A Deep Learning-Based Parameter Prediction Method for Coal Slime Blending Circulating Fluidized Bed Units
Round 1
Reviewer 1 Report
This manuscript is a nice work dealing with deep learning-based method for predicting parameters of coal-water slurry blending units.
THE WHOLE WORK IS VERY INTERESTING AND COULD HAVE POTENTIAL APPLICATIONS.
POINTS FOR IMPROVEMENT:
1. What's the limitations of the proposed method?
2. Please, address the on line control problem by defining manipulating variables, etc
3. Please, propose future directions for work.
Author Response
Response to Reviewer 1 Comments
Point 1: What's the limitations of the proposed method?
Response 1: The main limitation of this study is that the prediction method does not consider coal quality variation and coal slurry blending ratio. We have added the following in the summary section.
The limitation of this study is that the effect of coal quality variation and coal slime blend-ing ratio on model accuracy is not considered。
Point 2: Please, address the on line control problem by defining manipulating variables, etc
Response 2: We are very grateful to the reviewers for their suggestions. The prediction model in this paper predicts the target bed pressure from real-time data. The predicted value reflects the material condition in the future. The site operator uses the predicted information to make appropriate control adjustments, thus improving the safety of the unit operation. We have described this section in the Abstract and Introduction. The additions are shown below.
Abstract: An accurate bed pressure prediction model can reflect the future material quantity in the furnace, which helps adjust the operation of the unit timely.
Introduction: Accurate prediction models improve the safety of unit operation. By obtaining future information through forecasting, operators can make timely adjustments to their operations, thereby reducing accidents.
Point 3: Please, propose future directions for work.
Response 3: Thanks to the reviewers for their comments. We have included this section in the Conclusions section.
Future work includes the method to characterize coal quality data and coal slime blending ratio for the deep learn-ing model.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Comments applsci – Learning-based method - CFD
Manuscript ID: applsci-1760719
Title: A deep learning-based method for predicting parameters of coal-water slurry blending units
General comments:
In this paper, the authors proposed a prediction method for bed pressure fluctuations in circulating fluidized bed (CFB) reactors. The numerical approach was validated by using experimental operation data issued from Linhuan Zhongli 1 # 330mw CFB unit installed in China, treating coal-water slurry blending. The method includes an improved Attention layer, motivated by the successful application of Temporal Pattern Attention in the literature. The manuscript reports relevant results concerning the operation of CFD for different applications but also for applying in-depth learning on an industrial scale. The paper is well structured and duly illustrated. However, some concepts have been taken for granted, which makes reading the manuscript somewhat difficult. Some points must be clarified and a revision is needed before the paper could be considered for publication in “Applied Sciences”.
Here I include some comments to help the authors in revising the text of the manuscript and improving it.
Specific comments:
- Please, check the manuscript for some language imperfections.
- About the title: please, consider mentioning “circulating fluidized bed” in the title
- Graphical Abstract and Highlights: not provided.- Keywords: O.K.
- Abstract:
- Despite its required brevity, the abstract has the most impact on the average reader. Abstracts of papers published in the journal "Applied Sciences" should attract scientists across a wide spectrum of research areas as potential readers, and it is important to precisely define the domain of study of the manuscript. Then, please, introduce better the circulating fluidized bed application as used in the paper and the relevance of coal-water slurry blending for the operation of such a reactor. In the present version of the abstract, the CFD + coal-water slurry blending and fossil fuels are mentioned in a sentence. Improve the writing.
- Please, also add numerical findings in the Abstract.
- Introduction:
- Page 1, lines 32-33: “If the coal-water slurry is not correctly disposed of, it will harm the environment”. Please, be more precise about the potential damage to the environment.
- Page 1, line 38: As many mentions to different elements of the reactor and devices were made in the operation description, adding a global scheme of the CFB and auxiliary devices if necessary would be useful for readers to understand the eventual drawbacks caused by coal-water slurry blending.
- Page 2, line 51 …flowrate was (past tense)
- Page 2, lines 53-68: Please describe reported previous research in the past tense, and carefully revise the use of different conjugations in this segment of the text.
- Page 2, line 74: “in industrial processes” (eliminate “the”)
- Page 3: The description of the paper's content after presenting the set of main contributions of the manuscript is well appreciated.
- Check mechanistic model/mechanism model (it could mean different things !)
- Background Description
- Page 3, line 137: define better the pant-leg design for CFB, perhaps citing a previous work by the authors as a reference and including a brief definition.
- Line 139: Please, put this particle size in a numerical format of an appropriate range of particle size to be appropriately treated in the CFB.
- Line 142: The elutriation phenomena should be better defined. The work by Smolders and Baeyens can help as a reference: Smolders and Baeyens (1997), Powder Technol. Volume 92, Issue 1, June 1997, 35-46 https://doi.org/10.1016/S0032-5910(97)03214-2
- Figure 1: Wind? Air, fluidizing agent is a more appropriate denomination
- Page 4, lines 148-149: What does it mean "particular circumstances"?
- Page 4, lines 167-168: This sentence is not clear, please check it and eventually rephrase it to clarify the meaning of y(t+k)
- 2.2 Long Short-term Memory
- This section is O.K. However, check that all variable symbols and operators/functions are defined in the text or, it would be better to add a complete Nomenclature section.
- Eq. 8, page 5: check for format
- Page 5, line 186: "The" can be eliminated before the sigma symbol
- 2.4 Attention Layer
- Page 6, lines 225-226, check the style for "a, b and c.
- Fig. 3: Following the layer structure diagram and from the Timestep block to the HC matrix, could it be possible to indicate the arrowheads?
- Page 7: Fig. 4. When mentioning this figure in the text, could the authors please precise how was the influence of the data after convolution operation on the input at the later time evaluated/defined?
- 3.2 Segmented Training
- In this work, a segmented ... was used /applied. (“The paper” does not "use" anything)
- 3.3 Performance Measures: “Assessment” seems better than “Measures”.
- 3.4 Data Standardization and Differential Processing
- Page 12, line 357: Please, rephrase.
- 3.5 Variable Filtering
- Page 12, from line 372: Correlation Coefficient values seem to indicate a weak relationship between variables concerned. Is that correct?
- Page 12, Table 1: The order of importance, is it a result of the strength of the relationship?
- Page 12, line 375: Please, do not use the expression "this paper uses..."
- Page 13, line 392: volume? Are you referring to the volumetric flowrate?
4.1 Ablation Study 431
4.1.1. Differential Prediction Method
- Page 16, line 441: "It can be seen from the results". This sentence seems to have no sense here. Please, eliminate it.
5. Conclusions
- Page 20, line 546: Future work includes... (instead of "We hope to...")
- Add some relevant numerical findings to the Conclusions (section 5).
No further comments
Final assessment: Major revision required.
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The authors have considered the comments and suggestions made by this reviewer and significantly improved their manuscript. The paper is now suitable for publication in the journal Applied Sciences.
