Long-Lifetime Event-Driven Wireless Monitoring System for Pole-Mounted Transformers

Round 1

Reviewer 1 Report

In this paper, the authors propose an event-driven wireless monitoring system for 2 pole-mounted transformers. The proposed system substantially reduces operation and maintenance costs and enables low-cost diagnosis of components in pole-mounted transformers.

First, the conventional existing monitoring systems are presented and then the Event-driven Monitoring System is depicted using the block and timing diagrams.

I also remark that they tested the monitoring system using also the comparison between the simulations and experimental results.

Some things need to be clarified:

• Line 49 – a RMS not an RMS
• Please explain what does represent the time t_L and t_U – line 115
• You mentioned that this condition is satisfied P(RMS) + P(ADC(f_L)) + P(TX(f_L)) < P(ADC(f_H)) + P(TX(f_H)), please explain how does it result from eq. 3 and 5
• There are too much abbreviations in the text, there are difficult to follow, maybe you can add a table with the abbreviations at the beginning of the paper, before Introduction

Author Response

1. Thank you very much for your valuable comments. Authors would like to keep the original text as “an RMS”.
2. Explanation for t_L and t_U are in the text line 112. These variables representing the time that an abnormal signal exists. It is assumed that the abnormal signal starts to exist from t_L and ends at t_U. This is defined by the event controller in the proposed system.
3. Dynmaic power is a linear function of the sampling frequency. Also, it is represented in experimental results that P(ADC(f_L)) + P(TX(f_L)) is dominant over P(RMS). Therefore, P(RMS) + P(ADC(f_L)) + P(TX(f_L)) < P(ADC(f_H)) + P(TX(f_H)) can be guaranteed if f_L<<f_H.
4. Thanks for your comments. All the abbreviation in the paper are widely used for electronics, authors consider that adding a table can be redundant.

Reviewer 2 Report

- Keywords should be revised, since they are too general in the current form (eg. abnormality, event, monitoring).
- References cannot be cited in block (eg. [8]-[13]). For each of them, describe which aspects are significant for your paper.
- The difference between "sensor" in Figure 2.a and "integrated sensor" in Figure 2.c is unclear and should be described.
- From eq. (3), it is unclear the relationship between the energy required for monitoring the signal and the sample frequency.
- The meaning of alpha coefficient in eq. (9) and (10) should be explained.
- In Figure 15 and Figure 17, it is better to use colors to improve readibility.
- I suggest to support conclusions with some quantitative data.

Author Response

1. The keywords were revised as followed:

Original keywords: Abnormality; battery powered; pole-mounted transformer; event; monitoring; root-mean-square (RMS);

Revised keywords: Abnormality monitoring; battery powered; pole-mounted transformer; event-driven; root-mean-square (RMS);

2. Thank you very much for your valuable comments. The content is revised as commented.
3. An ‘integrated sensor’ is revised as ‘sensor’ in Fig. 2(c).
4. Dynmaic power is a linear function of the sampling frequency. Eq. (3) is an expression showing that the total power P(BPF)+P(ADC(f_H))+P(TX(f_H)) is multiplied by the total time (t_U-t_L) based on E=Pt.
5. α is the reciprocal of noise power P(RMS(v_NF(t_Li+t_Di))), which is described in line 171.
6. Thanks for your comments. Authors include colored figures only if emphasis is required. Authors consider that a legend is good enough for Figs.15 and 17.
7. Thanks for your comments. The conclusion was supplemented by adding quantitative data.

Reviewer 3 Report

The paper addresses many important issues related to event-driven battery-powered wireless monitoring system that monitors abnormalities of a transformer and transmits data only if an abnormality occurs.

The research has an innovative character in which are used an event controller monitors the abnormality of the input and allows for event-driven sampling when an abnormality is present in a transformer. It enables or disables each block and determines when to sample and transmit data. Without an abnormal signal, most of the components fall into sleep mode waiting for the next abnormality and the battery lifetime of the proposed monitoring system can be two orders of magnitude higher than that of a conventional monitoring system without an event controller.

The following remarks and recommendations can be made to the manuscript of the article:

To make a thorough analysis of the how to determine the time td (fig. 5) and in fact how much can be its minimum value?

 To give more detailed information about the experiments and in particular the parameters of the real transformer (if any).

Тhe operation of high-voltage transformers is associated with the presence of intense electric and electromagnetic fields. To what extent does this affect the operation of the sensors and the developed system?

Тhe presence of miniature solar and wind systems, as renewable energy sources, make a logical question - is it so important to make a system that could increase the lifetime of the battery?

In the conclusion it is concluded that a low-cost diagnosis system of the pole-mounted transformers components has been developed. Тhis statement would have been more emphatic if specific data had been provided.

Тhe system will operate at a fairly large temperature range - what measures are provided to maintain the necessary accuracy and speed of the developed monitoring system.

Author Response

1. In the conventional method, the maximum value of the RMS input is not determined at the precise time delay TD. Thus having an arbitrary value of TD between t_L and t_U may result in poor SNR depending on the input. This problem can be overcome by using the proposed method that finds the maximum value of the RMS input.
2. Acquiring abnormal inputs from the real transformer under normal operation can be very time consuming since it happens very rarely. Thus, we intentionally generated partial discharges by using a zig that mimics a pole mounted transformer. Authors hope that reviewers understand this limitation of experiments.
3. This paper is not considering problems by electromagnetic fields. However, proper shielding can reduce the effect of high electric field around the proposed monitoring system.
4. Even when renewable energy sources including energy harvesters are used, the proposed system is very advantageous. As an example, reducing the energy consumption by using the proposed method for monitoring a system requires smaller solar cells and smaller form factor. Moreover, battery power monitoring system can be widely used indoor or underground facilities.
5. Thanks for your comments. The conclusion was supplemented by adding quantitative data.
6. All components used in the paper work meet industrial standard. As an example, most of devices works at the temperature range between -40 degrees and 85 degrees. Components in the proposed work are available on the market and components name is already included in the main text in order to make reader access the datasheet easily. Accuracy and speed can be found from the data sheet of each component.

Round 2

Reviewer 2 Report

The Introduction section of the document must still be revised to avoid citing many references together (eg. [8-13]). For each reference, authors should clearly point out why it is mentioned in the text and how it is related to the their work. 

Author Response

Authors appreciate the reviewer's comments. Introduction was revised as followed:

Original contents: A wired sensor receives the physical quantities of a device in the form of raw    data at high-speed data acquisition (DAQ) for real-time continuous monitoring [8-13].

Revised contents: Wired sensors are widely adopted for high-speed continuous data acquisition (DAQ) of raw data from physical quantities of transformers such as bushing failures [8-9], winding deformation [10-11], partial discharge [12], and valve malfunctions [13].

Round 3

Reviewer 2 Report

Accept in present form.