Dynamic Modeling of Flue Gas Desulfurization Process via Bivariate EMD-Based Temporal Convolutional Network
Round 1
Reviewer 1 Report
The manuscript presents dynamic modeling of flue gas desulfurization process via biva-riate EMD-based temporal convolutional network. The manuscript compares the proposed bivariate empirical mode decomposition (BEMD)-temporal convolutional net-work (TCN) model and several reperesentative models for the operating data from the desulfurization system of a 600 MW coal-fired power station in China. The simulation results show that BEMD-TCN approach yields the best performance which demonstrates its effectiveness for industrial flue gas desulfurization (IFGD) dynamic modelling problem.
Abstract was prepared in accordance with the proposed research. There are few spelling mistakes, like frequent was mistyped as requent. Since line numbers were not provided, it is difficult to notify corrections in the particular locations.
Introduction presents about the effects of SO2 on environment, improtance of modelling and about first-principles and data-driven models, literature review, and organization of manuscript. The technical content is good. May be the long second paragraph could be splitted without disturbing the information flow of manuscript.
Section 2 describes IFGD process, problem statement, research gaps, and objectives. The whole section was well organized and presented in informative manner.
Section 3 provides stepwise approach of the BEMD-TCN modelling. The architecture, pseudocode, comparison of different convolution structures, and representation of BEMD-TCN approach were highlighted through illustrations. The sketches were very clear and understandable to prospective readers with supportive theoretical information.
Section 4 explores error analytical parameters to compare existing and proposed models. The case study was elaborated in details with images. The results with scientific interpretations were written in organized way.
The final conclusion section summarizes major findings of the research, which supports the research purpose.
References from 2023 could be added.
Language needs checking for typographical errors.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Dear Author(s)
The present paper discusses the modeling of industrial flue gas desulphurization (IFGD) via propose a new bivariate empirical mode decomposition (BEMD) approach based on the temporal convolutional network (TCN) to tackle the IFGD modeling task. In general, the research includes good novelty but, there are some points should clarify
In Abstract:
1. “The modeling of industrial flue gas desulfurization (IFGD) has become urgently necessary”, but the reasons for that are not explained.
2. “a novel bivariate empirical mode decomposition (BEMD) based temporal convolutional network (TCN) approach is proposed to address the IFGD modeling task”, but the reasons for proposing this new approach or what the previous problems were are not demonstrated, too.
The introduction needs to add a paragraph at the beginning, explaining the problem of study in detail before reviewing the references that dealt with the study of the subject. The countries that have studied this problem should also be specified, since it is clear that only Chinese references are mentioned.
A list of symbols mentioned in the manuscript should be provided with a full definition of each item and mention of the units (if any).
The conclusions paragraph requires more clarification of the results that have been reached, as the paragraph in its current status does not adequately reflect the results and objectives of the paper.
Dear Author(s)
The manuscript requires minor editing of English language by a native person
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
Dear Authors,
Authors worked really hard to find the most accurate modeling methods though various process are in the flue gas desulfurization process.
1. Pages on 4
Please could you add more detailed explanation where you installed sensors of pH and flow rate?
2. Pages on 4
Please could you explain how many hours time delay occurs?
3. Pages on 8
Please could you explain what is different between a two-dimensional kernel and a unidimensional kernel?
4. Conclusion (Comments)
It will be really interested in what will happen your model and the actual control fields in the future.
Sincerely,
Author Response
Please see the attachment.
Author Response File: Author Response.docx
