Research on Prediction Method of Bolt Tightening for Aviation Components Based on Neural Network
Round 1
Reviewer 1 Report
The aim of this paper is to develop a neural network model to reduce inaccuracies and non-uniform bolt preloads caused by elastic interaction due to varying gap sizes and inclination angle in components. The model predicts the initial preload tightening required for each of the bolts in the tightening sequence to result in uniform final preload across all bolts.
The main contribution is the development of the model which results in a reduction in time and computing power that would be required if using alternative models or software analysis.
General Concept Comments
Article:
The number of bolts that require tightening is few, however the practical results are used to validate the FEA model well. The FEA model is then used to create the neural network model, however the predicted bolt tightening requirements are not validated through practical application. This would create a more complete paper.
Review:
The references and knowledge are sufficient and appropriate, however after conducting a literature search myself I have found additional literature which could be included to form a more complete review. This would be in the area of neural modelling of bolt tightening, as well as aerospace assembly requirements and procedures.
-
"Development of an intelligent bolt tightening system using neural network approach" by A.K. Prasad, S. Nanda, and A. Ravi Kumar. This paper presents a neural network-based approach for developing an intelligent bolt tightening system.
-
"Bolt Tightening Using Neural Network Models and Fuzzy Logic Control" by S.A. Sallam and M.A. Megahed. This paper proposes a neural network model to predict the required torque for tightening bolts, which is then used in a fuzzy logic control system to adjust the torque applied to the bolts.
-
"A Bolt Tightening System Based on Neural Networks and Expert Systems" by F. Yang, X. Wei, and J. Liu. This paper describes a bolt tightening system that uses a neural network model to predict the tightening torque and an expert system to control the tightening process.
It is not clear why none of the bolts in the practical tightening sequence for the initial results do not reach the 72kN tightening. Based on experience, the final bolt tightened in the sequence will reach the desired preload?
Author Response
Thank you for editor’ and reviewers’ opinions, these comments are very helpful to improve the quality of the manuscript. We have carefully revised our manuscript, further clarify the logic of writing for improving the quality of the manuscript.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This paper has a good organizational framework and a smooth grammar. The research topic has important engineering application value. However, there are still some key issues that need to be addressed with caution.
(1) The “gaps” should be prominently marked in the Figure 1.
(2) Why only two sizes of gap were considered?
(3) What is the impact of gap size changes on the best tightening sequences?
(4) How to prove that this method can be used in situations other than the current structure with eight bolts tightening?
(5) In Lines 232-236, the maximum error between the simulation value and the experimental value is less than 450, but the differences between any two average values of these residual preloads do not exceed 900. How can we prove that the advantages of this simulation prediction method are not affected by model errors?
(6) A structure with only eight bolts cannot reflect the significant advantages of this method. It is recommended to consider a structure with more bolts. Detailed discussion of this section can be completed through numerical simulation.
There are some incorrect words in the text. For example, ‘proload’ in Line 235.
Author Response
Thank you for editor’ and reviewers’ opinions, these comments are very helpful to improve the quality of the manuscript. We have carefully revised our manuscript, further clarify the logic of writing for improving the quality of the manuscript.
Author Response File: Author Response.pdf
Reviewer 3 Report
Dear Authors. I am sending the comments as an attachment.
Comments for author File: Comments.pdf
Author Response
Thank you for your opinions, these comments are very helpful to improve the quality of the manuscript. We have carefully revised our manuscript, further clarify the logic of writing for improving the quality of the manuscript.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors answered the questions well. I suggest accepting the paper.
Author Response
Thank you for your advice
Reviewer 3 Report
First of all, I must state that I do not consider the incorporation of the comments to be relevant. The revised document does not indicate the modified parts, so I can only briefly consider the addition of two sentences, points 5 and 6, and the modification (addition) of some figures.
Having studied the manuscript again, I have concluded that it is necessary to clarify (add) the following facts:
Nowhere is it stated that the aim is to achieve a uniform loading of the bolts with the forces shown in Fig.18 (I assume that the values given were determined in accordance with the relevant standard ...).
Consequently, it should be stated whether the bolts are tightened sequentially or directly to the maximum force value given in Table 3 (L369). As far as I know, in practice bolts are tightened sequentially and not immediately to the maximum value. This is not mentioned anywhere in the text.
In L200 it should be „After tightening bolt 1, tightening bolt 8 or bolt 2 have the same effect.“
The text (L190 to L203) states that after tightening bolt 1, the gap is increased elsewhere. I required this to be documented in the numerical simulation results by plotting the displacement fields. The figures where the maximum deformations are do not say anything about where the gaps are increasing or decreasing.
L204 and L205 state that the maximum force of 72 KN is set in 96 steps. However, it is not clear from the text whether this refers to setting the maximum force in each bolt to the maximum in 96 steps or whether the 96 steps refer to all bolts. For example, according to Fig.6, after 96 steps the tightening process of the group of screws under consideration is completed, so the 96 steps are probably divided between 8 screws, giving me 12 per screw.
In my opinion, in practice, in similar cases, the bolts are first tightened by free hand until contact is made with the surface, and only then a torque wrench is used. So it is not clear to me how a situation can arise where just before the end of the „Interval two bolts tightening“ (Fig.6), the maximum deformation has increased to almost 1mm (at about the 90 step). At that time, most of the bolts were already tightened to the maximum ...
L254 there is the wrong figure number. It should be "... from Figure 9 ..."
The settings of the tightening sequence in the numerical simulation need to be explained in more detail because otherwise it loses meaning for the reader and, in addition, the results cannot be verified.
In Fig.9, three variants of bolt tightening are compared. Since each series starts by simultaneously tightening a pair of bolts 1 and 5, the start of all results should be identical. From Fig.9, this corresponds to approximately 3 steps, which is 1/16 of the entire tightening process. Since four pairs of bolts are being tightened, this seems to me to work out to a maximum tightening of the bolts in four stages ...
In Fig.18, it is not clear to me why the values predicted by you are not constant and, on the contrary, the experimental measurement is more accurate. I guess this is a flipped description.
In Table 3 you state (for a gap of 0.1mm), that the forces in bolts 3 and 7 need to be set to the same value of 65 576 N. These values are lower than the predicted and experimentally determined value according to Fig.18. Is this correct?
In the case of a smaller gap (i.e. 0.08mm), according to Table 3, the force in bolt 7 needs to be set more than 10,000 N higher and the force in bolt 8 needs to be set 6500 N lower at the same time (compared to a gap of 0.1mm). Is this OK?
Therefore, based on the above information, I suggest adding more detailed information about the bolt tightening sequence in the numerical simulation. Concurrently, within the appendix, it should be documented how the forces change in each bolt at each step of the simulation. Given that several variations have been performed, it is sufficient to provide only one for either 96 steps (single bolts) or 48 steps (pairs of bolts). I recommend presenting this in the form of a table that shows what the set force value was in the bolt and how it changed in the next step(s).
Author Response
We feel great thanks for your professional review work on our article. As you are concerned, there are several problems that need to be addressed. According to your nice suggestions, we have made extensive corrections to our previous draft, the detailed corrections are listed in the document.
Author Response File: Author Response.pdf
