Predicting turbidity and Aluminum in drinking water treatment plants using Hybrid Network (GA- ANN) and GEP

Alsaeed, Ruba; Alaji, Bassam; Ebrahim, Mazen

doi:https://doi.org/10.5194/dwes-2021-8

Preprints

https://doi.org/10.5194/dwes-2021-8

Preprints

31 Mar 2021

Review status: a revised version of this preprint is currently under review for the journal DWES.

Predicting turbidity and Aluminum in drinking water treatment plants using Hybrid Network (GA- ANN) and GEP

Ruba Alsaeed¹, Bassam Alaji², and Mazen Ebrahim³ Ruba Alsaeed et al.

Show author details

¹Engineer at the Department of Sanitary and Environmental Engineering, Faculty of Civil Engineering, Damascus University, Damascus, Syria
²Prof. Department of Sanitary and Environmental Engineering, Faculty of Civil Engineering, Damascus University, Damascus, Syria
³Prof. Department of Engineering Management and Construction, Faculty of Civil Engineering, Damascus University, Damascus, Syria

Received: 08 Mar 2021 – Accepted for review: 23 Mar 2021 – Discussion started: 31 Mar 2021

Abstract. Turbidity is the most important parameter needed to check the status of drinking water, as it is an integrated parameter because its high values indicate high values of other parameters related to water quality. Coagulation and flocculation are the most essential processes for the removal of turbidity in drinking water treatment plants. Using alum coagulants increases the aluminum residuals in treated water, which have been linked to Alzheimer's disease pathogenesis.

In this paper, a hybrid algorithm (GA-ANN) used to predict the turbidity values in the drinking water purification plant in Al Qusayr was used.

The models were constructed using raw water data: turbidity of raw water, pH, conductivity, temperature, and coagulant dose, to predict the turbidity values coming out of the plant.

Several models built and fitness detected for each model, the network with the highest fitness was selected, and then a hybrid prediction network was constructed.

The selected network was the most able to predict turbidity of the outlet with high accuracy with a correlation coefficient (0. 9940) and a root mean square error of 0.1078.

And 4 equations for determining the value of the residual aluminum was obtained using Gene expression method, and the best equation produced results with very good accuracy, in this regard it can be referred to RMSE = 0.02 R = 0.9 for the best model.

Ruba Alsaeed et al.

Status: final response (author comments only)

CC1:
'Comment on dwes-2021-8', fadi Asker, 02 Apr 2021

The article in general is good written and it is about a very important issue related to drinking water treatmentI noticed that the refrences are quite new. And tge equation gives the article more stringth, likewise the GUI .
But I have a question why the researcher used Ga and Ann and GEP
not fuzzy for example?
And at the end of the article I think it is good to mention the other 3 equation that reselted from the research, not just the best one
GOOD LUCK

Citation: https://doi.org/10.5194/dwes-2021-8-CC1
- AC1: 'Reply on CC1', Ruba Alsaeed, 02 Apr 2021
  
  We used GA-ANN Because there is some articles on the internet about predicting turbidity and other parameters by Fuzzy logic
  So to do something new , we used ANN Hybrid with GA, GA was used because many researches in other fields said that GA can improve the structure of ANN. And it gave great results comparing to similar researches.
  And for the other 3 equation, I can provide them in the final copy if I am allowed to add something to the article.
  GEP was great tool for predicting in a very amazing way.
  
  Citation: https://doi.org/10.5194/dwes-2021-8-AC1
CC2:
'Comment on dwes-2021-8', Salim Heddam, 03 Apr 2021
In the present work the authors propose and application of data driven approaches for predicting water turbidity and residual aluminum in drinking water treatment plant. The paper is within the scope of the journal, few similar studies can be found in the literature especially regarding the prediction of residual aluminum which is very interesting. However, the paper with the actual form is with low quality and needs to be in depth amended and reorganized before to be ready for publication.

English is below the required level, and the paper is hard to read and to understand what is really done by the authors. Native English speaker can help in improving the English style of the paper.

From the abstract to the conclusion, the paper needs to be completely reorganized, structured and presented in a clear manner. The actual form is unacceptable, and the paper is without a clear ideas.

Poor literature review and missing research gap.

Theoretical description of the models is out of scope.

Some information’s were provided by the authors and not related to the investigation, i.e., the GRNN model (line 145)???

Results should be deeply analyzed and discussed.
Citation: https://doi.org/10.5194/dwes-2021-8-CC2
- AC3: 'Reply on CC2', Ruba Alsaeed, 12 Apr 2021
  
  Thank you very much for your valuable notices, I have adjusted the literature review
  
  to be more clear and to summarize the previous studies in this field more clear.
  it is shown in the attached file.
  
  Citation: https://doi.org/10.5194/dwes-2021-8-AC3
- AC4: 'Reply on CC2', Ruba Alsaeed, 12 Apr 2021
  
  Thank you very much for your valuable notices, I have adjusted the literature review
  
  to be more clear and to summarize the previous studies in this field more clear.
  it is shown in the attached file.
  
  Citation: https://doi.org/10.5194/dwes-2021-8-AC4
CC3:
'Comment on dwes-2021-8', Lina khouri, 04 Apr 2021

In this paper, a Hybrid network (GA - ANN) which is a data-driven method, was proposed for predicting water turbidity, and GEP which is a promising tool in many fields, was used for the prediction of residual aluminum in drinking water treatment plants. There are a few similar studies in scientific journals, particularly for the prediction of residual aluminum, which was done in this study, and it is a very interesting topic and related to the journal reach.
- The English language should be checked by native English speaker .
- I think the abstract could be rearranged to give more information about the article.

Citation: https://doi.org/10.5194/dwes-2021-8-CC3
- AC2: 'Reply on CC3', Ruba Alsaeed, 07 Apr 2021
  
  I was obligated to a certain number of words in the abstract, I will take this notice into consideration
  Thank you
  
  Citation: https://doi.org/10.5194/dwes-2021-8-AC2
CC4:
'Comment on dwes-2021-8', Lilas Nabhan, 12 Apr 2021

The research is good and has innovative ideas

The issue is also important for water purification plants due to the importance of turbidity in assessing the stability of the plant’s work and controlling aluminum values

But it needs some modification and revision, especially with regard to the language, and it is preferred to be reviewed by a native English speaker

And although that there are some new references, It is also preferable to add other reference studies in the foreground

Citation: https://doi.org/10.5194/dwes-2021-8-CC4
- AC5: 'Reply on CC4', Ruba Alsaeed, 14 Apr 2021
  
  Thank you very much for your valuable notices, I have adjusted the literature review so it becomes more clear and summarizes the previous studies in this field.
  it is shown in the attached file
  
  Citation: https://doi.org/10.5194/dwes-2021-8-AC5
CC5:
'Comment on dwes-2021-8', amin hasanalipou shahrabadi, 13 Apr 2021

This paper is good in overall. The subject of this article is very important todays because of Lack of drinking water resources as well as Existence of pathogenic aluminum particle. In general, the following appear: 1. Almostly new and very interesting subject and new method for solving the problem (there are some article with Fuzzy). This is in scope of the journal. 2. The references are quite new.3. The English language should be checked. I think it will be better.4. The quality of some figs must be better. Ex: fig.85. The abstract must be written in 1 paragraph.

Citation: https://doi.org/10.5194/dwes-2021-8-CC5
- AC6: 'Reply on CC5', Ruba Alsaeed, 14 Apr 2021
  
  Doctor Amin Hasanalipou Shahrabadi
  Thank you for mentioning that methods that were used (GEP and hybrid GA -ANN) were new in solving this kind of problems
  
  as you said there are some articles that worked with Fuzzy systems and GMDH.
  You are right figure 8 needs to be more clear, I have fixed this
  and also adjusted the abstract in one paragraph
  I have attached a file showing the changes
  thank you
  
  Citation: https://doi.org/10.5194/dwes-2021-8-AC6
CC6: 'Comment on dwes-2021-8', Lina khouri, 14 Apr 2021

Good luck

Citation: https://doi.org/10.5194/dwes-2021-8-CC6
CC7: 'Comment on dwes-2021-8', Prakash Jagadeesan, 30 Apr 2021

The authors have developed a soft sensor to predict the turbidity level in the drinking water. The three years data from the water purificatoon plant has been used for validation. The efforst taken by the research scholar and his supevisors deserves special appreciation
Wishing the scholarr a very good luck

Citation: https://doi.org/10.5194/dwes-2021-8-CC7
CC8:
'Comment on dwes-2021-8', Adnan Sallom, 30 Apr 2021

This paper was well prepared.

The stringth points are:

- There are few articles that studied predicting turbidity and much less for the aluminum prediction.

- Also, using GEP in the field of water treatment plants is very promising.

- Predicting models are very important, especially when they are combined with Early Warning Systems.

- Comparing the results of MLR with ANN was very reliable

cause this parameter have nonlinear relations and it is hard to be described using linear ways

However, this paper still needs some further improvements.

1.There is some grammatical and structural errors.

2.The authors did not discuss the limitations of the data set they used.

3.In figure 3 what is the program used to make this figure?

4.figure 7 needs to be more clear.

The article is suitable for publication after some language editing.

Citation: https://doi.org/10.5194/dwes-2021-8-CC8
- CC10: 'Reply on CC8', Ruba Alsaeed, 09 May 2021
  
  professor Adanan
  the program used to make figure 3 is Orange 7. It helped to match 4 different things
  a number of neurons, MSE, the function of the hidden layer, the function of the output layer .
  and as you mentioned, figure 7 needs to be more clear
  I attached a photo of the edited figure
  Thanks for your comments
  
  Citation: https://doi.org/10.5194/dwes-2021-8-CC10
- AC7: 'Reply on CC8', Ruba Alsaeed, 09 May 2021
  
  professor Adanan
  the program used to make figure 3 is Orange 7. It helped to match 4 different things
  a number of neurons, MSE, the function of the hidden layer, the function of the output layer .
  and as you mentioned, figure 7 needs to be more clear
  I attached a photo of the edited figure
  Thanks for your comments
  
  Citation: https://doi.org/10.5194/dwes-2021-8-AC7
CC9: 'Comment on dwes-2021-8', Rana Muhammad Adnan Ikram, 02 May 2021

This subject addressed is within the scope of the journal. However, the manuscript in the present version contains several problems. Appropriate revisions should be undertaken in order to justify recommendation for publication. It is mentioned that GA-ANN and GEP models are used. What are the advantages of adopting these particular methods over others in this case? How will this affect the results? More details should be furnished.Why not tried MARS/OP-ELM/DENFIS/GMDH for comparison and validation? For readers to quickly catch your contribution, it would be better to highlight major difficulties and challenges, and your original achievements to overcome them, in a clearer way in abstract and introduction.

Citation: https://doi.org/10.5194/dwes-2021-8-CC9
RC1:
'Comment on dwes-2021-8', Sina Moharramzadeh, 24 Jun 2021

This manuscript proposes a new method for predicting the turbidity level in the drinking water treatment plant. A three-year dataset from the Al Qusayr water treatment plant was used for developing and validating the model. Although the article is in the scope of the journal and has some merits and innovations, the writing of the paper needs significant improvement. Also, there should be more information about the dataset used in the article and its limitation.

Citation: https://doi.org/10.5194/dwes-2021-8-RC1
- AC8:
  'Reply on RC1', Ruba Alsaeed, 02 Jul 2021
  
  Dear referee and editorial board
  the comment was uploaded as a supplement
  Thanks
  
  Citation: https://doi.org/10.5194/dwes-2021-8-AC8
  - RC2: 'Reply on AC8', Sina Moharramzadeh, 25 Jul 2021
    
    Yes, you should include that graph from your response to your article to improve it. By adding that data, your article should be suitable for publication.
    
    Citation: https://doi.org/10.5194/dwes-2021-8-RC2
    
    AC10: 'Reply on RC2', Ruba Alsaeed, 19 Aug 2021
    
    Dear referee, I am thankful for your notes
    I will add the figure you mentioned in the final version of the article
    Thanks a lot
    
    Citation: https://doi.org/10.5194/dwes-2021-8-AC10
RC3:
'Comment on dwes-2021-8', Anonymous Referee #2, 09 Aug 2021
The paper is about the modelling of the removal of turbidity and the concentration of residual aluminium during coagulation at a surface water treatment plant. This is an important topic, since accurate dosing of chemicals can lead to better performance of the plant and cost reduction. However, the quality of the paper is low. The use of references in the introduction and for the discussion of the results is poor, and the manuscript has severe language issues. In addition, the manuscript has redundant and superfluous information, including figures and tables (e.g. Fig 1, 2, 4, 8, 10; Table 1, 6). Moreover, a discussion lacks on the quality and usability of the data. Data are used from an existing plant that is operated. The obtained data are thus not independent and the results are biased. How can the model then be used in future to optimize coagulant dose?

General comments:

Let the language be checked by a native speaking person

Explain at the end of the introduction, what the “knowledge” gap is, what the “objective” of the study is and how it “contributes” to science/engineering.

Avoid too much theoretical background and explanation of the used models, but concentrate on the reason why they are used and how they can be used.

Avoid giving results of the modelling efforts in the Materials and Methods section

Conclusions section should only give short introduction with objectives and main findings (without references)

Specific comments:

Line 3-4, avoid too general introductions

Line 8, explain that turbidity also has to do with “organoleptics”

Line 10, delete (not a clear sentence)

Line 16, “bottom of the waterbed”?

Line 15-20, give values from small to large

Line 25-30, give references

Line 31, “standardize them”?

Line 32-35, biological processes are not relevant here

Line 38-39, delete (repetition)

Line 46-52, rephrase

Line 57, “raw water KMnO4 and PAC/KMnO4”?

Line 61, “GMDH”?

Line 62-63, not clear what is meant. Does it deal with a reference of with the present study?

Line 5-72, give references

Line 82, “water drained for the station”?

Line 84, “All chemical additives are added before the main dispenser”?

Line 85, “precipitations”? “double-exposed”?

Line 90-140, see in general comments

Line 143, “Mythology”?

Line 145-146, how can you train then for these situations?

Line 156, unclear why conductivity is a relevant parameter for coagulation (explain better)

Line 157, pH determines the solubility of what?

Line 164-166, how significant are the differences? It is also good to have a model with the lowest number of input parameters…

Line 246: what is the difference between validation and testing? ; Fig 5 = Fig 6

Line 258, is Tur-out and input parameter?

Line 265, explain sensitivity method in Materials and methods section. In addition, in the OAT the parameters is not removed but changed…

Line 271-275, not relevant for the paper

Line 278-285, rephrase.. (part should be to materials and methods). Here only results and discussion should be given

Line 301-302, rephrase

Line 321, rephrase… (very good in relation to what)

Line 322-323, give studies and discuss in this light

Line 329-351, see general comments
Citation: https://doi.org/10.5194/dwes-2021-8-RC3
- AC9: 'Reply on RC3', Ruba Alsaeed, 19 Aug 2021
  
  Dear Referee, I am very grateful for your comprehensive feedback and thankful for the time and effort you gave to the article
  I have attached a file with some points and questions
  I repeat my Thanks
  
  Citation: https://doi.org/10.5194/dwes-2021-8-AC9

Ruba Alsaeed et al.

Viewed

Total article views: 1,234 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
834	344	56	1,234	9	8

HTML: 834
PDF: 344
XML: 56
Total: 1,234
BibTeX: 9
EndNote: 8

Views and downloads (calculated since 31 Mar 2021)

Month	HTML	PDF	XML	Total
Mar 2021	15	2	1	18
Apr 2021	299	96	29	424
May 2021	189	34	10	233
Jun 2021	94	20	2	116
Jul 2021	98	32	4	134
Aug 2021	59	28	5	92
Sep 2021	43	43	4	90
Oct 2021	37	89	1	127

Cumulative views and downloads (calculated since 31 Mar 2021)

Month	HTML	PDF	XML	Total
Mar 2021	15	2	1	18
Apr 2021	299	96	29	424
May 2021	189	34	10	233
Jun 2021	94	20	2	116
Jul 2021	98	32	4	134
Aug 2021	59	28	5	92
Sep 2021	43	43	4	90
Oct 2021	37	89	1	127

Viewed (geographical distribution)

Total article views: 1,095 (including HTML, PDF, and XML) Thereof 1,095 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 27 Oct 2021

Short summary

A hybrid algorithm (GA-ANN) was used to predict the turbidity values in the drinking water plant in Al Qusayr. Several models built and fitness detected for each, the network with the highest fitness selected, then the network was constructed. The network was able to predict turbidity with high accuracy, R (0. 9940) and a RMSE of 0.1078. And 4 equations for determining the value of the residual aluminum was obtained using Gene expression, the best equation results were, RMSE = 0.02, R = 0.9.


Total:	0
HTML:	0
PDF:	0
XML:	0