On the Relationship of the Acoustic Properties and the Microscale Geometry of Generic Porous Absorbersâ€
Round 1
Reviewer 1 Report
This contribution shows an approach to gain insights into the relationship between the microscale geometry and the acoustic material parameters of a generic porous material using an artificial neural network. The applied global sensitivity analysis method allows to derive a one-to-one parameter impact relation, as well as their interdependencies.
The paper is quite well written; nevertheless, I stayed with some minor remarks, which are set out in the attached review.
MINOR REVISION
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
In this manuscript, the authors investigated the relationship between microscale geometry parameters and acoustic properties of complex porous sound absorbing materials through machine-learning method, which has necessary academic contributions and application reference values to the fields. The manuscript is well-written and can be recommended for accepting after a minor revision. My comments are listed as following:
1. Even if it is a simplified porous material structure, its geometric model is difficult to be directly established by program adaptively through simple structural parameter relations. In this way, is it necessary to manually modify the geometric model in thousands of iterative calculations?
2. In Fig. 2, the authors have fabricated dozens of samples for testing. Is it necessary? Is it possible to replace it with simulation results?
3. In order to further enhance the academic influence of the manuscript, the authors is suggested to add more detailed information about the machine-learning optimized method for the convenience of readers.
4. In terms of sound absorption, acoustic metamaterial is a hot topic in recent years and are closely related to this manuscript. So the authors need to make appropriate additions in the introduction (for example, Ann. Rev. Mater. Res. 47, 83-114 (2017); Compos. Struct. 239, 111978 (2020); Mater. Horiz. 9, 653-662 (2022); Adv. Mater. 33, 2104552 (2021)).
Author Response
Please see attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The manuscript (MS) describes acoustical absorption vs frequency with the models. The MS is but not an empirical study. Therefore, the model needs to be verified with the experiment for such a material in used in the different industrial fields.
The MS has a serious problem for writing style for a research article convenient for a publication in a journal.
The following marks would be taken into account of re-writing the entire MS:
The abstract does not reflect the results of the present study and includes aim of the study, problem of or difference from other studies published already and importance, material and methods, main results, recommendation.
The introduction has many repeats trough the entire MS, and must be shortened, and must be more clear and compressible. There are many citations such as Section XX, instead of goal or aim of the present study. The authors do avoid use "We" or "the authors".
Material and Methods has again the same problems aforementioned. Given in Introduction was repeated hereby. The model choosing is subjected without the reasons and property to the designed cylinder to measure the absorption in the theory. There are many assumption to solve model giving references.
According to the journal the MS was submitted to, there is no applied science, and include only modeling. In general, each model type has limitations for suitability to the material considered in the present study. Overall, the model results needs in many case, the verification or validity as compared to the experimental studies.
The result presentation has again the same problems aforementioned. For instance, "The influence of the different geometry parameters on the acoustic model parameters is investigated using the GSA framework described in Section 2.4. Therefore, the SALib implementation in python is used [43]. The underlying Monte Carlo procedure employs a Quasi Monte Carlo approach with a Sobol sequence [44] for equally sampling the parameter space. A total of 1 106 Monte Carlo runs is performed. Since the used python framework assumes a scalar valued model output, the analysis is performed for each output separately. However, since all inputs are processed in parallel, this is assumed to not limit the applied method. The parameter bounds for all inputs X used according to Table 1 and scaled to a uniform distribution X U(0, 1). The results of the GSA with all inputs and for all outputs are shown in the following. Thereby, in Section 3.1 the results for the main effect and total effect of all inputs are shown, the results for the second order effect are shown in Section 3.2. ". This is completely related to material and Methods.
The Discussion is overall not well-completed for comparison with other studies published. Usage of "Section XX" is not good manner to write a paper. The Discussion is too weak must be more compressible including the other publications. If no similar studies available, it is recommended to unit Results and Discussion.
Overall, the MS could be more attractive to address to wider audience.
After major revision, the MS could be submitted to journal “Acoustics” of MDPI, instead of “Applied Science” since the MS lacks of experimental application to valid model results for the ground-truth of the MS data output.
Author Response
Please see attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
The authors revised significantly the manuscript eventhough the context of the introduction is same to the previous version.
