Finite Element Simulation of HELICOIL^® Inserts

Round 1

Reviewer 1 Report

Finite Element Simulation of HELICOIL® Inserts, the novelty should be the FEA of this structure; however, I am sorry to say the present form is not. The paper presents many equations on multilinear curve by the the minimization method, this is just a method that well developed. The paper in its present form is too lengthy. In the revision, I suggest to focus on  the FEA of this structure, with the points such as:

1. The modelling method of this structure;

2. What is the key challenge in modelling of this structure;

3. Mesh size sensitivity is very important, at least three mesh sizes must be used.

Author Response

Thank you for the stimulating recommendations. We try to explain more about the problem and the goal of the article.

Finite Element Simulation of HELICOIL® Inserts, the novelty should be the FEA of this structure; however, I am sorry to say the present form is not.

We agree, but the basic idea has already been published, see [1].

The paper presents many equations on multilinear curve by the the minimization method, this is just a method that well developed.

We agree, but this procedure has not yet been submitted for the helicoil insert. Compared to the classical approach, there is a difference: Points identified from the values of the partial objective function are gradually added to the multilinear curve.

Identification is not so easy because  the parameters have different effects on the resulting curve; see, e.g. Figure 12. Parameters i = 1 and i= 2 affect regions of about 10 s and are not easy to find. Parameters i=8 and i=9 affect the region of 1 s and can be easily found. It is also clear from the figure that changing the parameter i=1 will also affect the value of parameter i=2, and parameter i=2 will affect the value of i=3, etc.

The problem is that identification in this case led to a sawtooth shape of the multilinear curve, which varied depending on the number of points of the curve and the method used.

In addition, we expect that the method will also be useful in other areas.

* Identification of material parameters in models with a large number of parameters, for example, Anand material model, Three network material model, etc.

* Some material models contain optional - extensible parts. These parts increase the number of parameters and the level of approximation of these models, but their determination is very complicated. We assume that adding a point to our simple model may correspond to adding a degree of approximation to the model, and the procedure may be similar.

*For parametric optimization problems with a large number of parameters, a simple method of selecting parameters according to the sign of the gradient reduces the number of parameters to half.

The following sections appear relatively rarely in the literature in a similar context.

* Partial objective function.

* Use the penalization function for the definition of an objective function.

* Normalized gradient estimation, usually evolutionary or gradient approaches, is used.

* Reduction in the number of parameters or selection of parameters for a given cycle of identification.

These parts are described in the paper, as they are not described elsewhere. It is about 10 pages long.

The paper in its present form is too lengthy.

The article contains a rough description of the algorithm in the first part; this is different from commonly used methods. In each part, we tried to indicate the thought process that led to the design of individual modifications (the time correction function, the parameter correction function, the penalty function, the sign vector, the search direction, the adding functions, the selection functions). Algorithm 1 and Algorithm 2 further describe the individual steps in the solution. We considered modifying this section, based on the recommendation of the opponent, eventually we just added one image to the attachment and slightly modified the algorithms. Based on the recommendation of the next opponent, we modified the introductory part of the article and added a new background section, and an example of the so-Sawtooth waveform is called Appendix 1.

 In the revision, I suggest to focus on  the FEA of this structure, with the points such as:

Commercial software (MSC.Marc) was used to solve the FEA, and a similar procedure could be used in Ansys, for example. In MSC.Marc two approaches can be used: directly interface elements - special types of elements for simulating cohesive zone, or as contact type (contact stiffness - slight penetration may occur between the two bodies - soft contact). We used a standard feature already implemented in the software, a more detailed description can be found, e.g., in [2]. Therefore, we did not want to describe too much the theoretical background of the cohesive zone.

The analysis itself contains standard elements that are available in many software programs. Although the idea of using the cohesive zone is quite original, we did not find it sufficient for publication in this journal. Therefore, the next suggested procedure was to first design a traction curve using a multilinear curve and then design and test a theoretical model for helicoil inserts. The procedure for identifying the multilinear curve proved to be more complex, and the proposed algorithm could be interesting in the field of material parameter and model identification or optimization problems. Therefore, this paper was written.

1. The modelling method of this structure;

A linear elastic material model was used, and a standard contact type in the cohesive zone setting was used.

2. What is the key challenge in modelling of this structure;

From our point of view, it is the proposed algorithm and partial objective function. A common procedure is to apply constraints to a multilinear curve, which leads to the design of a mathematical model of the application. In our case, constraints were applied to the solution method, which is not very common.

3. Mesh size sensitivity is very important, at least three mesh sizes must be used.

We agree and have added to the discussion.

 

 

Reviewer 2 Report

In my opinion, the article is interesting. It deals with current scientific and technical topics.
The content of the article and the description of the models are clear and sufficient.
The combination of modern measurement methods such as MES and DIC deserves recognition.

The work requires only editorial and graphic correction.
1) Please note the inconsistency in the description of the axes in the graphs. Please standardize the size and type of the font as well as the location of the description. (for example, compare the descriptions in fig. 5 and fig. 9, and so on)
2) Moreover, in many places there is no space between the drawing caption and the text of the article.

Author Response

Thank you for the positive review.

In my opinion, the article is interesting. It deals with current scientific and technical topics.
The content of the article and the description of the models are clear and sufficient.
The combination of modern measurement methods such as MES and DIC deserves recognition.

The work requires only editorial and graphic correction.
1) Please note the inconsistency in the description of the axes in the graphs. Please standardize the size and type of the font as well as the location of the description. (for example, compare the descriptions in fig. 5 and fig. 9, and so on)

We agree, we have unified the font format and sizes for the images. We hope it will be better in the new version.

2) Moreover, in many places there is no space between the drawing caption and the text of the article.

Agreed, appropriate whitespace added.

 

Author Response File: Author Response.docx

Reviewer 3 Report

In the present study, numerical simulation method of HELICOIL® Inserts was investigated and two contact models were compared. General speaking, this paper is written well. But there are still some doubts that need to be clarified.

(1) There are many grammar mistakes in the whole manuscript. So this paper need to be polished carefully.For example, In [1] are published experiments based on compression and torsion tests.

(2) It is recommended that a flowchart is given to show the proceedure of determining the cohesive model parameters clearly.

(3)In Conclusions, the authors said “Due to the use of the HELICOIL® inserts in parts created by 3D printing, it would be appropriate to include time-dependent behaviour in the simulation model.” Why?

Author Response

Thank you for the positive review.

In the present study, numerical simulation method of HELICOIL® Inserts was investigated and two contact models were compared. General speaking, this paper is written well. But there are still some doubts that need to be clarified.

 

(1) There are many grammar mistakes in the whole manuscript. So this paper need to be polished carefully.For example, In [1] are published experiments based on compression and torsion tests.

Thank you for the notice; the article will be sent to MDPI for language checking.

(2) It is recommended that a flowchart is given to show the proceedure of determining the cohesive model parameters clearly.

Thank you for the notice, we were not able to create it in short form, so it was placed in the appendix.  We have added a procedure to solve the search direction, which also includes the calculation of the correction functions. Other parts of the algorithm are roughly described in Section 4.5.  It was corrected only a bit. We hope that this addition will be sufficient.

(3)In Conclusions, the authors said “Due to the use of the HELICOIL® inserts in parts created by 3D printing, it would be appropriate to include time-dependent behaviour in the simulation model.” Why?

In our opinion, there is no simple answer to this question. At the moment, we expect to propose a very simple model, and if we succeed in promoting its use, we will proceed with a more complex model. Below, I will give some arguments why I think this is a good idea (YES) and some arguments why it is not a good idea (NO). (YES) The experimental data show [1] that a similar behavior is expected for the material. (YES) We assume that it would be possible to start from a Kelvin material model, and its design and parameter identification would be interesting. (NO) For the implementation, it would no longer be possible to use the standard cohesive zone model but a ‘user-defined’ procedure. This would also limit the possible use in other software. (NO) The applicability of such a model would be very limited due to its complexity. So in summary: I think it would be an interesting problem, but its economic benefit or applicability in practice would probably be negligible.

Author Response File: Author Response.docx

Reviewer 4 Report

Finite Element Simulation of HELICOIL Inserts – J. Rojicek et al.

General Comments: The paper attempts to describe a cohesive-based analysis of the Helicoil insert. However, the paper is very poorly written, especially the Introduction, and it has been very hard to follow. The manuscript may only be accepted after a thorough revision.

Specific comments:

1.      Introduce HELICOIL in a brief manner first, followed by methods described in the literature.

2.       Clearly state the objective of the paper prior to elaborating on the literary work

3.      The content seems to have been copy-pasted here. The first sentence of section 2 – Materials and Methods reads “This chapter describes best …”

4.      An introduction to the methods section is missing. It is hard for the reader to follow the multilinear curve that has been optimized using a cost function. Re-write this section after carefully describing the motivation first.

5.      What is the penalty stiffness utilized in your bilinear model?

6.      In the composite realm, the cohesive zone depicts a traction-separation curve with the area beneath the curve indicating the total strain energy released. This has no bearing on the global force-displacement curve. How are you correlating the force-displacement curve to the cohesive parameters here?

7.      What are the meshing guidelines for the cohesive parameters? The cohesive model is extremely mesh dependent.

 

Author Response

Thank you for the stimulating comments.  We tried to modify the introduction and better explain the problem (Chapter Background, Appendix 1).

The paper attempts to describe a cohesive-based analysis of the Helicoil insert. However, the paper is very poorly written, especially the Introduction, and it has been very hard to follow. The manuscript may only be accepted after a thorough revision.

Specific comments:

1. Introduce HELICOIL in a brief manner first, followed by methods described in the literature.

We agree, a paragraph at the beginning and a few sentences were added to clarify the use in the article. The order of the individual paragraphs was slightly modified, so that the logical continuity of the individual parts was improved.

2. Clearly state the objective of the paper prior to elaborating on the literary work

Added a task definition section, including simplifying terms.

 

 

3. The content seems to have been copy-pasted here. The first sentence of section 2 – Materials and Methods reads “This chapter describes best …”

      Yes, it is true. First, a rough text was written which was then inserted into the magazine template. We apologize for any confusion caused by this. The chapter has been fixed, and we hope it will be better.

4. An introduction to the methods section is missing. It is hard for the reader to follow the multilinear curve that has been optimized using a cost function. Re-write this section after carefully describing the motivation first.

The Introduction has been changed including the addition of a picture. A Background Section has been added to explain why the standard algorithm was not used. Added Appendix 1, where there is a simple example of a problem that appeared in some cases when solving.

5. What is the penalty stiffness utilized in your bilinear model?

When solving the given problem, we tried to distinguish between the search curve and the parameter search method. The curve itself falls within the area of identification of the mathematical model of the Helicoil insert. Therefore, the method is penalized - the value of the objective function, or parameters in individual cycles. We tried to influence the method of identification, not the curve itself.

6. In the composite realm, the cohesive zone depicts a traction-separation curve with the area beneath the curve indicating the total strain energy released. This has no bearing on the global force-displacement curve. How are you correlating the force-displacement curve to the cohesive parameters here?

Yes, at first, it was assumed that the correlation between these data is sufficient to identify the traction-separation curve. The DIC method was used to validate the result. Although the result turned out to be much better than with a fixed connection, the difference compared to the measurement is still significant. The shape of the curve is not based on any mathematical model. The initial estimate was made based only on visual comparison, and it is clear from the results that it is not correct. The same procedure can also be performed for a simple linear approximation. Then a bilinear model would be solved in the first step.

We considered using the so-called symbolic regression on the detected curve, which would reduce the number of parameters and simplify their identification. The mathematical model of helicoil insert itself will hopefully be the next step of the solution. So the shape of the curve is not based on any physical theory, we want to do that when designing a mathematical model.

7. What are the meshing guidelines for the cohesive parameters? The cohesive model is extremely mesh dependent.

We agree and have added to the discussion.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors have revised and improved the paper according to previous comments. It is recommended for publication.

Reviewer 4 Report

The revised manuscript has addressed lot of concerns that had risen during the review. This draft is recommended for publication w/ minor changes.