Next Article in Journal
A GAN-Based Video Intra Coding
Next Article in Special Issue
Adaptive Sliding Mode Neural Network Control and Flexible Vibration Suppression of a Flexible Spatial Parallel Robot
Previous Article in Journal
Modeling Power GaN-HEMTs Using Standard MOSFET Equations and Parameters in SPICE
Article
Peer-Review Record

A Segmented Preprocessing Method for the Vibration Signal of an On-Load Tap Changer

Electronics 2021, 10(2), 131; https://doi.org/10.3390/electronics10020131
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Electronics 2021, 10(2), 131; https://doi.org/10.3390/electronics10020131
Received: 8 December 2020 / Revised: 4 January 2021 / Accepted: 5 January 2021 / Published: 9 January 2021

Round 1

Reviewer 1 Report

A good paper. 

The proposed approach is easy to use.

Experimental set up has to be more clearly presented.

Author Response

Sincerely thank you for your valuable comments on the manuscript.

Q Experimental set up has to be more clearly presented

Answer: At the beginning of Section 4.1, the specific experimental device is introduced, and then the establishment of the sample is introduced in detail. Table 2 details the contents of the experiment and its implementation methods.

Reviewer 2 Report

This papers deals with an interesting pre-processing technique used to detect faults features robustly in the context of on-load tap changer monitoring. The context of the study and the state of the art are well introduced. The processing technique is unfortunately not quantitavely described, as parameters used to carry out processing steps like smoothing are not given. Nonetheless, the results obtained look convincing to me, so I recommend to publish the paper as it is.

Author Response

Sincerely thank you for your valuable comments on the manuscript.

Q The processing technique is unfortunately not quantitavely described, as parameters used to carry out processing steps like smoothing are not given.

Answer: In this step, we have tried FFT filtering, Kalman Filter, FIR filtering, Hilbert envelope and other technical solutions, but from the measured results of experimental data, this step actually does not use any smoothing technology to have little effect on subsequent processing (The purpose of smoothing is for the convenience of segmentation). Subsequent segmentation can be achieved ideally. Therefore, we describe here as " properly smoothed", that is, unsmoothing is also ok.

Reviewer 3 Report

This paper is well written. Readers can easily grasp the topic of the paper through the clear and logical expression of the abstract. Here are some areas for improvement:

The segmentation preprocessing method section can be expanded to fully illustrate this method.

Is the selection of training samples and continuous samples in 4.1 random or justified?

Why choose SVM for fault diagnosis?

Author Response

Sincerely thank you for your valuable comments on the manuscript.

Q1 The segmentation preprocessing method section can be expanded to fully illustrate this method.

Answer: This paper introduces the processing method in detail in section 3.2.

Q2 Is the selection of training samples and continuous samples in 4.1 random or justified?

Answer: The training sample uses the upshifts 1-2, 2-3, 3-4, 4-5, 5-6, 12-13, 13-14, 14-15, 15-16, 16-17 and downshifts 2-1, 3-2, 4-3, 5-4, 6-5, 13-12, 14-13, 15-14, 16-15, 17-16 in the continuous measurement sample.  The distribution and number of samples are shown in Table 1.

Q3 Why choose SVM for fault diagnosis?

Answer: SVM has the advantages of rigorous theory, strong adaptability, global optimization, high training efficiency and good generalization performance. It can handle small samples and nonlinear problems, so this paper chooses SVM for fault diagnosis.

Since the focus of the thesis is not here, there is no detailed discussion.

 

Back to TopTop