Estimating Calorific Value of Coal Using Laser-Induced Breakdown Spectroscopy through Statistical Algorithms: Correlation Analysis, Partial Least Squares, and Signal-to-Noise Ratio
Round 1
Reviewer 1 Report
The calorific values estimated on the basis of carbon elemental concentrations using three different analytical methods showed that the CA method is the most accurate (with a coefficient of determination of 0.971). The difference in estimating the calorific value of coal by the CA method and the calorific value of coal calculated by means of the Dulong equation is within the range of 0.16% - 6.99%, which is small, and it is quite acceptable, because in terms of the value of the multiple correlation coefficient makes up over 90%.
The methodology of the experiment was unqualified.
Author Response
I appreciate your comments.
Reviewer 2 Report
The title of submitted manuscript is Estimating calorific value of a coal using laser-induced breakdown spectroscopy by Kim et al. In the present paper, authors compared different statistical algorithms including CA, PLS, and SNR for estimating the calorific value of the coal based on quantitative analysis of elements. The intensity-concentration calibration curves were obtained for the elements in coal 20 and the elemental concentration correlations were estimated on the basis of calibration curve of each element. In my opinion, the novelty of the paper is quite good. However, major revision is necessary to improve the paper. My comments are as follows,
General comments,
1. What is a standard method used for estimation of calorific value of a coal. Authors should compare the result of calorific value estimation using the standard method and developed methods present in this work.
2. Authors does not explain in detail how to estimate the calorific value using the LIBS spectra obtained from the coal sample.
3. Please show in the figure the emission spectra of C, H, S, and O obtained from the coal samples used in this work? Emission lines of C and H are usually difficult to be identified using the LIBS due low-atomic mass and matrix effect characteristics of the coal sample.
Detailed comments,
- Title, please indicate three statistical algorithms used.
- Abstract, please explain the results obtained by using 3 statistical algorithms.
- Introduction, please review the standard method usually use to estimate the calorific value of coal and explain pros and cons of the methods.
- Results and discussion, please include in the figure the emission spectra of C, H, O, and S obtained from the coals.
- All figures did not match with explanation in the paragraphs.
- Figure 2, what calibration curves of elements used for Fig.2 and how to plot the curve, please show the example of the figure of calibration curves of elements.
- Figure 3, what is standard method used to measure the concentration in x-axis? Please add Carbon in x-axis. How many repetitions of sample bombardment each point plotted in the graph.
- Figure 4, The graph is bad. Could you distinguish the H lines from the sample with the H emission lines from air surrounding gas. Why is the H intensity almost the same with different concentrations.
- Figures 5 and 6, the graphs are also bad. What is the meaning of Figs. 4,5,6?
- Figs 7,8,9,10. Please explain in detail how to estimate the concentrations of C, H, O, and S using the emission spectra obtained by LIBS. Concentrations of elements in LIBS are usually at least estimated by calibration free if no elemental reference used.
Author Response
General comments,
- What is a standard method used for estimation of calorific value of a coal. Authors should compare the result of calorific value estimation using the standard method and developed methods present in this work.
Answer: Bomb calorimeter is a standard method used for estimation of calorific value of a coal at coal fired power plant. At table 2,3 and 4, the result of calorific value estimation using the standard method which is presented as “Chemical analysis” was compared to developed method presented as “Calorific value”. The error between the calorific value using standard method and developed method was presented in table2,3 and 4 as “Error”.
“The calculated calorific value was compared to the calorific value obtained from the bomb calorimeter presented as chemical analysis.” was added at line 324.
- Authors does not explain in detail how to estimate the calorific value using the LIBS spectra obtained from the coal sample.
Answer : The details to estimate the calorific value using the LIBS spectra was explained in the “4. Results and discussion”. The elemental concentration of seven different coal samples was determined from the conventional industrial chemical analysis which was represented in the Table 1. The correlation between the intensity of LIBS signal and the concentration of each element in the coal samples were analyzed and presented as figures (Figure 6 for C, Figure 7 for H, Figure 8 for O, Figure 9 for S). Using the correlation, the concentration of each element contained in the coal samples were measured and used to evaluate the Higher heating value (HHV) by Dulong’s equation. To enhance the understanding of contents, we added the following sentence in the manuscript (page 12 line 313).
- “Dulong’s equation uses the weight fraction of carbon, hydrogen, oxygen, and sulfur to predict the higher heating value of the coal which can be estimated as the calorific value.” Please show in the figure the emission spectra of C, H, S, and O obtained from the coal samples used in this work? Emission lines of C and H are usually difficult to be identified using the LIBS due low-atomic mass and matrix effect characteristics of the coal sample.
Answer: We included 4 emission spectra for carbon, hydrogen, oxygen, and sulfur as Fig. 5 in the manuscript. We obtained the emission spectra of at 247 nm for C and at 655 nm for H which are shown in the below figure. Both emission lines were clearly identified from adjacent emission lines.
- emission spectra for C b) emission spectra for H
- Title, please indicate three statistical algorithms used.
Answer: Based on the comments of the reviewer, we change the title as follows: “Estimating calorific value of a coal using laser induced breakdown spectroscopy and statistical algorithms: correlation analysis, partial least squares and signal-to-noise.”
- Abstract, please explain the results obtained by using 3 statistical algorithms.
Answer: In line 22, results from the 3 statistical algorithms were added. “The intensity-concentration calibration curves were obtained for the elements in coal and the elemental concentration correlations were estimated on the basis of calibration curve of each element. CA showed higher accuracy compared to PLS and SNR.”
- Introduction, please review the standard method usually use to estimate the calorific value of coal and explain pros and cons of the methods.
Answer: From line 39 to 45, standard method usually used to estimate the calorific value of a coal and its disadvantages were explained. “The LIBS can be used for analyzing solid [5], liquid [6], and gas [7] samples with the detection limit of ppm levels. Conventionally, loss on ignition method and CHN (carbon, hydrogen, and nitrogen)-analyzer were used for the knowledge of the coal component. In the aforementioned methods, the coal samples must be pre-treated and the analysis takes more than 10 hours. On the contrary, LIBS need little or no sample preparation process compared to conventional loss on ignition method and less time to analyze the result.”
- All figures did not match with explanation in the paragraphs.
Answer: We rearranged the figures well matched with the relevant explanation.
- Figure 2, what calibration curves of elements used for Fig.2 and how to plot the curve, please show the example of the figure of calibration curves of elements.
Answer : The details to make the calibration curves of elements was explained in the “2.2 Sample preparation” and “4. Results and discussion”. The elemental concentration of seven different coal samples was determined from the conventional industrial chemical analysis which was represented in the Table 1. The correlation between the intensity of LIBS signal and the concentration of each element in the coal samples were analyzed and presented as figures (Figure 6 for C, Figure 7 for H, Figure 8 for O, Figure 9 for S).
- Figure 3, what is standard method used to measure the concentration in x-axis? Please add Carbon in x-axis. How many repetitions of sample bombardment each point plotted in the graph.
Answer: For figuere3, 1 repetition at each point of the sample was used and accumulated.
- Figure 4, The graph is bad. Could you distinguish the H lines from the sample with the H emission lines from air surrounding gas. Why is the H intensity almost the same with different concentrations.
Answer: We included in H spectra of 7 coal samples in Fig. 5. The concentrations of hydrogen in the coal samples do not show much difference from the 4.81 to 5.27%. Therefore, the H intensity looks identical in the spectra plot. We included in the manuscript. In this experiment we did not isolate the effect of atmosphere. We will carry out the next experiment in an isolated chamber as a next work.
- Figures 5 and 6, the graphs are also bad. What is the meaning of Figs. 4,5,6?
Answer: Figure 4 is for the determination of delay time for analysis of coal sample. Figure 5 and 6 are correlation between the intensity and C, H concentration with maximum R2 in CA, PLS.
- Figs 7,8,9,10. Please explain in detail how to estimate the concentrations of C, H, O, and S using the emission spectra obtained by LIBS. Concentrations of elements in LIBS are
usually at least estimated by calibration free if no elemental reference used.
Answer: After measuring the intensity according to the concentration at the designated wavelength for each element. And draw the calibration curve of the intensity according to the concentration of the element.
Author Response File: 
Author Response.docx
Reviewer 3 Report
This Manuscript discusses the improvement of the LIBS method for determining the calorific value of coal.
The authors conducted and processed a large number of experiments.
At the same time, there are a number of questions and comments on the text of the Manuscript:
1. In the Introduction there is no detailed comparative analysis of the advantages and disadvantages of other methods for determining the calorific value of coal compared to LIBS.
2. Why was the integration time set to 1.05 ms?
3. What spectrograms were obtained from the spectrometer? Where are they? Without spectrograms, the justification for the choice of wavelengths looks unconvincing.
4. How exactly did the authors obtain calibration curves: at a single wavelength or in the spectral range?
5. Cited literature of 1985-2017. There are no sources of information for the last five years.
There are also design flaws in the Manuscript:
1. There are two pairs of figures numbered 1 and 2 in the Manuscript.
2. In lines 107 and 190-191 there is an actual repetition of the text.
Author Response
Reviewer 3
- In the Introduction there is no detailed comparative analysis of the advantages and disadvantages of other methods for determining the calorific value of coal compared to LIBS.
Answer: From line 39 to 42, the disadvantages of conventional method was mentioned. And the advantages of LIBS compared to other method were added as ”On the contrary, LIBS need little or no sample preparation process compared to conventional loss on ignition method and less time to analyze the result.”
- Why was the integration time set to 1.05 ms?
Answer: Typical integration time (gate width) used in LIBS range from 1 to 10ms. Integration time with delay time also affects the signal to noise ratio of spectrum. In this manuscript, only the effect of the delay time was considered. For the integration time, 1.05ms was borrowed from the reference paper number 25 in manuscript which was about the sulfur detection in soil by LIBS. Further studies would be needed to find the optimal integration time for coal analysis using LIBS.
- What spectrograms were obtained from the spectrometer? Where are they? Without spectrograms, the justification for the choice of wavelengths looks unconvincing.
Answer: The full spectrum from 180 to 884nm were obtained from the spectrometer for 100 points at 7 different sample. From the total of 700 spectrum, the averaged value of 100 spectrum was calculated for each sample. Shown below(left) is the averaged value of 100 spectrum at 7 different sample of full spectrum obtained from the spectrometer. And the figure on the right shows the spectrum at 247nm which is the representative wavelength of carbon.
We added the 7 emission spectra of 7 coal samples at 247nm for C, 655nm for H, 844nm for O, and 544 for H, full spectrum plot were in Fig. 5
- How exactly did the authors obtain calibration curves: at a single wavelength or in the spectral range?
Answer: Calibration curves were obtained at a single wavelength with the average intensity of 100 points for each sample. The wavelength of C, H, O and S at 247, 656, 844 and 545nm were used respectively.
- Cited literature of 1985-2017. There are no sources of information for the last five year
Answer: Reference number 25(2021) and 26(2020) were added.
- There are also design flaws in the Manuscript:
Answer: Above mentioned flaws were fixed.
- There are two pairs of figures numbered 1 and 2 in the Manuscript.
Answer: In figure 1, (a) represents the laser generation part and (b) represents the signal collection part. Line 89 “In Fig. 1, the LIBS system is composed of two main parts, which are laser generation part, and signal collection part.”was replaced as “In Fig. 1, the LIBS system is composed of two main parts, which are laser generation part , Fig 1 (a), and signal collection part, Fig 1 (b).”
In line 119 and 120, explanation about figure 2 left and right was added.
- In lines 107 and 190-191 there is an actual repetition of the text.
Answer: Yes, line 107 was deleted.
Author Response File: Author Response.docx
Reviewer 4 Report
Laser induced breakdown spectroscopy is applied in the qualification of the calorific value of the coal. The contained element in coal is respectively measured by CA model, PLS model, and SNR model. The obtained results are meaningful, but the manuscript should be revised before the further consideration. Followings are the detailed comments.
1. Inserted figures should be replaced by high-resolution version.
2. In Table.1, if the concentration is expressed in weight fraction?
3. The figure number in manuscript is really confusing.
4. In the figure of Variation in coefficient of determination and error, the top x-axis is dislocated.
5. What is the diameter of the laser beam used in your ablation experiment?
6. Whether the ablation-induced plume will act with the oxygen in the air and then affect the measurement? It should be illustrated in detail.
7. In Fig.9 and Fig.10, the relative error between measured and estimated value cannot be accepted.
Author Response
Reviewer 4
- Inserted figures should be replaced by high-resolution version.
Answer: Figure 1 and 2 was replaced by high resolution version.
- In Table.1, if the concentration is expressed in weight fraction?
Answer: Yes. The concentration was expressed in weight fraction. “Concentration(%)” in table 1 was replaced by concentration(wt%)
- The figure number in manuscript is really confusing.
Answer: Rearranged the figures after the relevant paragraph.
- In the figure of Variation in coefficient of determination and error, the top x-axis is dislocated.
Answer: Replaced the dislocated x-axis to the corrected version
- What is the diameter of the laser beam used in your ablation experiment?
Answer: Laser beam diameter was not measured.
- Whether the ablation-induced plume will act with the oxygen in the air and then affect the measurement? It should be illustrated in detail.
Answer: Since the experiment was conducted under atmospheric conditions, there would be and effect from the oxygen in the atmosphere. To confirm this effect, the next study will be conducted in nitrogen and helium condition.
- In Fig.9 and Fig.10, the relative error between measured and estimated value cannot be accepted.
Answer: The accuracy of relative error looks bad as you indicated. However, the result of CA is almost on the line. The minimum error is obtained with CA. We included in the manuscript.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
I can accept final revised manuscript
Reviewer 3 Report
In the new version of the Manuscript, most of the shortcomings have been corrected. The authors gave adequate answers. There are some small design notes: where are the names of Figures 3 and 4?
Reviewer 4 Report
All the questions have been well addressed by the authors, and it can be considered for publication now.
