Prediction of Wastewater Treatment Plant Using Genetic Algorithm

Volume 8, Issue 1, January – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Prediction of Wastewater Treatment Plant Using

Genetic Algorithm
AKSHATHA KAMATH
Assistant Professor
Dept. of Computer Science and Engg.,
Ramaiah Institute of Technology Bangalore, India
Abstract:- A prediction of the wastewater treatment plant If waste water is not adequately handled, it can have a
(WWTP) using a genetic algorithm based on historical severe influence on the environment and human health.
data. In Iran, the majority of treated wastewater is used Damage to fish and mammal populations, oxygen depletion,
in agriculture. As a result, using waste water with poor beach closures, and other limits on recreational water usage
quality attributes might be hazardous to one's health. The are only a few of the consequences. The goal of waste water
effectiveness of the neural network model in predicting treatment is to remove as many suspended solids as possible
performance was investigated in this study. To find before releasing the effluent into the environment.
relationships in the data, exploratory data analysis was
employed and evaluated at a dependency level. The neural The amount of dissolved oxygen (DO) consumed by
network models' proper architecture was identified biological organisms when they degrade organic substances
through a series of training and testing stages. The ANN- in water is measured by the biochemical oxygen demand
based models were discovered to be a useful and reliable (BOD). The amount of oxygen consumed when a water
tool for predicting WWTP performance. The activated sample is chemically oxidized is known as the chemical
sludge method will be considered as a replacement for the oxygen demand (COD). Low concentrations of BOD and
semi-mechanical treatment system. COD can create eutrophication and destroy aquatic life,
lowering the DO of lakes and rivers. Water with high
Keywords:- Wastewater, ANN, Neuralnetwork, COD/BOD levels can be produced by municipal wastewater
Geneticalgorithm. discharge and industrial activities, necessitating thorough
treatment before discharge to protect the health of rivers.
I. INTRODUCTION
Biological approaches, such as adding microorganisms
The process of transforming waste water into water that and creating favorable circumstances for the organic matter
can be discharged back into the environment is known as to break down quickly, are widely used to treat BOD. Return
waste water treatment. Wastewater treatment is one of the activated sludge (RAS) is a bacterium source often used in
most frequent ways of pollution control in the United States, sewage treatment plants to remove organics in the water. To
according to the US EPA. The goal of wastewater treatment forecast the influence of changes in the parameters of an
is to speed up the natural purification processes. anaerobic B-2 system on its performance, an artificial neural
network was employed, and the weights of the artificial
Bathing, washing, using the toilet, and rainwaterrunoff neural network were optimized using a parallel multi-
all contribute to the formation of wastewater. Wastewater is population genetically algorithm. These researchers' findings
simply used water that has been contaminated by home, revealed that combining these methodologies can provide a
industrial, and commercial activities. According to the Safe useful tool for anticipating changes in anaerobic system
Drinking Foundation, some waste waters are more difficult performance. The researchers also found that these
to treat than others. Industrial effluent, for example, can be techniques can beexpanded to other treatment systems due to
difficult to treat due to its high strength. On the other hand, their adaptations to various environmental circumstances. A
dealing with domestic waste water is quite simple. genetic algorithm and an artificial neural network were used.
Apart from oxygen, water will most likely be the most
Given that not all trash makes it to wastewater treatment important resource for interplanetary space flight. All living
plants, there are a variety of ways in which wastewater can things require water to survive. Most long-duration space
cause environmental issues. Combined sewer systems (CSS) journeys will require some form of water purification and
collect residential sewage in the same pipes as storm water recycling in the future. The application of genetic algorithms
runoff in many cities, particularly older ones. Following to the optimization of waste water treatment design is the
severe rains, street gutters gather more water than the system subject of this research article The Activated Sludge Model
can contain, and a mixture of raw sewage and storm water is no.1 of the International Association on Water Pollution
released directly into the environment, resulting in a Research and Control (IAWRC), as well as the mathematical
combined sewer overflow (CSO). solution to the model, will be reviewed first. The four-stage
Bardenpho procedure is then described as the design that will
be optimized in this situation. Finally, the genetic algorithm
IJISRT23JAN444 www.ijisrt.com 1302

ISSN No:-2456-2165
is investigated, with future research objectives taken into
account.
II. MATERIALS AND METHODS
Perkandabad waste water treatment facility No. 1

(Figure 1), which is located on the southern bank of the
seasonal river, is one of Mashhad's most important municipal
wastewater treatment plants. This treatment facility has a
nominal capacity of 15200 cubic metres per day, and it serves
a population of 100000 people. The raw wastewater is treated
by passing through the waste collecting unit, aeration
lagoons, sedimentation ponds, execution pond, and
disinfection unit in this treatment plant, which uses an
aerated lagoon with complete mixing. The design of this
treatment facility is based on surface water discharge since
effluent from the Perkandabad No.1 treatment plant is
discharged into the river at specific periods of the year. Statistical indicators such as the correlation coefficient,
mean square error, and root mean square error were employed
during the results analysis to describe the error rate and
compare the simulated parameters. The parameters affecting
the treatment plant's performance were determined using
quantitative and qualitative data from entering wastewater,
effluent, process conditions, climatic data, and, lastly, the
effluent. The concentrations of parameters in the effluent
were then predicted using the determined factors and the
neural network model. A genetic algorithm was utilized to
optimize the neural network in order to attain higher accuracy
in refinery modelling. Finally,the concentration of each of the
three parameters in the effluent was predicted using the
correlation coefficient R and the statistical criteria of mean
relative root mean squared prediction The mean absolute
percentage of relative error (rMAPE) and the root mean
squared percentage of relative error (rRMSPE) were
calculated with their actual values, and the model was
evaluated [17].
III. RESULTS
The genetic algorithm in this work looked for the best

response for 450 generations, evaluating 150 possible
solutions in the search space each generation. The effluent
Fig 1: Waste water Treatment Plant summarizes the results of five model runs for the BOD
parameter. Due to the semi-modern character of the
BOD5 (BiochemicalOxygenDemandin5days), COD treatment plant, the highest and minimum values of the
(Chemical Oxygen Demand), TSS (Total Suspended Solids), correlation coefficient of the findings derived from the
and pH were the parameters tested for waste water quality model were 0.93, 0.86, and 1.08, respectively. The average
assessment. Meteorological data, such as average daily air of these values was found to be a satisfactory number. It
temperature, sunshine duration, and daily rainfall, were also shows the TSS, BOD, and COD parameters of a waste water
utilized. To begin modelling the neural network, the data was treatment plan's predictive outcomes. Parameter
partitioned randomly for testing. Figure 2 depicts the input RMSEBOD0.892.3COD0.82TSS 0.83 2.07
and output data, as well as the model architecture. RMSEBOD0.892.3COD0.82TSS 0.83 2.07 The number of
selected neurons in the first and second hidden layers
revealed that the network had failed in 16% of situations.
There are two hidden layers, with an average of 16 neurons
in the first layer and 11 neurons in the second layer. The
network had only one hidden layer in the remaining 84
percent, with an average of 14 neurons. The optimum
network structure had two hidden layers, with 15 neurons in
the first layer and 2 neurons in the second layer on average.
At the greatest value of the correlation coefficient, the

ISSN No:-2456-2165
network comprised only one hidden layer with 11 neurons the highest correlation coefficient. For the structure and
(R). The GA-ANN model, which had a maximum correlation weight of the features in this scenario, the average number of
coefficient of 0.93 and rRMSPE and rMAPE error rates of neurons in each layer and the mean weight of the features in
10% and 7%, respectively, was an effective model for 15 times of model execution were used. In the first scenario,
predicting the concentration of the TBOD parameter and putting the model concerning the TBOD parameter into
produced correct findings. . As a result, it can be used to practise yielded superior results, as shown in Figure 3. These
simulate treatment plants. Parameters impacting microbe concentrations in the wastewater are substantially greater than
growth and activity, such as input TCOD in/TBOD in ratio, the usual TBOD value. As a result, it is critical to execute
dissolved oxygen content, aeration lagoon temperature, and appropriate strategies to improve effluent quality as soon as
input TBOD load, were also prioritised in the model feasible, as well as to ensure the efficacy of the solutions and
predictions due to biological treatment. The parameters the actual improvement of effluent quality at various phases,
affecting the performance of the specified treatment were as predicted by the model..
determined based on the results received from the optimised
artificial network model for the TBOD parameter. It shows a
summary of their findings, as well as the weights assigned to
each parameter. It shows a summary of the elements that are
successful in forecasting the effluent's TBOD parameter, as
well as the weights that are unrelated to it. Important factors
to consider The importance of feature weight in predicting
TBOD levels Aerator Num 0.66 TL 0.85 TAIR 0.73 Q 0.8
DO 0.88 Input pollution load TBOD 0.89 Aerator Num 0.66
TL 0.85 TAIR 0.73
Two characteristics (discharge and pollution load) are

among the input parameters in the table above that play a
crucial role in determining waste water quality. Based on the
available data, it was discovered that in most situations, the
flow rate exceeds the specified flow rate, and one of the
possible remedies is to reduce the input flow rate.Dissolved
oxygen (DO) and lagoon temperature were the most
important process parameters in predicting effluent TBOD.
The type of aeration system and its rate, in general, can have
a direct impact on dissolved oxygen. The temperature of the
lagoon and the rate of TSS precipitation are both effective.
The use of a deep aeration system rather than surface aeration IV. CONCLUSIONS
is one of the greatest ways to improve the performance of the
treatment plant because it efficiently improves the As a result, it is critical to monitor BOD and COD in
concentration of dissolved oxygen, raises the temperature of wastewater in order to prevent environmental pollution and to
the aeration lagoon content, and provides the appropriate save the lives of aquatic animals and fish. RLT offers modern
mixing. Due to the open nature of the aeration lagoon system, technology BOD and COD analyzers that can provide precise
it is practically and directly impossible to manage the effect data in real time. According to the findings, the inlet flow rate,
of air temperature (TAIR) on the treatment system. The effect TCODin/TBODinratio, temperature, and load of organic
of air temperature on the temperature of the aeration lagoon matter in the incoming wastewater, as well as the amount of
content can be decreased by utilizing deep aerators instead of dissolved oxygen, temperature, and pH in lagoon content, and
surface aerators. Surface aerators lose the most heat energy several active aerators, were the most important factors
since the air is more in touch with the aeration lagoon's affecting the Mashhad treatment plant's performance. Air
surface. The volume of inflow, according to the test results of temperature and the quantity of sunshine hours were two
this study, was the most relevant criterion in determining the climate elements that influenced performance. The neural
quality of waste water. The inlet flow rate to the treatment network model was optimized using a complete search
plant in the following years can be determined with a good genetic search technique, and the results showed a maximum
approximation based on the statistics and in the formation of correlation coefficient of 0.89 for the TBOD parameter and
the project. In the given plan, the rate of rise of TBOD and corresponding rRMSPE and rMAPE for the qualitative
TSS input concentration for every 10 years is equal to 5%. As parameter of 10% and 7%, respectively. Among process
a result of this rising trend, the annual concentration of factors - dissolved oxygen concentration, lagoon content
TBOD and TSS of waste water entering the treatment plant temperature, and several active aerators - the neural network
between 2021 and 2025 is anticipated to reach 1.005. The model singled out important parameters in predicting the
model was deployed in two modes to forecast the effluent concentration of TBOD parameter in the effluent -discharge
quality assuming no corrective action to improve the rate and a load of organic matter pollution of incoming
treatment plant's performance. The network topology and wastewater, and from climatic conditions - air temperature.
weight of the features were examined in the first case In comparison to other inlet characteristics, the inlet
scenario, which was based on the case where the model had discharge of the examined treatment plant had a higher weight

ISSN No:-2456-2165
in predicting the concentration of the TBOD parameter of the
effluent. As a result, in order to optimize the performance of
the aforementioned treatment plant, suitable steps should be
done to lower and control the incoming flow to various
treatment plant units.
REFERENCES
[1]. Optimization Of Pharmaceutical Wastewater Treatment

By Biosorption Using Genetic Algorithm Ola Belal
Hasan Abdallah1*, Priy Brat Dwivedi2
[2]. Qualitative evaluation of wastewater treatment plant
performance by neural network model optimized by
genetic algorithm Sara Dadar 1 , Atena Pezeshki 1 ,
Bojan Đurin 2,*, Dragana Dogančić
[3]. Zhao W, Beach TH, Rezgui Y. 2018 A systematic
mixed-integer differential evolution approach for water
network operational optimization.Proc. R. Soc. A 474:
20170879
[4]. Optimal control for wastewater treatment process based
on an adaptive multi_objective differential evolution
algorithm by Qian, Jun-Fie,Ying:Hong- Gui
[5]. Cristina Tuser is Associate Editor for WWD.
Tuser can be reached at ctuser@sgcmail.com.
[6]. Clean TeQ Water provides tertiary treatment of
wastewater, polishing nutrients or improving water
quality once the majority of COD and BOD has been
removed.
[7]. Xie, Y.; Chen, Y.; Lian, Q.; Yin, H.; Peng, J.;Sheng, M.;
Wang, Y. Enhancing Real- Time Prediction of Effluent
Water Quality of Wastewater Treatment Plant Based on
Improved Feedforward Neural Network Coupled with
Optimization Algorithm. Water2022, 14, 1053
[8]. OPTIMIZATION OF W ASTEW ATER DESIGN
USING GENETIC ALGORITHMS D. Starr Stanfield
by RLT Solutions | Jun 18, 2021 | Water Quality
Monitoring System

Prediction of Wastewater Treatment Plant Using Genetic Algorithm

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Prediction of Wastewater Treatment Plant Using Genetic Algorithm

Uploaded by

Copyright:

Available Formats

Volume 8, Issue 1, January – 2023 International Journal of Innovative Science and Research Technology

Prediction of Wastewater Treatment Plant Using

IJISRT23JAN444 www.ijisrt.com 1302

II. MATERIALS AND METHODS

Perkandabad waste water treatment facility No. 1

The genetic algorithm in this work looked for the best

IJISRT23JAN444 www.ijisrt.com 1303

Two characteristics (discharge and pollution load) are

IJISRT23JAN444 www.ijisrt.com 1304

[1]. Optimization Of Pharmaceutical Wastewater Treatment

IJISRT23JAN444 www.ijisrt.com 1305

You might also like