Measurement and Analysis of Inadequate Friction Mechanisms in Liquid-Buffered Mechanical Seals Utilizing Acoustic Emission Technique

Round 1

Reviewer 1 Report

This paper has presented detection and analysis of friction mechanisms inside a mechanical seal for fault diagnosis using an acoustic emission technique. Several f conditions were investigated using a test rig, including normal conditions, drying-running operation, abrasive particles and cooling failure. The AE results seem to align with the temperature progression indications over time of the mechanical seal’s housing. Some improvements could be made to the paper as follow. The introduction is rather long; perhaps the mechanical seals section and the AE section could be separated to make the structure clearer. The main point of this paper is to introduce AE technique for mechanical seal fault diagnosis, so the AE technique should be compared with other standard techniques in addition to vibrations, e.g. acoustic detection, see Acoustic-based engine fault diagnosis using WPT, PCA and Bayesian optimization, Applied Sciences, and thermal image analysis, see Intelligent fault diagnosis of rotor-bearing system under varying working conditions with modified transfer CNN and thermal images, IEEE TII. The experiments and experimental data were presented and explained comprehensively. Although vibration analysis was mentioned as the traditional monitoring technique, it was not compared with the AE analysis for every case study. In terms of signal analysis, only the amplitude of time signals and frequency spectrum were presented and analysed, which are rather standard. As AE technique provides high-frequency signals, which provide a large amount of data, deep learning techniques could be considered as future work, see Deep recurrent entropy adaptive model for system reliability monitoring, IEEE TII. In summary, the introduction could be re-structured. AE technique as the contribution should be made clearer; AE technique should be discussed with other monitoring techniques, including vibration, acoustic and thermal images, etc. Furthermore, potential further work using these high-frequency AE signals should be discussed towards the end.

Author Response

Thank you for your comments. Please find our replies in the attached file.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper presents an experimental research for detection and analysis of friction mechanisms in a mechanical seal using the acoustic emission technique. The results show that this technique allows to detect fault condition modes.

The paper is well written and structured. The research present a very interesting line for condition monitoring purposes.

Comments for improving the paper:

Line 46, authors claim that the typical measurement range goes up to frequencies of 1000 Hz.  It seems a very low used frequency as there are articles such as "Onboard Condition Monitoring Sensors, Systems and Techniques for Freight Railway Vehicles: A Review" that recommend using a sampling frequency of 50kHz for bearing analysis and in "Rolling element bearing diagnostics using the Case Western Reserve University data: A benchmark study "indicate finding the greatest presence of frequency components in the 11 to 14kHz band. Authors should add references to this claim.

Figure 2, the equation is not well seen.

I found interesting for reader to discuss about the economic cost for measuring at this very high sample frequency.

I encourage authors to include in Conclusions section next expectations about their research: new application fields, next steps…Also, the limitations or future works to improve the research.

Minor English editing changes:

Instead “Provoke”, I recommend to the authors use “cause”/”induce”

Author Response

Thank you for your comments. Please find our replies in the attached file.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The original comments were: The introduction could be re-structured. AE technique as the contribution should be made clearer; AE technique should be discussed with other monitoring techniques, including vibration, acoustic and thermal images, etc. Furthermore, potential further work using these high-frequency AE signals should be discussed towards the end.

It seems that only part of the comments were addressed. Besides, some of replies were doubtful, e.g. 'We know that Machine Learning techniques very often require a lot of marked data, which is something not yet available to us in the presented project.' -- there are many unsupervised machine learning techniques which do not require labelled data, such as clustering techniques, autoencoders, etc.

Author Response

Thank you for your comments and suggestions. Please see the attachment.

Author Response File: Author Response.docx

Round 3

Reviewer 1 Report

Optional: Some response to comments could be included in the revised paper (Introduction or Discussions).

Author Response

Thank you very much for your suggestion. Please see the attachment.

Author Response File: Author Response.docx