Compilation and Extrapolation of Load Spectrum of Tractor Ground Vibration Load Based on CEEMDAN-POT Model

Round 1

Reviewer 1 Report

The manuscript describe extrapolation method of axial load induced by ground vibrations during different tractor operation conditions.

The content is really interesting for agricultural engineer who engage in tractor design.

But, the way you present your data and analysis result is not adequate.

So, you should modify the following point.

 

Line 120 and 168

Table 1 and Table 2 are the same. You should delete Table 2.

 

Line 376

It is difficult to understand the titles and legends in Figure 9.

You should enlarge font or change figure style.

I think Figure 9 has too many subplots.

 

Line 403

In Figure 11, 9 IMF components are derived from raw data.

Please describe each component meaning.

I think each IMF component can be linked with specific tractor motion, such as vertical, longitudinal or lateral motion.

 

Line 403

You should add unit and name for raw data.

Is raw data a axial load or acceleration?

Please be specific.

 

 

Author Response

Compilation and Extrapolation of Load Spectrum of Tractor Ground Vibration Load Based on CEEMDAN-POT Model

agriculture-2097165

Response to Reviewer #1 Comments

Thank you very much for your kindly comments on our manuscript. There is no doubt that these comments are valuable and very helpful for revising and improving our manuscript. In what follows, our responses are given point by point. We have uploaded a revised version with all the changes highlighted by using the track changes mode in MS Word.

Point 1:

Line 120 and 168

Table 1 and Table 2 are the same. You should delete Table 2.

The Authors’ Response: Thank you for pointing this out. We have modified the contents of Table 2, which is the specification parameters of the tractor in the experiment.

Original text:

Table 2. Field experiment condition parameters. Parameters Value Weather Sunny Atmospheric pressure ≥96.6kPa Environmental temperature 18.5 ℃ Gear and speed of field farming Gear: 2, 3, 4; Speed: 2~8 km h-1 Gear and speed of transportation Gear: 5,6; Speed: 10~30 km h-1

Corrected text:

Table 2. The specifications of the tractor in the experiment. Parameters Value External dimensions (mm) 4910×2430×3030 Vehicle weight (kg) 4980 Speed range (km·h-1) 2.84-38.58 Tire specifications 14.9-38 Rated power (kW) 103

Point 2:

Line 376

It is difficult to understand the titles and legends in Figure 9.

You should enlarge font or change figure style.

I think Figure 9 has too many subplots.

The Authors’ Response:

Appreciate for your advice. Per your comment, we have deleted all the raw data plots as provided in the previous figure. Instead, we keep the 6 representative raw data for presenting and comparing the data obtained in different operating conditions. As shown below, the corrected figure can now clearly display all the necessary information such as the legend, title, axis labels, and all the axis ticks.

Original text:

Figure 9. Original load data: (a) The original vibration load data curve of six kinds of ground, (b) Tractor rear axle (right) data curve for six kinds of ground.

Corrected text:

Figure 9. Tractor rear axle (right) data curve for six kinds of ground.

Point 3:

Line 403

In Figure 11, 9 IMF components are derived from raw data.

Please describe each component meaning.

I think each IMF component can be linked with specific tractor motion, such as vertical, longitudinal or lateral motion.

The Authors’ Response: Thank you for your questions and suggestions. In this paper, we collect the ground vibration acceleration data of the tractor when it is running on the ground. Due to the influence of the working environment, the signal will have noise interference. CEEMDAN decomposition algorithm is implemented on the basis of EMD algorithm (an adaptive decomposition method proposed for nonlinear and non-stationary signals). Based on the local characteristic time scale of the original signal, CEEMDAN decomposition algorithm decomposes the original signal into characteristic mode functions. Which means, the processed signal is decomposed into a series of IMF components from high frequency to low frequency. CEEMDAN method adds white noise (or IMF component of white noise) to the residual value every time the first order IMF component is calculated, calculates the mean value of IMF component at each step, and iterate step by step. The former IMF components cover the main vibration load information of the tractor during ground driving, while the latter IMF components and residual functions (IMF 19) mainly reflect the load change trend during driving.

 

Without an effective denoising method, the collected signals suffer from redundant and great noise interference. Applying a denoising algorithm can reduce the noises. We found and demonstrated that CEEMDAN-wavelet denoising algorithm can greatly reduce the residual noise in IMF components, while mainly retaining the characteristics of the vibration acceleration data. To achieve such a goal, we employ the CEEMDAN decomposition algorithm by adding adaptive white noise in the decomposition process to improve the decomposition efficiency. We also combine the CEEMDAN with a wavelet threshold denoising algorithm. The result shows that the processing can effectively reduce the noise signal of the original data and ensure that the signal is not distorted as much as possible.

In this paper, the data we only collected are the vibration acceleration, (Z-axis, Figure 2) which is perpendicular to the ground when the tractor is running. According to the existing theoretical analysis, such a signal may not be possible to analyze the vertical, longitudinal or transverse motion through the IMF components. Nevertheless, we indeed  value your suggestions, and will it in our following studies.

Point 4:

Line 403

You should add unit and name for raw data.

Is raw data a axial load or acceleration?

Please be specific.

The Authors’ Response: Thanks for your comments. The original data is the vibration acceleration (m/s²). The description and explanation can be found in both context (Line 153) and Figure 2.

In line 403, the data in Figure 11 is the vibration acceleration (m/s²). Because of the space limitation of the figure, we wrote the name and unit of the coordinate axis in the caption below the figure.

 

 Please see the attachment ‘Word file’.

Author Response File: Author Response.docx

Reviewer 2 Report

This article proposes a CEEMDAN-POT (Peak Over Threshold) model to comprehensively build a full life-cycle ground load spectrum of the tractor vibration with six ground conditions and different field operations. In addition, this paper realizes the effective denoising of ground load signal, load preprocessing, load spectrum compilation, and extrapolation. The description in Method regarding the load spectrum compilation is comprehensive. The manuscript employs a comparative novel method for denoising signals and compiling load spectrum, leading to interesting results. The Results and Discussion part also provides an effective comparative analysis. Overall, the paper is well-written 

1. Generally, the thesis is written clearly, with a complete theoretical framework and clear methods.

1.1 However, the authors should correct some minor editing errors: some subheadings are bold, such as 2.1 Preparatory tasks, while some subheadings (i.e., 3.2) are italics. All the subheadings should be unified and revised into the same format.

1.2. There are some minor typo errors, such as 4.1.2. ‘According to the verification and analy-sis, the first seven IMF components subject to wav elet denoising’. And some sentences in the text are repeated, which harms the conciseness of the paper.

1.3 Some symbols in content have inappropriate size, i.e. F(x) in the line 294, and these should be refined and elaborated.

2. What is the novelty of employing the CEEMDAN wavelet denoising algorithm? Please specify. The introduction should contain some necessary investigations and citations on the current and state-of-art signal processing for the ground load data.

Please explain the applicability of the CEEMDAN wavelet denoising algorithm to the preparation and extrapolation of the load spectrum of the POT model. What aspects of the load spectrum have been improved or enhanced?

3. Some references have minor errors, i.e., they miss page numbers, or volumes or are in an incorrect format, such as Refs. 9, 10, 18, 31, 35. Please check the reference and correct any errors. There is no need to add ‘.’ After the ‘[CrossRef].’. The authors should correct these.

4. Figure 17 presents a similar histogram between the load data and the 1-time extrapolation w.r.t mean and amplitude. What does that mean? Please explain in the content. In addition, the two subfigures in Fig. 17 have different y ticks. The legend format is also different from other figures. Please revise and ensure that the legends are unified.

5. Some symbols in the content are over-defined, i.e., “SNR” and “RMSE” are defined in both lines 260 and 423. Please correct.

Author Response

Compilation and Extrapolation of Load Spectrum of Tractor Ground Vibration Load Based on CEEMDAN-POT Model

agriculture-2097165

Response to Reviewer #2 Comments

Thank you very much for your kind comments on our manuscript, which is very helpful for revising and improving our manuscript. In the following reply, our answers are given point by point. At the same time, we have uploaded a revised version while highlighting all the changes by using the Track Changes mode in MS Word.

Point 1:

1.1 The authors should correct some minor editing errors: some subheadings are bold, such as 2.1 Preparatory tasks, while some subheadings (i.e., 3.2) are italics. All the subheadings should be unified and revised into the same format.

The Authors’ Response: Thank you for your suggestion. We have checked all the headings and subheadings in the article and revised them according to the template.Comment 1.2:

1.2. There are some minor typo errors, such as 4.1.2. ‘According to the verification and analy-sis, the first seven IMF components subject to wav elet denoising’. And some sentences in the text are repeated, which harms the conciseness of the paper.

The Authors’ Response: Thanks for your comment. We have found errors or questionable sentences in the text, and revised it accordingly .Comment 1.3:

1.3. Some symbols in content have inappropriate size, i.e. F(x) in the line 294, and these should be refined and elaborated.

The Authors’ Response: Thank you for your question. We have modified the size of  F(x) in line 294 and some other symbol words.

Point 2:

2. What is the novelty of employing the CEEMDAN wavelet denoising algorithm? Please specify. The introduction should contain some necessary investigations and citations on the current and state-of-art signal processing for the ground load data.

Please explain the applicability of the CEEMDAN wavelet denoising algorithm to the preparation and extrapolation of the load spectrum of the POT model. What aspects of the load spectrum have been improved or enhanced?

The Authors’ Response: Thank you for your questions and comments. In this paper, we collect the ground vibration load signal of the tractor when it is running on the ground, and analyze its vibration acceleration data. Affected by the working environment, the signal will have noise interference. CEEMDAN-wavelet denoising algorithm can effectively reduce the redundancy and residual noise in IMF components, while retaining the characteristics of the original signal. This algorithm can well process the acceleration signal, attempt to eliminate the signal distortion, and reduce the interference of ground and other factors on the acceleration data. In the ‘3.1. Data preprocessing method’ of this paper, the relevant papers of this study have been referred to, and the research progress has been described and investigated.

As presented by the paper, the introduced CEEMDAN wavelet denoising algorithm denoises the acceleration signal of the ground load, which can effectively improve the accuracy and authenticity of the load spectrum data, and make the load spectrum closer to the real field ground load data.

Point 3:

3. Some references have minor errors, i.e., they miss page numbers, or volumes or are in an incorrect format, such as Refs. 9, 10, 18, 31, 35. Please check the reference and correct any errors. There is no need to add ‘.’ After the ‘[CrossRef].’. The authors should correct these.

The Authors’ Response: Thank you so much for your careful check. We checked the list of references and corrected the errors found, some of which were in the revised manuscript. The revised contents are as follows.

Original text:	Corrected text:
16. Yan, J. G., Wang, C, G., Xie, S. S., Wang, L. J. Design and validation of a surface profiling apparatus for agricultural terrain roughness measurements. R and D National Institute for Agricultural and Food Industry Machinery - INMA Bucharest(3) 2019.	16. Yan, J. G., Wang, C, G., Xie, S. S., Wang, L. J. Design and validation of a surface profiling apparatus for agricultural terrain roughness measurements. R and D National Institute for Agricultural and Food Industry Machinery - INMA Bucharest 2019, 59, 169-180.
35. Fei, H.; Shan, J. Application of CEEMDAN-Wavelet Threshold Method in the Signal Processing of Blasting Vibration. Blasting. 2022, 39, 8.	35. Fei, H.; Shan, J. Application of CEEMDAN-Wavelet Threshold Method in the Signal Processing of Blasting Vibration. Blasting. 2022, 3, 41-47,164.

Point 4:

4. Figure 17 presents a similar histogram between the load data and the 1-time extrapolation w.r.t mean and amplitude. What does that mean? Please explain in the content. In addition, the two subfigures in Fig. 17 have different y ticks. The legend format is also different from other figures. Please revise and ensure that the legends are unified.

The Authors’ Response: Thank you for your question. Figure 17 has been explained in this paper. In order to better analyze the Mean and Amplitude counts information before and after extrapolation, the Mean and Amplitude counts diagram of 1 time extrapolated data and original data is compared and analyzed.

It can be seen from the figure that the counts of the extrapolated load spectrum is basically consistent with the Mean and Amplitude of the original data, and the fluctuation is small, which can verify that the load spectrum compilation method is feasible.

Original text:

Corrected text:

Point 5:

5. Some symbols in the content are over-defined, i.e., “SNR” and “RMSE” are defined in both lines 260 and 423. Please correct.

The Authors’ Response: Thank you for your suggestion. We have deleted the duplicated part (423 lines).

 

Please see the attachment 'Word file'.

 

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Thank you for your modification based on my comments.

The paper is well-written and should be published in this journal.

Author Response

Thank you again for your valuable comments on our manuscript.