooding in urban drainage systems. Journal of Network and Computer Applications, 78, pp.30-42.

Lennon, M., Scott, M. and O'Neill, E., 2014. Urban design and adapting to flood risk: the role of green infrastructure. Journal of Urban Design, 19(5), pp.745-758.

Meyer, D., Molle, P., Esser, D., Troesch, S., Masi, F. and Dittmer, U., 2013. Constructed wetlands for combined sewer overflow treatment—Comparison of German, French and Italian approaches. Water, 5(1), pp.1-12.

Niazi, M., Nietch, C., Maghrebi, M., Jackson, N., Bennett, B.R., Tryby, M. and Massoudieh, A., 2017. Storm water management model: performance review and gap analysis. Journal of Sustainable Water in the Built Environment, 3(2), p.04017002.

Rathnayake, U., 2015. Migrating storms and optimal control of urban sewer networks. Hydrology, 2(4), pp.230-241.

Sadler, J.M., Goodall, J.L., Behl, M., Morsy, M.M., Culver, T.B. and Bowes, B.D., 2019. Leveraging open source software and parallel computing for model predictive control of urban drainage systems using EPA-SWMM5. Environmental Modelling & Software, 120, p.104484.

Thorne, C.R., Lawson, E.C., Ozawa, C., Hamlin, S.L. and Smith, L.A., 2018. Overcoming uncertainty and barriers to adoption of Blue‐Green Infrastructure for urban flood risk management. Journal of Flood Risk Management, 11, pp.S960-S972.

Vazquez-Prokopec, G.M., Vanden Eng, J.L., Kelly, R., Mead, D.G., Kolhe, P., Howgate, J., Kitron, U. and Burkot, T.R., 2010. The risk of West Nile virus infection is associated with combined sewer overflow streams in urban Atlanta, Georgia, USA. Environmental health perspectives, 118(10), pp.1382-1388.

Zhao, W., Beach, T. H., & Rezgui, Y. (2017). Automated model construction for combined sewer overflow prediction based on efficient lasso algorithm. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 49(6), 1254-1269.