Review Reports

Round 1

Reviewer 1 Report

This paper investigated the vibration signal (mechanical wave during trip process), and used it to diagnoise circuit breaker fault types. Overall, the paper is well presented and written. Below are some comments to the author:

1. For the simualtion test in section 4.2, 80 samples were used for training, and the other 80 samples for test.  But according the paper, for each fault condition (4 types inculding normal) under each power supply voltage (4 voltages), the tests were repeated 10 times. That's how the 160 samples were obtained. So it is doutful if the training group and test group are higly correlated.  Then it doesn't make too much sense to test the model which is trained by higly correlated data. Has the author tried any measure to remove the correlation between repeated tests?

2. Page1 line 29, come --> came

3. Page1 line 30 occur --> occurred

4. Page2 line 96, what is CC short for? Same thing for the Page3 line 98.

5. Page3, line 99 should be "Synchronously received"

Author Response

Dear Reviewers,

We are deeply grateful for your thoughtful comments. We have taken your suggestions into careful consideration, and our detailed point-by-point responses to the comments are as follows.

Point 1:

For the simulation test in section 4.2, 80 samples were used for training, and the other 80 samples for test.  But according the paper, for each fault condition (4 types including normal) under each power supply voltage (4 voltages), the tests were repeated 10 times. That's how the 160 samples were obtained. So, it is doubtful if the training group and test group are highly correlated.  Then it doesn't make too much sense to test the model which is trained by highly correlated data. Has the author tried any measure to remove the correlation between repeated tests?

Reply:

Thank you very much for your suggestion. The correlation between repeated tests is necessary in this paper to reveal the effectiveness of the feature sets.

The trip mechanism of circuit breaker is a complex nonlinear system, and its vibration signal is also dispersive. Even if repeated tests are carried out, the characteristics of the vibration signal may change due to slight changes in parameters. The purpose of this paper is to propose a feature set that features are distributed in different regions under different faults, so as to realize fault diagnosis. In repeated tests, samples under different faults can be well separated by the features, as can be seen in Figure.8 in the paper. The results verify the effectiveness of the feature set.  In the actual application, the effective fault diagnosis model can be established by combining the vibration signal under normal conditions obtained from the factory test of the circuit breaker and the defect simulation test carried out for the same type of circuit breaker.

 

Point 2~5:

2. Page1 line 29, come --> came
3. Page1 line 30 occur --> occurred
4. Page2 line 96, what is CC short for? Same thing for the Page3 line 98.
5. Page3, line 99 should be "Synchronously received"

Reply:

Thank you very much for your suggestion. CC is short for coil current, and is added in Page2 line 78. These parts have been revised in the revised manuscript.

Author Response File: Author Response.pdf

Reviewer 2 Report

Notes in the attachment.

Comments for author File: Comments.pdf

Author Response

Thank you very much for your suggestion. The related part has been revised in the revised manuscript.