FO82AA0EFAB46

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 36

Literature Review

Amongst all the issues of improving sanitation that the governments are struggling with, slum

developments bring in a unique combination of logistical, financial and social complexities that

in turn may negatively affect the public health, the environmental sustainability and finally, the

overall well-being of the people in the impoverished areas. In tackling these issues, the literature

views each constrained setting of slums necessitating best-suited approaches that embrace the

innovative, adaptability, and cost-effectiveness factors and which must include the practice of

community participation. The literature review here conceals the comprehensiveness of the

contemporary knowledge, techniques, and gaps in these contexts, thus forming a ground for our

study to focus on an auxiliary denouement in the realm.

The key theme of investigation which has come up in recent studies on the question of how to

introduce infrastructure facilities in slum areas is logistical complexity. Earlier research (Article

written by Sayed et al. 2021 and Solanki and Salgude, 2019) has pointed out how hard it is to

navigate the pre-planned narrow road networks in slums, which are the major factors of such

projects being unable to commence making a sour environment that is not conducive to

construction. The problem of insufficient or with unprecise infrastructure maps which couldn’t

be used for the planning makes the mentioned problems even worse causing difficulties with

finding optimal routes and methods for excavation and installation works.

The financial restraints are another undergoing factor that determines whether the upgrade

project can be realized and what scope it can have. The literature continues to stress the fact that

there has to be resource and cost-effective strategic and project implementation methods that will

minimize expenditure, but maximize the impact on the people (Zengin & Işik, 2019). This is

especially concerning since the high costs of manual excavation methods that entropy normally
occur due to methodological constraints mentioned earlier are the case. Furthermore, too much

file Burden of slum dwellers mean who living on informal or disadvantaged habitat, is not

enough wording. In affordability step slum dwellers affordability is one of the crucial elements

of the plan and implementation.

The community resistance aspect in infrastructure development in urban slums as a matter of

social dimension among the many local efforts to address these situations became a topic of great

interest to current practitioners. The confusion often caused by an erection including the noise,

dust and the interim loss of passage ways may sometimes lead to strong resistance from the

community locally affected. Also, it is imperative to note that the longer benefits of sanitation

upgrade can be sandwiched between the reality of constructions thus making it to be a concern to

residents enduring the nuisance. Research has put forward the recommendation that the

involving of local community members in the planning and realization of the slum upgrade plan

could help to overcome the resistance and to enhance among them the sense of belongingness

and cooperation.

In the middle of these adversities, the use of multi-objective optimization models that use these

compiling algorithms, genetic algorithms (GAs) as well as Non-Dominated Sorting Genetic

Algorithm II (NSGAII), that is what the literature explores. Such models are advanced tools that

allow them to ensure the balance among the different project objectives which could be for

example cost effectiveness, livability and neighborhood livability.

The OP studies allow displaying Pareto contrasts that represent the specific unfavorable

outcomes of distinct objectives. As a result, such decision-makers are equipped with something

being more than just a simple tool; they have something giving them the power to handle

complexities which emerge in slum upgrading projects.


In essence, what literature gives is the multi-layered obstacles of upgrading the drain system in

slum areas that encompass the logistic, financial and social barriers. However, after huge

progress in the implementation of techniques designed to address these issues, some holes persist

in the application of those solutions in the actual environment. The presented research is an

attempt to fill the void of experience and knowledge which are lacking in effective slum

upgrading strategies by delineating a holistic framework that considers logistical and financial

difficulties imperative to the slum environment and fully incorporates servicability and

participation of communities in the cyclical improvement. This research has effectively used the

most advanced optimization technique in order to help living conditions in the slum improving

while reflect the intricate nature of the challenges and opportunities associated with sanitation

network upgrading projects properly.

Methodology

Model Formulation

The construction of the prototype has been hinged on the creation and use Multi-Objective

Optimization Model that has been developed to address the issues of maximizing the

Serviceability Index (SI) and minimizing the Cost Index (CI) simultaneously. The first objective

(quality and accessibility) and second objective (making it affordable financially and subject to

long-term planning) are the two key factors in dealing with the complexities involved in

upgrading sanitation networks in slum areas. This subsection addresses the modeling component

of the study through the demonstration of mathematical and computational backgrounds that are

used to achieve those goals.


Objectives

1. Serviceability Index (SI): The SI, having been invented recently, is a startling metric deployed

to rate the effectiveness and frequency of services' delivery in an improvement of sanitation

networks. It is based on the number of the population which gets access to improved sanitation

as the stated figures are for segment to segment completion at any given time. This project index,

which serves as a public health indicator, shows how the interventions affected the living

conditions of the inhabitants of the designated slum area.

2. Cost Index (CI): The CI, as opposed to the D&A cost, incorporates the comprehensive

expenditure that equates to the entire projects such as the direct and indirect costs. The cost

includes such things as material cost, labor cost, equipment use, and all the other expenditures

from the project start to end. In addition, CI values effect of money, which means that discount

rate is calculated in order to evaluate financial limitations related to the project timeline.

Decision Variables

The model's decisions involve the choice of sequencing and time allocation to the upgrading of

different sections of the treatement and sewage pipes. They are the deciding factor on the urban

area's development path, determining the order in which the road segments are upgraded and

hence the SI and CI as they directly impact road quality. The algorithm is designed such that
entire checkout alternatives are considered enabling the optimization tool which can then

discover the most effective process.

Constraints

Many of the constraints which are vital in the design are the factors that are responsible for the

feasibility and practicality of the formulated solutions for the optimization problem.These

include:

• Technical Constraints: This will help achieve alignment with engineering and safety

benchmarks to guarantee the work is carried out in a professional way.

• Logistical Constraints: These involve in an assessment of the capability to reach up any portion

of the road and the possibility of using necessary materials and machines.

• Social Constraints: They take into account the general community needs and any potential

dissenting opinions but seek to minimize this intrusion on the residents' daily life.

Mathematical Formulation

The model is designed to reproduce reality in a mathematical model and integrate the objectives

and constraints with the multi-objective optimization approach.The optimization seeks to:

Maximize SI (X ), Minimize CI ( X)

Subject to:

Technical constraints

Logistical constraints

Social constraints
where

Decision variables of X specify the list and a certain order of project’s tasks.

Optimization Methodology

For the solution of this complicated multi-objective programming problem, the study puts the

state-of-the-art of computational methodology into practice, including genetic algorithms (GAs)

and Non-Dominated Sorting Genetic Algorithm II (NSGAII). They are brilliant in solving the

multi-objective problems, by showing the trade-off values among controlling the spread of the

disease and assuring people’s confidence with the community, as exemplified by the Pareto

optimal solutions set.

Formulating the model thus plays a paramount role in searching for the solution that facilitates

the design of a practical, functional and community - friendly system for the improvement of the

sewage network in the informal settlements. This model takes into consideration the challenge of

satisfying both the need to maximize serviceability and the need to plan cost management, which

serves as the cornerstone to evidence-based and effective decisions concerning infrastructure

development in the urban settings, where poverty is prevalent.

Problem Formulation

Objectives
 The primary objective is to reduce the overall cost of upgrading sewage pipes while

maintaining or improving system performance and longevity.

 Another key objective is to minimize the duration of the construction phase to mitigate

disruption to residents and expedite the availability of the upgraded sewage system.

 Ensuring that the upgraded sewage system has a long operational lifespan is crucial to

minimize the need for frequent maintenance or future upgrades, thus reducing long-term

costs.

 Enhancing the performance of the sewage system is essential to meet the community's

capacity, efficiency, and reliability needs.

Constraints

• Aged residential areas frequently can't fit excavation machinery, leading to the need for

effective site planning and the use of the available space with the purpose of minimizing project

disruption.

• Modern cities often continue to have older communities, and some areas within may have

historical significance. Hence, there will be restraints to prevent construction activities that

would remove or destroy cultural heritage and architectural integrity.

• During the upgrading program, the new sewer system design and implementation, the in-situ

system, and other relevant features such as the existing pipelines, sewage points, and stairways

must be considered.

• Adherence to the adjacent area rules and norms guiding the sewage system upgrading, such as

environmental norms, safety rules, and building laws, are be-all.

Assumptions

• The assumption is that data indicating things like existing infrastructure maps, community
population data, and construction regulation are all available to our case study area.

• Since the community settings, population concentration, economic status, and social situation

will likely remain relatively unchanged, the project implementation period is assumed to be

stable.

• Access to a quantum platform capable of quantumly performing the computationally complex

quantum genetic algorithm is ideal.

Metrics

Cost: Estimated total project cost, which consists of the material cost, the labour cost,

consumption of equipment and any other expenses related to the construction process.

Time: The construction phase goes from the project kickoff to closure, taking into account the

scheduling, material availability, and weather conditions.

Longevity: Designated lifetime of the enhanced sewage system for maintenance procedures

through material durability, construction quality, and maintenance requirements.

Performance: Analysis of the key system performance indicators like implementation capacity,

operational effectiveness, and reliability parameters that vary considering the different situations.

Quantum Genetic Algorithm (QGA) Development

1. Quantum Representation

Encoding: Represent candidate solutions (chromosomes) using qubits, where each qubit

corresponds to a decision variable or gene.

Superposition: Utilize superposition to explore multiple candidate solutions simultaneously.

Measurement: Employ quantum measurement to collapse the qubits into classical states, yielding

potential solutions.
2. Genetic Operations

Selection: Employ quantum-inspired selection methods, such as tournament or roulette wheel

selection, to choose parent chromosomes for reproduction.

Crossover: Apply quantum-inspired crossover operators to exchange genetic information

between parent chromosomes.

Mutation: Utilize quantum-inspired mutation operators to introduce diversity into the

population.

3. Quantum Evolution

Evolutionary Cycle: Execute the evolutionary cycle, including selection, crossover, and

mutation, iteratively to evolve the population towards optimal solutions.

Fitness Evaluation: Evaluate the fitness of each chromosome in the population based on

predefined objective functions.

4. Convergence and Termination

Convergence Criteria: Define termination conditions based on the convergence of fitness scores

or a maximum number of generations.

Output Generation: Extract the best chromosome(s) from the final population as the optimal

solution(s) and return relevant output metrics.


Framework Schematic Diagram

Preparing a schematic sketch encompassing the pipelines of the executing framework for

optimizing the emplacement of sewage pipes for the older cottage suburbs concerns applying

together the process flow, data inputs, algorithm modules, and result outputs. A framework

within the schema is a structure involving different significant components.

At first, this approach combines data inputs, including information on the existing

infrastructures, such as new installation of sewage pipes, new drainage piping, or the current

sewage pipe network's replacement, repair, or maintenance; the material of the pipes; their

diameter; and their locations and layout. Besides, the information is classified by the population

subgroups data, housing density, as well as the spatial location of the old apartments. Besides,

construction behaviour rules, zoning laws, and historical preservation factors are built into the

framework and integrated into the framework design.

The heart of the framework resides in the Quantum Genetic Algorithm (QGA) part, which is

composed of several major submodules. The initialization phase sets the quantum chromosome

population into an intermediate state at first and it subsequently moves towards better quality

values as the result. Consequently, the evolutionary cycle of QGA proceeds processes like

selection, cross-over, and mutation, which are used in an iterative way to facilitate the solution's

improvement. The fitness function of the evaluation rounds each chromosome based on a given

objective while tracking the convergence controls the criteria that end the optimisation process.

Additionally, the framework consists of a Construction Time/Cost Analysis module that is

required to do modelling and analysis of the construction activities within time and budget

constraints. It analyses labour market conditions, time management, and material acquisition
issues. It is impossible to do this without considering many variables such as those highlighted.

Furthermore, the combination of the Optimization Model and the QGA (Quantum Genetic

Algorithm) promotes a cyclic model for the optimization of construction schedules and resource

allocation. Thus, the integration loop results in the generation of optimised construction visas of

up materials and cost estimates that are based on the QGA answers.

The timetables for practical construction, where the consolidated resource allocation plans, as

well as cost estimates, are manufactured, form a vital part of the framework. The outputs, which

are helpful and practical, provide the basis for informed decisions for project stakeholders

participating in the sewer pipe upgrade. These can enable them to make cost-effective,

environmentally friendly, sustainable decisions.

Also, in addition to all that is stated in the previous sentence, a feedback mechanism that permits

the assessment of the efficiency of the upgraded pipeline will be constructed. It covers systemic

aspects like durability, scope, and running efficiency. Finally, the feedback from the performance

evaluation is further integrated into the future iteration of the optimization phase, which in turn

will refine the framework for continuing the betterment process.


Case study

For the case study, we will focus on a portion of the drainage system in Hohhot, located in the

Inner Mongolia Autonomous Region of Northern China. The selected watershed covers an area

of 1,398 hectares, with approximately 71% of impervious surfaces, including buildings, roads,

and industrial districts. The remaining 29% comprises permeable surfaces, such as green spaces

and agricultural areas. Land use within the watershed is classified into five categories:

residential, industrial, agricultural, green space, and others, including municipal facilities,

commercial areas, institutions, and hospitals.

The city's planned land use map for the 2020s indicates a shift towards increased residential and

"other" category spaces, replacing industrial districts with residential areas. This transition
presents an opportunity for implementing Low Impact Development (LID) designs to mitigate

the impervious surfaces in the region and improve stormwater management practices. The area

experiences a cold semi-arid climate characterized by dry winters and hot, rainy summers, with

an annual mean precipitation of approximately 396 mm.

The existing drainage system in the watershed dates back to the early 1970s, with limited

upgrades conducted over the years. The pipeline coverage rate in most areas is relatively low,

contributing to drainage issues during heavy rainfall events. The watershed is divided into 53

subcatchments, with runoff generally flowing from north to south. The total length of pipelines

in the system is approximately 36.6 km, with the River Xiao Hei and its tributaries serving as the

primary receiving water bodies for drainage discharges.

Despite efforts to improve water quality, including a river restoration project in 2013, the

drainage system still faces challenges. Approximately 70% of untreated water was discharged

directly into the river before 2007, leading to significant water pollution. While joint pluvial and

fluvial floods are unlikely, there is an increasing frequency of flooding within the watershed due

to the limited capacity of the drainage system.

In response to these challenges, decision-makers want to upgrade the drainage system to achieve

a 3-year return period service level. Short-term plans focus on pipeline capacity expansion, while

long-term strategies explore the potential of infiltration-based LID measures, such as porous

pavements, green roofs, infiltration trenches, and vegetation swales, for integrated stormwater

management.

The project aims to optimize the combination of LID and pipe measures to upgrade the drainage

system while considering both physical and economic performance criteria. Genetic algorithms
will be employed to generate optimal strategies that balance system performance improvements

with project budget constraints.

Construction Time/Cost Analysis Model

The construction duration/cost analysis model is being developed to estimate the duration of

construction projects and the related costs concerning sewage pipe upgrading in older residential

communities. It starts with identifying activities like site preparation, excavation, pipe laying,

backfilling, compaction, testing and restoration. For each corresponding activity, the duration

and cost are estimated according to the appraisal, the offered workforce productivity, the

availability of equipment, the covered material costs, and other overhead expenses. Next, the

model gets trained on achieving project completion as fast and cost-effective as possible, does

the activity sequencing using algorithms or techniques focused on reducing the critical path

activities time, allocates the resources in a balanced way, and mitigates the project

risks. Simulation instruments or mathematical tools are designed to evaluate the construction

sequences and dig out possible effects on project time, cost, and budget. Comparison with

historical data or indicators is to provide support for the model and make it reliable. On its part,

the sensitivity analysis examines whether any change in the inputs leads to any change in the

outputs. The outcome is a recommendation and plan on the sequence of the construction phase to

improve duration and cost and meet performance standards.


Optimization Model

The QGA (Quantum Genetic Algorithm) based Optimization Model is utilized in an iterative

way to align the projects of sewage pipe upgrading in older residential areas to have optimal

construction order. The QGA passes through many generations of candidate sequences of

construction processes and considers such features as activity dependencies, resource

toughening, and performance constraints. Each iteration produces a pool of potential sequences

to construct an entire chromosome for each sequence that encodes the construction

activities. The QGA will assess the efficiency of each chromosome in turn, according to the

objective functions specified beforehand, which include estimated project cost and duration;

however, further system performance criteria remain in place, including the system's reliability

and longevity. Unlike traditional genetic algorithms that utilize classical operators like selection,

crossover, and mutation, the QGA employs quantum-inspired evolutionary operators such as

quantum selection, probability cross-talk, and mutation that enable its population to converge
upon optimal solutions in terms of the low budget and time requirements as well as performance

parameters. This optimisation process keeps repeating until there is a convergence of the target

solution or up to the maximum number of generations is attained. Getting checked against

conventional methods ensures that using the QG algorithm will succeed as an innovative

construction sequencing solution that can solve sewer pipe repair task-specific problems in such

residential communities' old districts.

2 Framework for describing sewer system control

Controllable devices, such pumps, gates, valves, and the like, are known as actuators. The

variables being modified are known as actuators. The control group has the ability to change

these factors, which are called manipulated variables (MV). Measured variables (MeV) are the

variables that are captured by sensors like level meters, flowmeters, rain gauges, and so on.

Controlled variables, abbreviated as CV, are those that can be managed. Consider the volume of

water in a storage tank or the pressure within a pipe as examples. We talk about setpoints when

we want the controlled variables to have certain values. Larsson and Skogestad (2000) state that

control is the act of modifying the available degrees of freedom to help achieve a satisfactory

system functioning. This is the procedure for figuring out how to operate the sewage system's

actuators in a manner that satisfies the control's setpoints and rejects disruptions. A disturbance

to the system is any new input that affects the controlled variables in a way that the

uncontrollable variables do not. In the context of a sewage system, this includes both raw sewage

and precipitation runoff. As an example, let's say we want to control the volume of water in a

tank. In this case, you may reject a disturbance by employing the right control mechanism to

change (manage) the outflow from the tank. So, even if the inflow into the tank is changing,

which is a disturbance to the system, the controlled variable—the water level—remains at its
fixed setpoint value. The eighth definition of a control method is a control legislation or process

that calculates the adjustment of the controlled variable. The PID5, MPC6, and rule-based

control schemes are only a few examples. In addition to storing the control technique, the

controller computes the actuator's corrective action. A control loop consists of one system or unit

that has to be controlled, together with a sensor, a controller, and an actuator. A control loop is

said to have a single input and one output when it is possible to keep a controlled variable at its

desired setpoint by changing a single manipulated variable. Therefore, it is called a SISO control

loop since it only has one input and one output. Adjustments to both control loops must be made

simultaneously and in coordination if there are two or more control loops that interact with each

other in order to obtain the needed setpoints. Multivariable control, sometimes called multiple-

input, multiple-output control (MIMO control), is another term for this layout of controls. All of

the actuators in a control loop are controlled by a single control mechanism when multivariable

control is used. Centralized control is important when thinking about the layout of the existing

control system. The existence of a single optimum controller that stabilizes the system's

processes and appropriately coordinates all the modified variables is a defining feature of

centralised control, as stated by Larsson and Skogestad (2000). It is typically impractical to build

a controller of this sort for large systems, however. The alternative is a hierarchical or

decentralized or distributed system of "blocks" that contain the control logic (Larsson and

Skogestad, 2000). You have the option of doing it vertically or horizontally. Local controllers,

often called SISO control loops, are typically given more authority to oversee the sewage system

as a whole in a decentralized fashion.

2.2 Timescale dependent control hierarchy


These days, a spatial decomposition is the cornerstone of modern sewage system management,

which is generally approached horizontally. This shows that the control difficulty is subdivided

into smaller difficulties that are handled locally. This is known as distributed or decentralized

control when seen from a systems perspective (see the section preceding this one), and local

control when dealing with urban drainage. However, this kind of decomposition may not be the

best choice if the subproblems are really interacting. A control system that is too complex and

hard to comprehend might be the outcome of optimizing it using logic switches or cascading

feedback loops. As shown in Figure, another possible approach is to vertically decompose the

control problem. By delegating certain tasks to various levels of the hierarchy, such as

monitoring setpoints, decoupling interacting control loops, and determining setpoints, the

complexity of the control problem is reduced and the interactions are regulated in this way.

When it comes to controlling wastewater treatment plants (WWTPs), the idea of plant-wide

management has been around for a while (Olsson and Newell 1999). It is considered beneficial

for urban water management to incorporate control of the sewage system in the hierarchy while

working toward integrated control between WWTPs and sewer systems. However, the sewage

system cannot operate with continuous processes that may be described by linear models, in

contrast to wastewater treatment plants (WWTPs). Since the sewage system's dynamics are

unpredictable and transient, they defy linear modeling, making an instant transfer of the

framework impossible. Hence, before the framework can be transferred, it has to be altered. A

control hierarchy for sewage systems that depends on time was established in Paper I. This

hierarchy was developed by reviewing existing control hierarchy frameworks and analyzing

three separate control systems for European sewage systems. What sets this approach apart from

previous offerings for sewage systems is its focus on the communication of information across
hierarchies and the frequency of computation updates. You may see a suggested schedule in

Figure 2. The response and transient times of the system, in addition to the number of layers,

determine the ranges of possible action frequencies at each level. This is because the amount of

layers in a sewage system determines how these frequencies could change from one system to

another. The layers are linked by the information that is passed down from higher levels to lower

ones.

Managing the objectives is the fourth layer. The highest level of the hierarchy is occupied by the

management of objectives layer. The sewage system's overarching scope and operational

objectives are defined at this tier. Along with operational goals, it may also think about

regulatory necessities. The constraints and expenses used by the goal function are supplied

herein as a component of the optimization procedure. Possible causes of fluctuations in the

restrictions and expenditures include seasonal changes, such as entering or exiting the bathing

season or undergoing a diurnal pattern of rising or falling electricity costs. Even though no

existing sewage system control system incorporates such a layer into its structure, it is

nevertheless considered an important component of the control hierarchy and is so included here.

For example, CLABSA7, Barcelona's wastewater utility, has found that the two main objectives

—reducing CSOs and preventing floods—are mutually exclusive in their current context. On top

of that, it still can't provide dynamic prioritizing by changing the relative importance of various

goals. The primary focus of the aim function should always be the prevention of floods. Now

that they have their control system in place, they are thinking about adding the management of

objectives layer. Depending on the predicted rainfall, they would be able to move between the

two goals in terms of priority (Paper I). This would transform the optimization process

formulation from continuous to event driven. The third level is optimization. In addition to
deciding on the desired setpoints or setpoint trajectories for the control layers, the optimization

layer is in charge of discovering the optimum operation. The statuses of a sewage system may

change rapidly and often during a rainstorm. Regardless matter the medium, this process always

occurs in a sewage system. Consequently, there is a temporal restriction on how long a setpoint

can provide optimal performance. Because of this, you'll have to adjust the setpoint often;

alternatively, optimization might give you a series of setpoints instead of just one big number.

The second layer is control coordination. As a sublayer of the optimization layer, a coordinating

control layer may exist. One name for the supervisory control layer in plant-wide control systems

is the coordinating control layer. However, when discussing sewage system management, the

term supervisory control takes on a new meaning; here, it describes the level of automation

included in the control system (Schütze et al. 2004a). This is why the layer is now known as the

coordinating control layer.

One must have the coordinating control layer if the control loops are interdependent or if the

controlled or manipulated variables are subject to immutable constraints. While actuator capacity

is often the limiting factor in a sewage system, flow or water levels at critical locations may also

be a problem. Interdependent control loops and limitation management are also the purview of

the coordinating control layer. The likelihood of this layer being used independently in real-

world scenarios is low. The alternative is to have the limitations preserved within the

optimization process itself, or to see it integrated directly into the controllers alongside the

regulatory control. The first level consists of regulatory oversight. The controller situated at the

lowest layer is the regulatory control layer. It is the job of the regulatory control layer to reject

disturbances and make sure that the setpoints or trajectories are followed. The operations at

higher levels are discrete from those at lower ones because each layer runs on its own timeframe.
Larsson and Skogestad (2000) state that this means the system can't achieve really optimum

functioning since finding the appropriate setpoints would be an ongoing process. An important

part of a steady-state system is the determination of the timeframes. Because the determination

of the timeframes is of paramount importance for a system in a steady state, it is reasonable to

assume that this will be especially true for a system of a transitory nature, like the sewage

system, this aspect is related to the frequency of the disturbances operating on the system and the

rate of change of the restrictions and costs (e.g., the number of orders, the cost of materials, the

available personnel, etc.). Choosing a chronology becomes a delicate balancing act, as this

sentence implies. One option is to approach the ideal behavior by often updating setpoints; this

requires speedy computation of setpoints, which in turn puts demands on the techniques utilized.

In addition, the setpoints might be set using more efficient but time-consuming techniques,

which would need more frequent updates. How the coordinating layer works is one of the

subjects up for debate. An interdependent control loop's layer is the one in charge of handling

constraints and decoupling them. One may make the case that the optimization or rule-based

control that is a part of the regulatory control layer really does these tasks rather regularly.

Consequently, it is necessary to remove the layer. On the other hand, it's possible that a generic

control hierarchy may become limiting if it were only based on the recorded cases of sewage

system control in the literature. Quantitative aspects of CSO are now the focus of most control

efforts for sewage systems. Also, studies are looking at CSO's qualitative features (e.g.,

Vanrolleghem et al., 2005; Langeveld et al., 2013). Therefore, in order to keep the maximum

amount of flexibility, the hierarchy is kept as general as feasible. With any luck, this will also

help it handle the sewage system's control hierarchies down the road. The time-scale dependent

framework's ability to graphically represent control structures in a manner that permits


comparison and contrast is one of its most significant features. Not only may this be helpful

during the design process, but it will also be helpful throughout the rest of the system's

development. Documenting the control structure will help the utility company with two things:

first, comparing the system to others; and second, maintaining and developing the control

system, which is an important but often neglected issue after implementation. We provide a

complete inventory of the control loops and methods used in the control system, together with

the layer in the hierarchy at which these control approaches are implemented, by using the

framework in a thorough way. This data is really helpful for a number of reasons, including

figuring out whether updates are necessary and getting a head start on examining the control

system. Furthermore, it serves as an effective means of communication by facilitating future

collaboration among specialists from various disciplines who will be designing control systems

for sewer systems (e.g., hydrologists, environmental engineers, chemical engineers, electrical

engineers, and control engineers), all of which often have their own specialized jargon and

definitions.

Designing sewer system control

As operators have gained knowledge about the system's dynamics and interactions (via the

HOFOR internal papers that detail the controls), the current rule-based control has been refined

and enhanced over time. The water levels in both the P2 and P3 basins, as well as the water

levels further downstream, determine how these pumping stations empty the basins. P1 is the

wastewater pumping station that uses gravity to lift the wastewater to a higher level, where it
may continue to flow towards the WWTP. A PID9 controller is used at point P1 for control. The

controlled variable, u, is limited by a set of constraints imposed by a chosen control mechanism

in conjunction with the PID control. Here, the controlled variable is the discharge from the

pumping station. Consequently, this measure has been implemented to ensure that the

downstream system can handle the flow from the pumping station. The selection helps choose

between four options, one of which being the minimum value of the PID controller's output. All

three options are based on conditions farther downstream. The first two are trying to limit the

flow to CSO structure UO38 to reduce the likelihood of an overflow scenario based on level

readings made at or near UO38. The part of Copenhagen's wastewater system that is being

studied here is somewhat small. Hence, it is often assumed that the catchment region has a

consistent distribution of rainfall. The amount of time the system can concentrate is mostly

determined by runoff routing.

3.2 Designing a regulatory control layer

According to the data in Figure, the most basic level of control is the regulatory control. It

ensures the setpoint trajectories are reached by adjusting the actuators. To lessen the disparity,

the actuators are adjusted based on the computation of the difference between the setpoints and

the observed values (as shown in Figure 4). Someone else may call this "feedback control" as

well. Determining the measurement locations, regulating objects, control techniques, and

setpoint origins is essential for constructing the regulatory control layer.


When using a gain matrix, many classical control theory-based tools need a transfer function

model to depict the relationship between the system's inputs (𝒖𝒖𝑖𝑖) and outputs (𝒚𝒚𝑖𝑖) (Seborg et

al. 2011). The transfer function gain matrix is preferred over the steady-state gain matrix for

studying the operation of a sewage system. Because system dynamics are a crucial component

that must be considered. So, the very first step is to transform the VT model into a gain matrix

for the transfer function. The concept of a gain matrix transfer function is shown in Figure 7,

where 𝐺𝐺(𝑦𝑠) is the transfer function gain matrix that represents the system dynamics in the
Laplace domain (frequency domain). Because of the abundance of resources for controllability

analysis in the classical control toolkit, work in the frequency domain is preferred. This is the

rationale for the decision.

Results

In order to find the optimal network model, the training step included testing several artificial

neural network (ANN) topologies with different numbers of hidden layers. According to the

findings, the ten-neutrino network was the most effective for the modeling tasks. We succeeded

after five rounds of iteration. Using the performance assessment tools, it was found that the

network structure (4-10-1) had the best Regression Value (R = 0.9) and the lowest Mean Square

Error (MSE = 0.4). As a consequence, the Regression Value is maximized.


In the graphic, we can see the network model's regression values shown. For the training
instance, the R-value was 0.997; however, since there was insufficient data for the validation and
testing instances, the R-value was not a number (Nan). R, the overall regression coefficient, is
0.994, however, so it's clear the network fits the data quite well.

Forecasting MSW in chosen municipalities

The quantity of trash that will be produced in eight of the selected cities is projected in the
accompanying chart. A technique for gauging future events based on historical data is known as
time series forecasting. One simple approach that does not need a mountain of data is time series
forecasting. The values that the models predicted were compared to the original data that was
used. It is reasonable to say that the result is satisfying since the actual and predicted numbers
were so close to each other. The forecast produced an acceptable outcome according to the
performance indicator assessment, with regression values of 0.9 and a mean squared error (MSE)
of 0.43.
Furthermore, in order to get the most optimal results when predicting the total quantity of trash
created using the historical data shown in Table 2, many ANN topologies were examined. The
eight-neurolencer network outperformed the others in terms of model accuracy, and this was true
across the board in the experiment and training process, where each hidden layer was evaluated
in turn. At last, after 67 rounds, it was mastered. We found that the network structure (2-8-4) had
the lowest Mean Square Error (MSE) and the best Regression Value (R) when we utilized the
performance assessment tools to evaluate the model. Two distinct models were derived from the
investigation. The first model estimated annual trash production for a handful of randomly
selected ian cities using socioeconomic variables. Based on data collected from 1999 to 2015, the
second model projected the overall amount of trash created in . It was observed that the quantity
of rubbish created was affected by the selected socio-economic characteristics. Furthermore, the
second model proved that the quantity of trash generated in is directly related to the rate of
population growth there. In order to assess the efficacy of the models, two performance metrics
were employed: regression (R) and mean square error (MSE). In order to find the optimal match,
the top-performing models were selected. The regression and MSE values for the best model
(based on socio-economic factors for garbage production) are 0.994 and 0.43, respectively.
These numbers came straight out of the model. The regression and MSE values for the best
model for total waste based on historical data were 0.993 and 0.2, respectively. The data was
used to get these values.

Discussion
The inquiry led to the addition of 150mm and 250mm pipes to the network. Figure 3 shows the

altered structure of the Water Distribution Network in the study region; the final network

configuration consists of pipes with four different sizes. Figure 2 shows that the prior design just

considered 100 mm and 200 mm diameters.

Based on the desired parameters and the optimization's relevance, the results are presented in a

statistical analysis that was deemed essential for determining the optimization's importance.

How much: A statistical comparison was made between the pre- and post-optimization costs of

pipes using non-parametric analysis, which is necessary since the populations in question do not

follow a normal distribution. The Wilcoxon test type and a 95% Monte Carlo confidence interval

were used for the comparison. Tables 1.0 and 2.0 display the outcomes of this calculation.

Assuming no change from the initial pipe cost in our sample previous to Network optimization to

the final pipe cost after optimization, we find no evidence to support the null hypothesis. The

"Positive Ranks" have a much greater total than the negative ones when using two

interconnected sampling techniques. If pipe costs were higher before optimization than after,

then the phrase "positive" describes this situation.

Given that the observed data do not support the null hypothesis being true, it is clear from

examining the test statistic summary that P=0.007 < 0.05=α. Hence, the hypothesis must be

rejected. But the Monte Carlo confidence interval showed that, with 95% certainty, the lower

limit was 0 and the upper bound was 0.259. This finding lends credence to the idea that the two

prices are somewhat close to one another. Consequently, there is insufficient evidence to reject

the null hypothesis.

A statistical analysis of the node pressures before and after optimization shows that the average

pressure is 12.4 after optimization, up from 12.3 previously. The pressure that existed before to
optimization was different from this. Assuming "there is no difference between the two pressures

before and after optimization of the water distribution network," we will use the null hypothesis.

According to the other theory, "there is a difference between the pressure before optimization

and after optimization." Table 3.0 shows that after adding up all the negative rankings, the

"Positive Rank" comes out on top. "Zero tie" stresses that the two pressures being compared are

completely unrelated, but "positive" implies that the Nodes' pressures were higher after

optimization than they were before.

When the value of P is less than the significance level of 0.05—that is, 5%—the level of

significance of 0.05 is chosen, according to the summary of the test results presented in Table

4.0. With a p-value less than this level, we can say with confidence that the result is statistically

significant; because we can all agree that the null hypothesis should be rejected, we may accept

the alternative hypothesis, which says that there is a difference.

Flow rate: The population data are not parametric, as shown by the testing of normality, which

yielded a P-value of 0.001, which is less than 0.05. With the null hypothesis reading "There is no

difference between the two flow rates with respect to pre and post optimization," a statistical

comparison of the flows in the pipes before and after optimization was carried out in line with

the Wilcoxon Rank test. You can see the results of the Wilcoxon Rank test in Table 5.0, and it

shows that the "positive rank" sum is higher than the "negative rank." This means that the

optimization method reduced the flow rate from its pre-optimization state to its post-optimization

state.

After considering the test statistic summary, which is P=0.003<0.05=α, it is clear that the

observed data support the null hypothesis being true, therefore rejecting the hypothesis. This

strongly suggests that the two flows are separate from one another. When all factors are
considered, the alternative hypothesis, which proposes that the flow after optimization differs

from the flow before to optimization, has a good chance of being true.

If optimization has no effect on the head loss, then the null hypothesis asserts that the two sets of

estimates are identical. In Table 7.0, we can see the statistical rank test findings, which show that

the "Positive Rank" total is far more than the sum of the negative rankings. That is, the Headloss

in the links increased after optimization as compared to pre-optimization.

The facts that have been seen clearly do not support the null hypothesis being true, when we

evaluate the test statistic summary, which is P=0.001<0.05=α. Hence, the hypothesis must be

rejected. This proves that there is a difference in the Headloss. The Headloss after optimization is

different from the Headloss before optimization, albeit the change is not statistically significant.

Finally, we shall accept the null hypothesis.

Following the optimization of the network, the study of the results produced may be used to

draw a number of conclusions. These include the following: 1. From a mathematical standpoint,

the program that is written in MATLAB can provide the relevant diameter combinations for each

connection. What this implies is that a comparable pipe may deliver the best possible outcome at

each of the connections that are visible.

2. At the conclusion of the final run of the solver, the chosen diameters and the hydraulic

analysis result of the network's node and link values are shown. In accordance with the

limitations, all of the velocity and pressure head data fall within the acceptable range.

3. The pressures (m) of the hydraulic results of the nodes after the optimization process have

been found to be higher than those obtained before the optimization process. It may be seen from

this that the pressure rises as the flow rate increases while the diameter of the pipes inside the

nodes decreases. In the results of the pre-optimization process, the highest pressure was found at
node J2, with a value of 14.79 meters. However, following the optimization process, the highest

pressure on node J13 was 15.71 meters. For the former, the minimum pressure in the Network is

7.59 meters, whereas for the latter, it is 9.42 meters. This pressure is measured at J52. This

demonstrates that the pressure is at its lowest point at the same node, but the pressure is also at

its highest point at a different node. A statistical comparison utilizing two related sample

methods of non-parametric distribution (Wilcoxon Signed Rank Test) reveals that there is a

significant difference between the pressures that were compared. To summarize, the increase in

pressure makes the network more hydraulically efficient. This is due to the fact that during

periods of peak demand, the pressure will be adequate to meet the demand, and the issue of a

scarcity of supply will be addressed. In light of this, the minimum pressure of 9.42 meters that

was attained after optimization is enough.

4. Taking into account the statistical result, there was a slight increase in head loss after

optimization; however, when direct comparison is made, it is observed that the maximum head

loss occurs at Link P95 (7.62m/km), in contrast to the same Link before optimization, which was

21m/km. This indicates that the head loss decreases by a factor of three at Link P95, and it is

closer to the maximum head loss of 7.7m/km of pipes ranging from 100mm to 400mm in size.

Once again, the minimal head loss at connection P92 was 0.05 meters per kilometer after

optimization, while it was 0.08 meters per kilometer before optimization.

5. Since the goal of optimization is to reduce the overall cost as much as possible, the total cost

that was previously estimated was $716501.868 for 100mm and 200mm pipe. After

optimization, the total cost fell by 7.15 percent (as a direct comparison). The use of statistical

analysis reveals that the cost of optimizing the network is, to a certain extent, distinct from the

cost that existed prior to the optimization process, which is coupled with the enhancement of the
hydraulic parameters (properties) of the whole water distribution network.

Conclusions

In addition to the costs of operation, the capital expenditures and maintenance expenses

associated with a water distribution network are quite high. In order to achieve the best possible

design of water distribution networks, design engineers are seeking for novel approaches in

addition to the conventional techniques that are currently being used. During the course of this

investigation, a technique is developed that makes use of an optimization approach that makes

use of linear relationship of the model parameters (LOP), while the objective function ensures

that the capital cost of the pipes is minimized.

A water distribution system that was constructed by Network of Dukku is presented as an

example in this case study. Because the majority of the water came from bore holes, a single

period simulation (SPS) is utilized to run the solver rather than an extended period simulation

(EPS). This is because of the sources of the water.

Due to the fact that the least pipe size that is considered acceptable for any WDS is 100mm, the

optimization process is carried out with pipes that have diameters of 100mm, 150mm, 200mm,

and 250mm. This is despite the fact that the WDS is originally designed with two different kinds

of pipes, namely 200mm and 100mm pipes. Over the course of the direct comparison, there has

been a total reduction of 7.15 percent in the original estimate of the pipes. Statistical analysis, on

the other hand, reveals that there are very minor (minimum) significant variations between the

cost before and after they were optimized. When it comes to hydraulic qualities, the network is
more efficient, and pressures are able to meet peak demand at every portion of the distribution

network. This would alleviate the issues of water scarcity in Dukku if it were implemented.

It is essential to be aware of the fact that a node isolation strategy may be used, particularly in the

process of decision making for the purpose of enhancing the network, in the event that there is a

restricted quantity of money accessible.

References

Kerwin, S., & Adey, B. T. (2020). Optimal intervention planning: A bottom-up approach to

renewing aging water infrastructure. Journal of water resources planning and

management, 146(7), 04020044.


Seyedashraf, O., Bottacin-Busolin, A., & Harou, J. J. (2021). Many-objective optimization of

sustainable drainage systems in urban areas with different surface slopes. Water Resources

Management, 35(8), 2449-2464.

Chen, P., Zhao, W., Chen, D., Huang, Z., Zhang, C., & Zheng, X. (2022). Research progress on

integrated treatment technologies of rural domestic sewage: a review. Water, 14(15), 2439.

Calle, E., Martínez, D., Brugués-i-Pujolràs, R., Farreras, M., Saló-Grau, J., Pueyo-Ros, J., &

Corominas, L. (2021). Optimal selection of monitoring sites in cities for SARS-CoV-2

surveillance in sewage networks. Environment International, 157, 106768.

Su, B., Lin, Y., Wang, J., Quan, X., Chang, Z., & Rui, C. (2022). Sewage treatment system for

improving energy efficiency based on particle swarm optimization algorithm. Energy Reports, 8,

8701-8708.

Jayasooriya, V. M., Ng, A. W., Muthukumaran, S., & Perera, C. B. (2020). Optimization of

green infrastructure practices in industrial areas for runoff management: A review on issues,

challenges and opportunities. Water, 12(4), 1024.

McClymont, K., Cunha, D. G. F., Maidment, C., Ashagre, B., Vasconcelos, A. F., de Macedo,

M. B., ... & Imani, M. (2020). Towards urban resilience through Sustainable Drainage Systems:

A multi-objective optimisation problem. Journal of Environmental Management, 275, 111173.

You might also like