Integrated Modeling of Hybrid Nanofiltration/Reverse Osmosis Desalination Plant Using Deep Learning-Based Crow Search Optimization Algorithm

Round 1

Reviewer 1 Report

The title of your research paper, "Integrated modelling of hybrid nanofiltration/reverse osmosis desalination plant using deep learning-based crow-search optimization algorithm," suggests a complex and technically advanced study in desalination and optimization. Here are some points to consider for your paper: here are some comments for the authors.

The authors have plotted several graphs, but knowing how the graph data was calculated is challenging. How different error was calculated? The clear objective and conclusion are not mentioned.

1.       Explain what you mean by a "hybrid nanofiltration/reverse osmosis desalination plant"?

2.       What is the research gap?

3.       What is the objective of the current study?

4.       Highlight the advantages of using deep learning in the desalination process.

5.       What components or aspects of the desalination plant are being integrated, and why is this integration important?

6.       Explain how deep learning and the crow-search optimization algorithm were applied. This will help readers understand the technical aspects of your work.

7.       Authors effectively present the results using performance statistical criteria like RMSE and MAE. Are mentioning that these measures of prediction accuracy would be helpful for readers unfamiliar with these terms?.

8.       Kindly explain in detail about Figure 2. What is presented in the y-axis and x-axis?

9.       Draw the flowcharts of the genetic algorithm (GA)? Also, provide code.

10.   What is the plot's actual (Input) data presented in Figure 4?

11.   How the MAE-value and RMSE-value were calculated?

12.   How the value of the plot are calculated?

13.   What is the clear output of the current study?

Comments for author File: Comments.pdf

Need to improved. 

Author Response

Manuscript ID: water-2616766 “Integrated modelling of hybrid nanofiltration/reverse osmosis desalination plant using deep learning-based crow-search optimization algorithm”

The authors would like to sincerely thanks the editor and reviewers for the time spent on reviewing the and suggesting the positive comments. All suggestions have been incorporated in the revised manuscript and highlighted by the red color. The manuscript is formatted as per the Water. The point-by-point response to all the comments is given below:

Reviewer #1: Comments and Suggestions for Authors

Comment:  The title of your research paper, "Integrated modelling of hybrid nanofiltration/reverse osmosis desalination plant using deep learning-based crow-search optimization algorithm," suggests a complex and technically advanced study in desalination and optimization. Here are some points to consider for your paper: here are some comments for the authors. The authors have plotted several graphs, but knowing how the graph data was calculated is challenging. How different error was calculated? The clear objective and conclusion are not mentioned.

Response: Thanks for owing the positive comments and minor revision; we appreciate the respected reviewer’s thoughtful comments.  All the comments/suggestions have been incorporated in the revised version of the MS.  We added a strong conclusion in the revised version.

Comment 1: Explain what you mean by a "hybrid nanofiltration/reverse osmosis desalination plant"?

Response 1: Response: Thanks for the positive feedback. A "hybrid nanofiltration/reverse osmosis (NF/RO) desalination plant" as detailed in the provided research, refers to an advanced desalination system that integrates both nanofiltration and reverse osmosis technologies to optimize water desalination. In the context of the study, this hybrid approach is modeled using deep learning and optimized metaheuristic algorithms to enhance the efficacy of the desalination process, particularly in seawater treatment. The NF process effectively reduces the presence of organics, pollutants, and ionic strength, and softens the water, serving as an efficient pretreatment stage, thus enhancing the efficiency of the subsequent RO process. The combined NF/RO system aims to overcome the limitations of using NF or RO in isolation, offering improved desalination results, energy optimization, and cost-effectiveness in water treatment. The research employs advanced artificial intelligence models to further refine the hybrid NF/RO desalination process, aiming to achieve higher accuracy and reliability in permeate conductivity predictions and overall desalination performance.

Comment 2:   What is the research gap?

Response 2: Thank you for the positive comments. We already added the research gap in the new manuscript file. However, the research gap provided in the paper appears to be centered around the challenges and limitations associated with existing desalination methods, particularly reverse osmosis (RO) and nanofiltration (NF). These challenges include high energy and cost requirements, the need for periodic membrane maintenance, the issue of membrane fouling, and inefficiencies in handling high salinity, among others. The existing methodologies for desalination are often limited in terms of accuracy and efficiency, as traditional linear or polynomial correlations may oversimplify the complex dynamics involved in the desalination process. The research aims to address these gaps by employing an integrated approach that combines NF and RO with advanced artificial intelligence (AI) models. Specifically, the study utilizes Long Short-Term Memory (LSTM) networks and an optimized metaheuristic Crow Search Algorithm (CSA) to improve the performance and reliability of the hybrid NF/RO desalination process. This innovative approach seeks to overcome the identified limitations and enhance the effectiveness and sustainability of desalination technologies, contributing to global water sustainability goals.

Comment 3: What is the objective of the current study?

Response 3: Response 2: Comments well appreciated. The objective of the current study has been elaborated upon in the new manuscript file.

Action:

The goals and objectives of the research outlined in the provided research are primarily centred around tackling global water scarcity by enhancing desalination processes, specifically focusing on the hybrid nanofiltration/reverse osmosis (NF/RO) process. The research aims to develop a model for evaluating the performance of NF/RO based on permeate conductivity, utilizing deep LSTM integrated with an optimized metaheuristic CSA and LSTM-CSA. By adopting uncertainty Monte-Carlo simulation, the study evaluates the uncertainty attributed to prediction, aiming to prove the reliability of both LSTM and LSTM-CSA in terms of various performance statistical criteria. The research further examines addressing the key challenges faced in the RO desalination technique, aiming to enhance its efficiency by integrating it with NF, and subsequently optimizing various parameters involved in desalination processes. The study also seeks to leverage the advanced capabilities of the LSTM network and the CSA to refine the model's performance, particularly for the uncertainty analysis and prediction of the hybrid NF/RO desalination process. The paper is committed to maximizing the accuracy of the hybrid algorithms and providing a comprehensive evaluation, contributing to the optimization of energy usage, identification of energy-saving opportunities, and the development of more sustainable operating strategies in desalination processes.

Comment 4: Highlight the advantages of using deep learning in the desalination process.

Response 4: Thanks for the positive comments. We already added the advantages of using deep learning in the desalination process in the revised manuscript file.

Action

It is worth mentioning that the use of deep learning in the desalination process, as illustrated in the given abstract and introduction, holds significant promise in revolutionizing the efficacy and sustainability of water desalination. The application of LSTM networks integrated with an optimized metaheuristic CSA in the hybrid NF/RO desalination process enhances the overall performance, offering notable advantages. One of the primary benefits is the increased accuracy in predicting the desalination process outcomes, with substantial reductions in root mean square error (RMSE) and mean absolute error (MAE), demonstrating the reliability of AI models. This integration also optimizes energy usage, enabling the identification and exploitation of energy-saving opportunities for more sustainable operation. Additionally, AI contributes to the development of advanced brine treatment techniques, minimizing waste and maximizing resource utilization by facilitating the extraction of valuable resources from the brine. The models accurately evaluate and mitigate the uncertainty associated with predictions, ensuring more consistent and reliable desalination results. The comprehensive optimization of various desalination factors, including permeate rate and conductivity, underscores the significant enhancement AI brings to water treatment and desalination, addressing and overcoming many challenges inherent in traditional methods.

Comment 5: What components or aspects of the desalination plant are being integrated, and why is this integration important?

Response 5: Authors appreciated the positive comments. Based on the paper’s methodology, the research integrates artificial intelligence (AI) models with the hybrid nanofiltration/reverse osmosis (NF/RO) desalination process to enhance the performance of water desalination. Specifically, deep learning Long Short-Term Memory (LSTM) networks are combined with an optimized metaheuristic Crow Search Algorithm (CSA), resulting in an advanced LSTM-CSA model. This integration is essential as it aids in developing a more accurate, reliable, and efficient desalination process. Before model development, an uncertainty Monte-Carlo simulation is adopted to evaluate the uncertainty attributed to the prediction, ensuring the robustness and reliability of the AI models. The outcomes showcase that this integration successfully optimizes energy usage, identifies energy-saving opportunities, and proposes more sustainable operating strategies. Moreover, AI assists in developing advanced brine treatment techniques, thus minimizing waste and maximizing resource utilization by facilitating the extraction of valuable resources from the brine. The integration plays a crucial role in overcoming the challenges faced in the NF/RO desalination process, demonstrating improvements in accuracy and reliability in predicting the permeate conductivity, essential for the evaluation of the desalination process performance.

Comment 6: Explain how deep learning and the crow-search optimization algorithm were applied. This will help readers understand the technical aspects of your work.

Response 6: We appreciate the positive feedback. We have added details explanations.

Actions

In the paper, deep learning and CSA are applied to improve the performance and efficiency of hybrid nanofiltration/reverse osmosis (NF/RO) desalination processes. Specifically, a LSTM network, a type of recurrent neural network, is used to model the desalination process. The LSTM network is adept at learning from sequences of data, which makes it suitable for modelling complex processes such as desalination where numerous variables interact over time. The LSTM is integrated with CSA, a metaheuristic optimization algorithm. CSA is inspired by the behavior of crows and is used to optimize the parameters of the LSTM network, ensuring that the model makes the most accurate predictions possible. Before developing this LSTM-CSA model, an uncertainty analysis is conducted using the Monte Carlo simulation to evaluate potential uncertainties related to the predictions made by the model. This innovative approach utilizing both LSTM and CSA is designed to enhance the prediction of the performance of the NF/RO desalination process based on permeate conductivity, a key parameter in assessing the effectiveness of the desalination process. The research aims to optimize energy usage and suggest more sustainable operating strategies for desalination, ultimately leading to enhanced water recovery rates, reduced energy consumption, and improved overall efficiency of the desalination process. In a more technical understanding, the LSTM network is trained using historical data from the desalination process, learning to identify patterns and relationships that are not easily discernible. It then uses this learning to make predictions about the performance of the desalination process under various conditions. The CSA is applied to fine-tune the parameters of the LSTM network, ensuring that it operates as effectively as possible. The integration of deep learning with CSA thus presents a robust and innovative approach to optimizing the performance of NF/RO desalination processes, contributing to advancements in addressing global water scarcity challenges.

Comment 7: Authors effectively present the results using performance statistical criteria like RMSE and MAE. Are mentioning that these measures of prediction accuracy would be helpful for readers unfamiliar with these terms?

Response 7: Thank you for the comments.

The authors present the results of their research effectively by employing performance statistical criteria, specifically the Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) and other visualization. These metrics judge the model's prediction accuracy, highlighting the reliability of the LSTM and LSTM-CSA models in predicting permeate conductivity in the NF/RO desalination process. For readers unfamiliar with these terms, providing a brief explanation or context about RMSE and MAE would indeed enhance the comprehension and accessibility of the research findings. Although the use of these terms underpins the scientific rigor and precision of the research, ensuring that the study is understandable to a broader audience by elucidating these terms would augment the impact and reach of the authors’ work, making it more inclusive and comprehensible to readers from varied backgrounds and levels of expertise.

Comment 8: Kindly explain in detail about Figure 2. What is presented in the y-axis and x-axis?

Response 8: Authors thank the reviewer for the suggestions. We have added more details.

Comment 9: Draw the flowcharts of the genetic algorithm (GA)? Also, provide code.

Response 9: Thank you very much the GA is not part of the main work hence we remove it.

Comment 10: What is the plot's actual (Input) data presented in Figure 4?

Response 10: Thank you we have updated the input and the target variable in updated manuscript.

Action

For modelling purposes, the six most important parameters were accurately sourced from 73 distinct data points. The chosen independent parameters covered a diverse range, temperature (⁰C), time duration (h), pressure levels (kg/cm²), feed flow rate (m³/h), and feed conductivity (μS/cm). On the other hand, our target or dependent parameter was identified as the permeate conductivity (μS/cm). Through the feature selection insights collected from the PSO algorithm, two distinct datasets were crafted. The result of our study is expected to show a comprehensive evaluation of these models, with both training and testing datasets subjected to performance metrics like the root mean square error (RMSE) and the mean absolute error (MAE). The findings were explaining models powered by the metaheuristic algorithms notably outperformed the singular LSTM model in terms of accuracy.

Comment 11: How the MAE-value and RMSE-value were calculated?

Response 11: Thank you we have calculated using statistical equations.

Action

The integration of deep learning with CSA thus presents a robust and innovative approach to optimizing the performance of NF/RO desalination processes, contributing to advancements in addressing global water scarcity challenges. The performance criteria used to calculate the model’s accuracy are presented in the following equations:

                                                                               (1)

                                                                                             (2)

MSE =                                                                                                  (2)

where      and  indicate the predicted and computed values   with N as means for the data points.

Comment 12: How the value of the plot are calculated?

Response 12: Thank you very much we used the above formula which also presented in Table.

Comment 13: What is the clear output of the current study?

Response 13: Thank you the output has been mention in several place which is “permeate conductivity (μS/cm)”

Author Response File: Author Response.docx

Reviewer 2 Report

The article is devoted to mathematical modeling.

Notes:

1. What are the goals and objectives of your research? In the introduction, it is necessary to clearly formulate the purpose of the study.

2. It is necessary to add a section devoted to the analysis of existing methods of mathematical and computer modeling. In this section, it is necessary to refer to the critically important work aimed at the methods of mathematical modeling and artificial intelligence.

-Ilyushin, Y.V.; Kapostey, E.I. Developing a Comprehensive Mathematical Model for Aluminum Production in a Soderberg Electrolyser. Energies 2023, 16, 6313. https://doi.org/10.3390/en16176313

-Pershin, I.M.; Papush, E.G.; Kukharova, T.V.; Utkin, V.A. Modeling of Distributed Control System for Network of Mineral Water Wells. Water 2023, 15, 2289. https://doi.org/10.3390/w15122289

-Martirosyan, A.V.; Martirosyan, K.V.; Chernyshev, A.B. Investigation of Popov’s Lines’ Limiting Position to Ensure the Process Control Systems’ Absolute Stability. In Proceedings of the 2023 XXVI International Conference on Soft Computing and Measures (SCM), Saint Petersburg, Russia, 24–26 May 2023; pp. 69–72. https://doi.org/10.1109/SCM58628.2023.10159089.

To search for additional works, I recommend using the search engine mdpi.com or similar.

3. lines 144-155 must be supplemented with a mathematical description of the method used. Describe in more detail its difference from existing methods.

4. Figure 2 Figures A and B. must be broken down into separate components. The principle of grouping makes it difficult to understand these drawings. In the figure with temperature, there is a sharp drop in temperature. What is it connected with? Experiment error?

5. The work must be supplemented with an analysis of the results. It is necessary to consider whether errors and shortcomings of the experiment are possible.

6. Add a discussion section before the Conclusions section. It is necessary to give a critical assessment of their work.

Conclusion. I characterize the article as positive and in need of serious improvement.

Author Response

Manuscript ID: water-2616766 “Integrated modelling of hybrid nanofiltration/reverse osmosis desalination plant using deep learning-based crow-search optimization algorithm”

The authors would like to sincerely thanks the editor and reviewers for the time spent on reviewing the and suggesting the positive comments. All suggestions have been incorporated in the revised manuscript and highlighted by the red color. The manuscript is formatted as per the Water. The point-by-point response to all the comments is given below:

Reviewer #2: Comments and Suggestions for Authors

Comment: The article is devoted to mathematical modeling.

Response: We appreciate respected reviewer’s thoughtful comments. All the comments/suggestions have been incorporated in the revised version of the manuscript.

Comment 1: What are the goals and objectives of your research? In the introduction, it is necessary to clearly formulate the purpose of the study.

Response 1: Thank you for the positive comments.

Action:
The goals and objectives of the research outlined in the provided research are primarily centered around tackling global water scarcity by enhancing desalination processes, specifically focusing on the hybrid nanofiltration/reverse osmosis (NF/RO) process. The research aims to develop a model for evaluating the performance of NF/RO based on permeate conductivity, utilizing deep LSTM integrated with an optimized metaheuristic CSA and LSTM-CSA. By adopting uncertainty Monte-Carlo simulation, the study evaluates the uncertainty attributed to prediction, aiming to prove the reliability of both LSTM and LSTM-CSA in terms of various performance statistical criteria. The research further examines addressing the key challenges faced in the RO desalination technique, aiming to enhance its efficiency by integrating it with NF, and subsequently optimizing various parameters involved in desalination processes. The study also seeks to leverage the advanced capabilities of the LSTM network and the CSA to refine the model's performance, particularly for the uncertainty analysis and prediction of the hybrid NF/RO desalination process. The paper is committed to maximizing the accuracy of the hybrid algorithms and providing a comprehensive evaluation, contributing to the optimization of energy usage, identification of energy-saving opportunities, and the development of more sustainable operating strategies in desalination processes.

Comment 2: It is necessary to add a section devoted to the analysis of existing methods of mathematical and computer modeling. In this section, it is necessary to refer to the critically important work aimed at the methods of mathematical modeling and artificial intelligence.

-Ilyushin, Y.V.; Kapostey, E.I. Developing a Comprehensive Mathematical Model for Aluminum Production in a Soderberg Electrolyser. Energies 2023, 16, 6313. https://doi.org/10.3390/en16176313

-Pershin, I.M.; Papush, E.G.; Kukharova, T.V.; Utkin, V.A. Modeling of Distributed Control System for Network of Mineral Water Wells. Water 2023, 15, 2289. https://doi.org/10.3390/w15122289

-Martirosyan, A.V.; Martirosyan, K.V.; Chernyshev, A.B. Investigation of Popov’s Lines’ Limiting Position to Ensure the Process Control Systems’ Absolute Stability. In Proceedings of the 2023 XXVI International Conference on Soft Computing and Measures (SCM), Saint Petersburg, Russia, 24–26 May 2023; pp. 69–72. https://doi.org/10.1109/SCM58628.2023.10159089.

To search for additional works, I recommend using the search engine mdpi.com or similar.

Response 2: The authors are grateful for this comment.  We have added the above suggested references where appropriate.

Action

Generally, membrane processes in desalination have predominantly been based on linear or polynomial correlations, even though these methods may generalize the complex dynamics of the desalination procedure [15]–[17]. In contrast to these mathematical and traditional paradigms, which often settle for lower precision, the modern technological landscape, marked by the beginning of Industry 4.0 and the growing field of the Industrial Internet of Things (IoT), has seen artificial intelligence (AI) algorithms making substantial inroads into various industrial spheres [18].

<<References>>

[15]    Y. V. Ilyushin and E. I. Kapostey, “Developing a Comprehensive Mathematical Model for Aluminium Production in a Soderberg Electrolyser,” Energies, vol. 16, no. 17, 2023, doi: 10.3390/en16176313.

[16]    I. M. Pershin, E. G. Papush, T. V. Kukharova, and V. A. Utkin, “Modeling of Distributed Control System for Network of Mineral Water Wells,” Water (Switzerland), vol. 15, no. 12, 2023, doi: 10.3390/w15122289.

[17]    N. AlSawaftah, W. Abuwatfa, N. Darwish, and G. Husseini, “A comprehensive review on membrane fouling: Mathematical modelling, prediction, diagnosis, and mitigation,” Water, vol. 13, no. 9, p. 1327, 2021.

[18]    M. S. Bhati, “Industrial Internet of Things ( IIoT ): A Literature Review,” no. 03, pp. 304–307, 2018.

Comment 3: lines 144-155 must be supplemented with a mathematical description of the method used. Describe in more detail its difference from existing methods.

Response 3: The authors thank the reviewer for this comment.  We have added more explanation based on the above comments.

Actions
In the effort to handle challenges faced in the field of desalination modelling, we sought to leverage the advanced capabilities of the Long Short-Term Memory (LSTM) network. The primary focus of this study was to refine the model's performance, particularly for the uncertainty analysis and prediction of the hybrid NF/RO desalination process. To realize this aim, we employed a state-of-the-art metaheuristic optimization technique, Crow Search Algorithm (CSA). A noteworthy originality of our methodology was the inclusion of the Particle Swarm Optimization (PSO) algorithm. This algorithm was instrumental in recognizing the optimal combination of features, serving as a foundation to our in-depth modelling, statistical evaluations, and graphical analysis. Such a strategic approach was intended to maximize the accuracy of the hybrid algorithms (LSTM-CSA) as well as the LSTM in its single form. For modelling purposes, six most important parameters were accurately sourced from 73 distinct data points. The chosen independent parameters covered a diverse range, encompassing temperature (⁰C), time duration (h), pressure levels (kg/cm²), feed flow rate (m³/h), and feed conductivity (μS/cm). On the other hand, our target or dependent parameter was identified as the permeate conductivity (μS/cm). Through the feature selection insights collected from the PSO algorithm, two distinct datasets were crafted. The result of our study is expected to show a comprehensive evaluation of these models, with both training and testing datasets subjected to performance metrics like the root mean square error (RMSE) and the mean absolute error (MAE). Our findings were illuminating models powered by the metaheuristic algorithms notably outperformed the singular LSTM model in terms of accuracy. Among the hybrids, the LSTM-GA variant emerged as a standout, displaying unparalleled precision and the least uncertainty and error margins.

The goals and objectives of the research outlined in the provided research are primarily centred around tackling global water scarcity by enhancing desalination processes, specifically focusing on the hybrid nanofiltration/reverse osmosis (NF/RO) process. The research aims to develop a model for evaluating the performance of NF/RO based on permeate conductivity, utilizing deep LSTM integrated with an optimized metaheuristic CSA and LSTM-CSA. By adopting uncertainty Monte-Carlo simulation, the study evaluates the uncertainty attributed to prediction, aiming to prove the reliability of both LSTM and LSTM-CSA in terms of various performance statistical criteria. The research further examines addressing the key challenges faced in the RO desalination technique, aiming to enhance its efficiency by integrating it with NF, and subsequently optimizing various parameters involved in desalination processes. The study also seeks to leverage the advanced capabilities of the LSTM network and the CSA to refine the model's performance, particularly for the uncertainty analysis and prediction of the hybrid NF/RO desalination process. The paper is committed to maximizing the accuracy of the hybrid algorithms and providing a comprehensive evaluation, contributing to the optimization of energy usage, identification of energy-saving opportunities, and the development of more sustainable operating strategies in desalination processes.

Comment 4: Figure 2 Figures A and B. must be broken down into separate components. The principle of grouping makes it difficult to understand these drawings. In the figure with temperature, there is a sharp drop in temperature. What is it connected with? Experiment error?

Response 4: The authors are grateful for this positive feedback.  We have added more explanation based on the above comments. For the figures there is an idea of making them two actually.

Actions
In the scenario described where a figure indicating temperature in an NF/RO system shows a sharp drop, several factors could be responsible, and it's crucial to examine the experimental setup and conditions to determine the exact cause. It might be related to experimental error, where an issue with the equipment, such as a malfunctioning temperature sensor, could give inaccurate readings. Another possibility could be a sudden change in the system parameters or operating conditions, such as a sudden influx of feed water at a different temperature, or a change in the ambient conditions. An unexpected drop in temperature could also relate to the system's performance, indicating potential issues or inefficiencies within the NF/RO process itself, such as unexpected heat loss. It is crucial to thoroughly investigate the system, review the experimental procedure, and check the equipment to accurately diagnose the reason behind the sudden temperature drop.

Comment 5: The work must be supplemented with an analysis of the results. It is necessary to consider whether errors and shortcomings of the experiment are possible.

Response 5: The authors are grateful for this comment. In analyzing the results of the research detailed in the abstract and introduction, it is paramount to scrutinize both the LSTM and LSTM-CSA models’ performance based on permeate conductivity in the NF/RO desalination process. Given the technological complexity and the multifaceted components involved in the desalination process, various errors and shortcomings could potentially emerge during the experiment. One significant area to consider is the uncertainty analysis, which has been addressed in the study using Monte Carlo simulation before the model development. It is vital to examine how effectively this simulation has been executed and if it comprehensively evaluates all possible uncertainties. Even a minor oversight in this phase can lead to erroneous predictions and assessments in later stages. Another potential error could stem from the integration of the Crow Search Algorithm (CSA) with LSTM. The precise tuning of these advanced algorithms is crucial for optimizing their performance. Any misconfiguration or integration issues could adversely impact the results, leading to possible inaccuracies in the prediction of permeate conductivity in the NF/RO process. The study's reliance on various parameters like temperature, time duration, pressure levels, feed flow rate, and feed conductivity for modeling purposes also opens avenues for potential errors. Ensuring the accuracy and consistency of these parameter measurements is essential to avoid propagation of errors in the model predictions. Any inconsistency or error in these measurements can significantly skew the results. In the context of the sharp drop in temperature in the NF/RO system, thorough investigation is warranted to ascertain whether it's an experimental error, a system performance issue, or a valid observation. It is crucial to consider all potential sources of errors, including equipment malfunction, abrupt changes in system parameters, and external factors, to ensure the reliability and validity of the experimental results. While the research employs advanced AI models to enhance the efficiency of the NF/RO desalination process, a meticulous and comprehensive analysis of the results, considering all potential errors and shortcomings, is imperative to validate the findings and ensure their applicability in real-world desalination processes.

ActionsTop of Form

In the analysis regarding NF/RO desalination utilizing LSTM and LSTM-CSA models, it's vital to precisely assess potential errors and experiment shortcomings. Uncertainty analysis, performed using Monte Carlo simulation prior to model development, should be critically evaluated for comprehensive and accurate execution using various error criteria. The integration of LSTM with CSA requires precise tuning; any integration issues could negatively affect the results, leading to inaccuracies in permeate conductivity predictions. The research's reliance on diverse parameters for modelling introduces another possibility for potential error. Accurate and consistent measurements are fundamental to ensure the reliability of model predictions. Concerning the observed temperature drop (Fig. 2a) in the NF/RO system, a thorough investigation is necessary to determine if it's an experimental error, a system issue, or a valid result. Comprehensive analysis, considering all potential errors, is crucial for validating the research findings and their real-world applicability in desalination processes.

Comment 6: Add a discussion section before the Conclusions section. It is necessary to give a critical assessment of their work. Conclusion. I characterize the article as positive and in need of serious improvement.

Response 6: The authors are grateful for the feedback. We have added several discussions in results and discussion part.

Author Response File: Author Response.docx

Reviewer 3 Report

Dear Editor,

Water

In this research, the performance of NF/RO based on permeate conductivity was developed using deep learning long short-term memory (LSTM) integrated with an optimized metaheuristic crow search algorithm (CSA) (LSTM-CSA). Before model development, uncertainty Monte-Carlo simulation was adopted to evaluate the uncertainty attributed to the prediction. The subject addressed is interesting and within the scope of the Water. Also, the novelty of manuscript is acceptable.  Nevertheless, some major revisions have been found:

Major revisions:

1-     What is the mean of “the performance of NF/RO based on permeate conductivity was developed using deep learning long short-term memory (LSTM)”? Which variables are modeled and what are input variables? Please state it clearly. 

2-     The literature review is limited to the NF/RO related studies. However, what is the application of AI methods in other fields? Why do you employ LSTM and CSO for modeling PC? Hence, the literature review should be completed with new and applicable studies such as:

a-     Modeling the optimal dosage of coagulants in water treatment plants using various machine learning models

b-     Prediction of groundwater table and drought analysis; a new hybridization strategy based on bi-directional long short-term model and the Harris hawk optimization algorithm

c-     Predicting the ammonia nitrogen of wastewater treatment plant influent via integrated model based on rolling decomposition method and deep learning algorithm

3-     The time series plot of simulated data together with observed data should be presented in the results section.

4-     In Figure 4, the difference between observed and simulated data is significant.

5-     It is recommended that the significant difference between simulated and observed data is evaluated based on the Wilcoxon test.  

6-     In this study which figure is related to the Tylor plot?

7-     The main quantities results should be presented in the conclusion.

8-     Please add the limitations of your study in the conclusion section. Minor revision:

9- Is correct the world of “Taylor” or “Tylor”?

10- In Figure 4 there are six CDF in each plot, while two of them are referenced in legend. This figure must be revised.

11- Table 1 and Table 2 are referenced however, are not presented in the text of the manuscript.

12- The quality of Figure 1.c and 6 are low.

Considering the mentioned points, this study in the current version needs major revisions.

With kind regards,

Minor editing of English language required

Author Response

Manuscript ID: water-2616766 “Integrated modelling of hybrid nanofiltration/reverse osmosis desalination plant using deep learning-based crow-search optimization algorithm”

The authors would like to sincerely thanks the editor and reviewers for the time spent on reviewing the and suggesting the positive comments. All suggestions have been incorporated in the revised manuscript and highlighted by the red color. The manuscript is formatted as per the Water. The point-by-point response to all the comments is given below:

Reviewer #3: Comments and Suggestions for Authors

Comment:  In this research, the performance of NF/RO based on permeate conductivity was developed using deep learning long short-term memory (LSTM) integrated with an optimized metaheuristic crow search algorithm (CSA) (LSTM-CSA). Before model development, uncertainty Monte-Carlo simulation was adopted to evaluate the uncertainty attributed to the prediction. The subject addressed is interesting and within the scope of the Water. Also, the novelty of manuscript is acceptable.  Nevertheless, some major revisions have been found.

Response: Thank you for the positive comments and minor revision. We appreciate respected reviewer’s thoughtful comments.  All the comments/suggestions have been incorporated in the revised version of the MANUSCRIPT.

Comment 1: What is the mean of “the performance of NF/RO based on permeate conductivity was developed using deep learning long short-term memory (LSTM)”? Which variables are modeled and what are input variables? Please state it clearly?

Response 1: Thank you for the feedback. We have added a clear answer in the revised version
Actions

For modelling purposes, the six most important parameters were accurately sourced from 73 distinct data points. The chosen independent parameters covered a diverse range, temperature (⁰C), time duration (h), pressure levels (kg/cm²), feed flow rate (m³/h), and feed conductivity (μS/cm). On the other hand, our target or dependent parameter was identified as the permeate conductivity (μS/cm). Through the feature selection insights collected from the PSO algorithm, two distinct datasets were crafted. The result of our study is expected to show a comprehensive evaluation of these models, with both training and testing datasets subjected to performance metrics like the root mean square error (RMSE) and the mean absolute error (MAE). The findings were explaining models powered by the metaheuristic algorithms notably outperformed the singular LSTM model in terms of accuracy.

Inputs: temperature (⁰C), time duration (h), pressure levels (kg/cm²), feed flow rate (m³/h), and feed conductivity (μS/cm).

output: permeate conductivity (μS/cm).

Comment 2:   The literature review is limited to the NF/RO related studies. However, what is the application of AI methods in other fields? Why do you employ LSTM and CSO for modeling PC? Hence, the literature review should be completed with new and applicable studies such as:?

a-     Modeling the optimal dosage of coagulants in water treatment plants using various machine learning models

b-     Prediction of groundwater table and drought analysis; a new hybridization strategy based on bi-directional long short-term model and the Harris hawk optimization algorithm

c-     Predicting the ammonia nitrogen of wastewater treatment plant influent via integrated model based on rolling decomposition method and deep learning algorithm

Response 2: Authors appreciates this comment. We have added several literature including the 3 suggested papers were appropriate.

Action

Achite et al. [1] introduced a hybrid model known as M5-GTO, built upon a blend of the M5 and Gorilla Troops Optimizer (GTO) algorithms. This research employed nine diverse parameters including raw water production (RWP), water turbidity, Conductivity, TDS, Salinity, pH, water temperature (WT), SM, and O2 as inputs for CD modeling. The comparative analysis highlighted in the study robustly positions the M5-GTO model as superior in CD modeling accuracy against numerous established models such as multiple linear and nonlinear regression, artificial neural network, multivariate adaptive regression splines, M5 model tree, k-nearest neighbor, least-squares support vector machine, general regression neural network, and random forest (RF). This was exemplified by the recorded values of various error metrics and correlation coefficient criteria for the M5-GTO, all demonstrating significant improvements over the least effective algorithm, LSSVM. The findings further ranked M5 and RF algorithms second and third, respectively. Moreover, the research provided insight into the significant impact of RWP and WT on CD alterations, showcasing the inverse relationship of RWP with CD and the direct relation of WT with CD. In the study conducted by Firzin et al., [2], a novel BLSTM-HHO algorithm is introduced for improved groundwater table (GWT) and drought predictions. This innovative algorithm notably outperformed other benchmark algorithms such as standalone BLSTM, LSTM, ANN, SARIMA, and ARIMA in terms of prediction accuracy and performance criteria. Similalrly, in the research conducted by Yan et al., [3] an innovative approach to predicting influent ammonia nitrogen (NH3-N) concentration in wastewater treatment plants is proposed, leveraging the synergistic capabilities of the rolling decomposition method and deep learning algorithms. The integrated model delineated in the study notably circumvents information leakage during the decomposition process, showcasing enhanced performance compared to standalone GRU models as evidenced by reduced RMSE, MAE, and MAPE values. The study robustly underscores the superiority of the proposed model over other integrated models trained with information leakage, highlighting notable strides in prediction accuracy.

<<References>>

[1]      M. Achite, S. Farzin, N. Elshaboury, M. Valikhan Anaraki, M. Amamra, and A. K. Toubal, “Modeling the optimal dosage of coagulants in water treatment plants using various machine learning models,” Environ. Dev. Sustain., 2022, doi: 10.1007/s10668-022-02835-0.

[2]      S. Farzin, M. V. Anaraki, M. Naeimi, and S. Zandifar, “Prediction of groundwater table and drought analysis; a new hybridization strategy based on bi-directional long short-term model and the Harris hawk optimization algorithm,” J. Water Clim. Chang., vol. 13, no. 5, pp. 2233–2254, May 2022, doi: 10.2166/wcc.2022.066.

[3]      K. Yan, C. Li, R. Zhao, Y. Zhang, H. Duan, and W. Wang, “Predicting the ammonia nitrogen of wastewater treatment plant influent via integrated model based on rolling decomposition method and deep learning algorithm,” Sustain. Cities Soc., vol. 94, Jul. 2023, doi: 10.1016/j.scs.2023.104541.

Comment 3: The time series plot of simulated data together with observed data should be presented in the results section.

Response 3: Thank you. The time series is not for results but for only the raw data showing the pattern of the raw data.

Comment 4: In Figure 4, the difference between observed and simulated data is significant.

Response 4: Yes, it is in combination one only LSTM-M1 but in other approach the results generated remarkable accuracy. 

Comment 5: It is recommended that the significant difference between simulated and observed data is evaluated based on the Wilcoxon test. 

Response 5: Thank you for the suggestion. The authors will consider this in the next proposed work. We do appreciate indeed.

Comment 6: In this study which figure is related to the Tylor plot?

Response 6: Thank you for your kind observation. The Figure referring to Taylor plot has been changed to fan-plot and CDF plot in the revised manuscript.

Comment 7: The main quantities results should be presented in the conclusion.

Response 7: We added the reviewer’s suggestion in the revised file.

Conclusions

In the face of ever-growing ecological challenges and the critical importance of sustainable development goals, it is evident that innovative solutions are a necessity. This research highlights the transformative potential of artificial intelligence in advancing desalination processes, particularly in addressing the global issue of water scarcity. By focusing on the hybrid nanofiltration/reverse osmosis (NF/RO) process, the study introduces an AI model using deep learning long short-term memory (LSTM) combined with the metaheuristic crow search algorithm (CSA), termed LSTM-CSA. This model was rigorously tested based on several performance criteria and validated using both external and internal validation techniques, with a preliminary uncertainty assessment through the Monte-Carlo simulation. The statistical performance of the models was commendable. The LSTM model displayed noteworthy accuracy, but the LSTM-CSA outperformed it, suggesting that the integration of the crow search algorithm with deep learning augments predictive capability. This was further corroborated by innovative 2D-graphical visualization methods, such as the fan plot, which added depth to the accuracy evaluation by accounting for various other assessment indicators. Beyond just the technical accomplishments, the broader implications of this research are profound. AI, as demonstrated, can be critical in making desalination processes more energy-efficient, opening avenues for substantial energy savings. It can further guide the development of advanced techniques to treat brine, allowing for resource extraction and thereby ensuring minimal wastage. In essence, AI model is not only reliable promise for enhancing desalination processes but also propels the sector towards a more sustainable and efficient future. The quantitative summary is presented in the following points.

• LSTM-M1: This model has an MSE of 0.3168, an RMSE of 0.5628, and an MEA of 0.1468. Among the models listed, it performs better than LSTM-M2 but is outperformed by both LSTM-CSA variants in all three metrics, indicating it has good predictive accuracy and reliability.
• LSTM-M2: This model demonstrates the highest error rates across all three metrics with an MSE of 0.4284, RMSE of 0.6545, and MEA of 0.1884, indicating it has the least predictive accuracy and reliability among the listed models.
• LSTM-CSA-M1: This model outperforms the LSTM models in all metrics with an MSE of 0.1985, RMSE of 0.4456, and MEA of 0.1985. Its error rates are lower, showcasing enhanced predictive accuracy and reliability, substantiating the effectiveness of integrating CSA with LSTM.
• LSTM-CSA-M2: This model reveals error metrics very close to LSTM-CSA-M1 with an MSE of 0.1992, RMSE of 0.4463, and an MEA of 0.1292. It has the lowest MEA among all models, indicating it has the smallest average prediction error, making it the most reliable model for predictions among the ones listed.

Comment 8: Please add the limitations of your study in the conclusion section.

Response 8: Thank you for the suggestion. The limitation has been added in the revised manuscript as can be seen below.

While the study provides a comprehensive overview of the research conducted, it does not explicitly highlight the limitations of the study, which is a crucial element for balanced and credible outcomes. The absence of this information could leave readers with unanswered questions regarding the potential challenges or limitations faced during the research. For instance, issues related to the generalizability of the AI models used (LSTM and LSTM-CSA) to other desalination processes or real-world scenarios are not addressed which is part of the limitation of this study. Additionally, the study does not mention any limitations regarding the data used for training the deep learning models or the potential impact of these limitations on the models’ predictive accuracy. Further, the study does not discuss the computational complexity or resource requirements of the proposed models, which could be significant factors for practical implementation. Acknowledging and outlining these limitations would provide a more holistic view of the research conducted, allowing readers to assess the applicability and relevance of the research findings to their own work or contexts.

Comment 9: Is correct the world of “Taylor” or “Tylor”?

Response 9: Thank you for the insightful observation. We have corrected the word “Tylor to Taylor” all through the revised manuscript.

Comment 10: In Figure 4 there are six CDF in each plot, while two of them are referenced in legend. This figure must be revised.

Response 10: Thak for the positive comment. The legend is not the priority of this software and our paper rather the Parten between the observed and predicted value. 

Comment 11: Table 1 and Table 2 are referenced, however, are not presented in the text of the manuscript.

Response 11: Thank you for the insightful observation. This is an oversight from us. We have incorporated the Tables in the revised manuscript.

Actions:

Table 1: Results of PC in desalination plant for LSTM

Table 2: Results of PC in desalination plant for LSTM-CSA

Comment 12: The quality of Figures 1.c and 6 are low.

Response 12: Thank you. We have improved the quality in the revised manuscript.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Authors has responded comments satisfactorily, thereofre manuscript may be accepted  

Seems fine. 

Reviewer 2 Report

no comments

Reviewer 3 Report

Dear Editor,

 Atmosphere

This manuscript is previously been carefully evaluated. The current version of the manuscript is acceptable.

With kind regards,