Study of the Optimal Waveforms for Non-Destructive Spectral Analysis of Aqueous Solutions by Means of Audible Sound and Optimization Algorithms

Round 1

Reviewer 1 Report

In this study, the authors proposed a biomimetic approach in the audible range to overcome some limitations in acoustic sensors such as penetration depth and the use of coupling gels. A total of 364 samples of water and fructose solutions with 28 concentrations between 0 g/L and 9 g/L have been analyzed inside an anechoic chamber using audible sound configurations. The spectral information from the scattered sound is used to identify and discriminate the concentration with the help of an improved grouping genetic algorithm that extracts a set of frequencies as a classifier. An Extreme Learning Machine implements the fitness function able to classify the mixtures.

The obtained results are promising and reasonable. In addition, the paper’s subject could be interesting for readers of journal. Therefore, I recommend this paper for publication in this journal but before that, I have a few comments on the text that should be addressed before publication:

 

Comments:

1)In my opinion, figure 4 doesn’t include significant data. Just 2 laptops are seen in this figure. Please substitute it with a better experimental image or remove it.

2)The axis titles in figure 3 are so small and are not readable. Please enlarge them.

3)The subfigures in figure 5-8 are small. Please enlarge them.

4)The font size in Tables are not compatible with the text. Please correct them.

5)In page 15 line 480 sentence “This work has allowed us…”: Please rewrite it in this way “This work allowed us…”

6)Authors have used 80% dataset for training and 20% for testing machine learning. Why didn’t use a part of dataset for validating (validation set)?

7)How authors evaluated the accuracy of machine learning performance?

8)Since recently it has been proved that artificial intelligence (AI) has a numerous applications in all of engineering fields, I highly recommend the authors to add some references in this manuscript in this regard. It would be useful for the readers of journal to get familiar with the application of AI in other engineering fields. I recommend the others to add all the following references, which are the newest references in this field of computer engineering [1], electrical engineering [2], petroleum engineering [3], fluid mechanic engineering [4,5], energy engineering [6]

In recent years, in addition to Venturi he authors are remore new references in the paper in the mentioned field. Some suitable and new references are listed in the following:

[1] Jahanshahi, A.; Sabzi, H.Z.; Lau, C.; Wong, D. GPU-NEST: Characterizing Energy Efficiency of Multi-GPU Inference Servers. IEEE Comput. Archit. Lett. 2020, 19, 139–142.

[2] Jahanshahi, A. TinyCNN: A Tiny Modular CNN Accelerator for Embedded FPGA. arXiv 2019, arXiv:1911.06777.

[3] Roshani, M.,, et al., 2021. Evaluation of flow pattern recognition and void fraction measurement in two phase flow independent of oil pipeline’s scale layer thickness. Alexandria Engineering Journal.

[4] Nabavi, M.; Elveny, M.; Danshina, S.D.; Behroyan, I.; Babanezhad, M. Velocity prediction of Cu/water nanofluid convective flow in a circular tube: Learning CFD data by differential evolution algorithm based fuzzy inference system (DEFIS). Int. Commun. Heat Mass Transf. 2021, 126, 105373.

[5] Roshani M, Application of GMDH neural network technique to improve measuring precision of a simplified photon attenuation based two-phase flowmeter. Flow Measurement and Instrumentation. 2020

[6]Arabi, M.; Dehshiri, A.M.; Shokrgozar, M. Modeling transportation supply and demand forecasting using artificial intelligence parameters (Bayesian model). J. Appl. Eng. Sci. 2018, 16, 43–49.

 

 

Author Response

Please see attached file.

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript describes a new method of determining the fructose concentration in liquid mixtures with fructose concentration in the range 0–9 g/L based on the spectral analysis of audible scattered signals measured in an anechoic chamber by reducing the problem to a set of 9 frequencies in the range 3–15 kHz. 

I think that the results obtained have limited applicability, because the specific features of the spectra observed by the authors are determined mainly by the peculiarities of experimental configuration and may change depending on the installation parameters, in particular, the material and volume of the cylindrical glass and the volume of the water solution. In addition, when describing the anechoic chamber, the authors said nothing about its floor that to my mind also affects the results obtained. 

In addition, the symbols of the coordinate axes of the figures are too small. Figure 4 is uninformative and should be deleted.

Section 3 "Algorithm for Clustering Problem" is too general, it repeats the material of the references. It should be focused on the specific features of the author's algorithm rather than on the enumeration of the general principles of algorithms for solving clustering problems and should be shortened.

The language should be improved: 

Line 31: with other methods like electroacoustic measurements

Lines 34-35: are devoted to ultrasound techniques and devices, due to its larger energy and bandwidth than sounds in the audible range.

Line 39: 0,2%

Line 68: are based in the previous determination

Line 75: we shall show

Line 88: with regards the determination

and so on.

Based on the foregoing, I consider that the manuscript can be published after careful consideration of the foregoing comments.  

  

Author Response

Response to comments from Reviewer 2

The manuscript describes a new method of determining the fructose concentration in liquid mixtures with fructose concentration in the range 0–9 g/L based on the spectral analysis of audible scattered signals measured in an anechoic chamber by reducing the problem to a set of 9 frequencies in the range 3–15 kHz.

I think that the results obtained have limited applicability, because the specific features of the spectra observed by the authors are determined mainly by the peculiarities of experimental configuration and may change depending on the installation parameters, in particular, the material and volume of the cylindrical glass and the volume of the water solution. In addition, when describing the anechoic chamber, the authors said nothing about its floor that to my mind also affects the results obtained.

We are appreciating the reviewer's observation which is necessary and convenient for us to clarify. Obviously, any modification of the experimental arrangement and boundary conditions changes the results of the experiment. For example, no perfect anechoic chamber or microphone is possible, and any liquid container will change the frequency response. In most spectroscopic ultrasound studies, experimenters try to minimize this variability calculating the speed of sound. However, such approach is complicated by numerous theoretical and statistical uncertainties, as commented in our literature survey. Thus, our work describes a new general method, which tries to bypass the calculation of the speed of sound, using only relative, not absolute, measurements. In this way, the differences in the spectral response due to the experimental conditions can be controlled by reference to a standard sample, for example, pure water, in the same conditions. The use of artificial intelligence algorithms for pattern recognition makes far easier to identify key spectral information that permits sample classification and then fructose concentration without a rigid specification of laboratory conditions. This biomimetic approach in essence is not new and can be related to the differences between a physical treatment of sonar source identification in contrast with the training of a human sonar operator which must listen to ocean sounds in very variable circumstances.

 

In addition, the symbols of the coordinate axes of the figures are too small. Figure 4 is uninformative and should be deleted.

Thank you for your appreciation. Figures have been enlarged and the font of the axis titles enlarged as requested by the reviewer. We hope that the new figures will satisfy all readers.

We have considered removing the image from the manuscript because, as indicated by the reviewer, it does not provide added information with respect to the text. All figure references have been verified and updated.

 

Section 3 "Algorithm for Clustering Problem" is too general, it repeats the material of the references. It should be focused on the specific features of the author's algorithm rather than on the enumeration of the general principles of algorithms for solving clustering problems and should be shortened.

We have slightly reduced subsections 3.1 and 3.2 because they are, as the reviewer points out, general descriptions. We think that it is not appropriate to shorten it further because it could hinder the understanding of some readers in their first reading. On the other hand, we have elaborated subsection 3.3, which focuses on the specific characteristics of the algorithm developed by the authors for the application described in the manuscript.

“Not all groups of an individual are useful for classification, only those with better accuracy. Note that considering a specific individual, each feature of the 655 is only present in a single group. In the GGA fitness function, the ELM algorithm is applied over each group of the individual to classify the testing set data from the knowledge of the training data. The group with the best classification accuracy is selected as a candidate classifier. Also, the fitness of the individual takes the value of the classification accuracy of this highlighted group, which is the best accuracy obtained among all the groups of the individual.”

 

The language should be improved:

Line 31: with other methods like electroacoustic measurements

Lines 34-35: are devoted to ultrasound techniques and devices, due to its larger energy and bandwidth than sounds in the audible range.

Line 39: 0,2%

Line 68: are based in the previous determination

Line 75: we shall show

Line 88: with regards the determination

and so on.

 

Thank you very much for these remarks. We have corrected these errors and reviewed the draft manuscript to improve the quality of the editing. We surely appreciate the detection of these language errors.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

All the comments have been addressed correctly. In my opinion, the paper in the present form is ready for publication.

Author Response

Thank you very much for your dedication. 

Reviewer 2 Report

The authors has significantly improved the text of the manuscript. It can be recommended for publication after two more mandatory corrections: 

Line 221: "in the frequency range 0 Hz–22.050 kHz" 

Line 343: "over the audible frequencies range [0—22.05] kHz"

The lower boundary cannot be 0 Hz. 

Author Response

We have been able to improve the text of the manuscript thanks to your comments and effort. We sincerely thank you for your work.

Thank you very much for pointing out this detail, we completely agree. We have removed 0 Hz from the whole document and written instead the microphone limit (20 Hz).