Design study of on-line monitoring system for power equipment operation status

In order to ensure the normal power supply work, the state of electric power equipment must be scientifically monitored, but also do a good job of diagnosis, so as to fundamentally eliminate the emergence of higher equipment operation and maintenance costs, and better enhance the safety of the power system [1-3]. However, the current technical impact, coupled with cost constraints, resulting in power equipment operation online monitoring is difficult to achieve high efficiency, so the need to strengthen the design of power equipment operation status online monitoring system [4-5].

From the power equipment operation status online monitoring system itself, it is the information that has been sterilized deterioration, there may be defects in the electrical, physical, chemical and other characteristics of the change in the collection, and then through the overall analysis and processing, real-time grasp of the specific change in the data, which can effectively assess the use of electric power equipment time and reliability [6-8]. If the data, trends have problems, can determine the power equipment operation is abnormal, more able to determine the factors of failure, failure severity, specific parts, etc. [9]. Therefore, during the design of power equipment online monitoring system, should accurately grasp the power equipment operating conditions, especially some of the inconspicuous abnormal information before the failure, so as to warn in advance, and then make a correct assessment [10-12].

Characteristics of condition monitoring and fault diagnosis system. In the process of power system operation, the overall equipment status of the system should be monitored at all times, but also the scientific use of appropriate fault diagnosis techniques to ensure that the power system can work properly [13-14]. However, the actual operation of the system is often due to some internal and external reasons, resulting in various types of faults. In addition, the power system for a long time in the working state, some construction materials will be in the electric heat climate and other factors under the influence of the aging situation, if the power system failure problems, the consequences are extremely serious, will cause widespread blackouts occur, which not only affects people’s daily life, but also will cause great economic losses [15-18]. In order to prevent the formation and development of faults, electric power enterprises need to do a good job of monitoring the status of the power equipment system, and more effectively determine the faults that may occur [19].

Power equipment operation status online monitoring system has a scientific monitoring role, can be continuous and real-time monitoring of the operation status of power equipment, in order to achieve better monitoring quality and effect, need to strengthen the design of the power equipment operation status monitoring system, for the diversified monitoring needs, and constantly improve the monitoring technology of the system [20-22]. However, the improvement of technology is affected by more factors, resulting in more system problems, which not only fails to meet the monitoring standards, but also causes power equipment operation faults, so it is more important to pay attention to the achievability of the design of the online monitoring system for the operation status of power equipment [23-24].

This paper designs and optimizes an online monitoring system for power equipment operation status to improve its data collection and analysis capability. The decision tree classification algorithm is utilized to diagnose fault conditions in electric power equipment to improve the level of equipment fault prediction and diagnosis accuracy. By conducting control experiments, the advantages of the online monitoring system designed in this paper are compared to the fractional order system and the improved association rule system. Targeted power equipment operation monitoring system performance testing and analysis is carried out with transformer power equipment as a case study, which proves that the monitoring accuracy of the power equipment operation monitoring system designed in this paper is high, and it can realize the effective monitoring of the state of power equipment. Adopt the method of real-time monitoring of energy-consuming data to verify that the monitoring system designed in this paper has the advantage of low energy consumption. Through the experiment of equipment fault diagnosis algorithms, it is clear that the diagnosis accuracy of the decision tree classification algorithm is high.

2

Overall design and optimization of the monitoring system

This part of the overall design of the power equipment operation status online system, analyze the effectiveness of the data collection method, improve the decision tree classification algorithm for fault diagnosis needs, and improve the overall monitoring effect.

2.1

Overall design program for the monitoring system

1)

The overall logical design of the system adopts the method of gradual refinement in graded chunks.

2)

In terms of hardware design, considering that the signal interface circuit of each monitoring parameter, the communication module interface circuit and the keyboard interface circuit are all mature technologies before this, they can be directly transplanted and utilized in order to help shorten the development cycle. We mainly concentrate on designing a strong compatibility with the sensing detection data acquisition board and preparing the corresponding driver.

3)

In software design, because this system is based on database design, but there are still certain requirements for real-time, can be used in assembly language, VB mixed programming programming methods.

4)

As this system is a performance testing equipment, the design should focus on the safety and practicality of the system, reliability, fault tolerance of the system.

Technical Route:

After a comprehensive analysis, we propose a remote data transmission scheme using GPRS information technology, which can be effective for remote transmission. This wireless monitoring system for the working status of electric power equipment is mainly composed of the following five functional modules, and each component is briefly introduced below: 1)

Data Acquisition Terminal

Through the microcontroller control system with built-in TCP/IP protocol, the information collection terminal can complete the centralized collection and processing of signals from various sensors, and complete the real-time information exchange with the data monitoring management center in the integrated control room through GPRS network, so as to achieve real-time monitoring of the environment.

2)

GPRS wireless Modem

Using GPRS wireless modem, which has the function of processing TCP/IP protocols and can support both short message and GPRS wireless data transmission, is the core communication module of the system.

3)

Data processing host (upper computer) in the monitoring center

A computer connected to the network, able to use Socket interface and information collection terminal to complete the GPRS transmission, the monitoring center should be equipped with short message module, in order to complete the information collection terminal and information collection between each other’s fast and stable transmission.

4)

GPRS wireless mobile network channel

GPRS is an important communication link that can connect the control center with the data information collection terminal, making remote communication more convenient and efficient. Through dial-up connection, the monitoring center can easily access and collect data.

5)

SIM card

In GPRS wireless Modem, SIM card is an important user identification card, which can store all kinds of data, authentication methods and keys of users, so that GSM operating system can accurately identify the user’s identity, and can provide called data service service to realize the important task of transmission. Users can connect with the monitoring system through SIM cards and can realize the exchange of information.

This management system uses GPRS network system, which can complete the remote transmission of historical data information, instant data information and alarm messages, and at the same time, the SMS service can also transmit parameter changes and alarm messages, so that the monitoring and management center can obtain them in time. 1)

Server Side

When the host side starts up, it will start listening at specific devices, and once it finds a client accessing, it will immediately create a network connection and accept the information uploaded by the user’s data terminal.GPRS network uses TCP/IP technology to transmit the data, so the server side can use Socket to complete the communication easily.

2)

Client Side

Once the client is powered on, it will establish a connection with the server according to the pre-specified IP address and port number, and is equipped with a wireless Modem with GPRS function in order to realize two-way data transmission, either in server mode or client mode, and can realize the function of sending and receiving short messages.

This control system contains two levels: information acquisition unit and control unit. The first is formed by the user equipment, while the latter is formed by the controller and the GPRS wireless modem. The unit that receives and processes data at the highest level is connected to the Internet and has a GSM short message transceiver. The core part of the whole control system is formed by GPRS communication module, TCP/IP protocol module, short message processing module and data acquisition module, in which the GPRS communication module can be accomplished by a single wireless Modem, while the TCP/IP protocol module and the short message processing module can be accomplished by a single wireless Modem as well. Figure 1 shows the overall architecture of the control system.

Using GPRS wireless Modem to connect to the GPRS network system, the monitoring center gives information processing commands, and the information acquisition terminal receives the commands, processes the digital information output from the sensor and encapsulates the protocols, and then sends it instantly to the GSM/GPRS network system, and ultimately transmits it to the Internet through the network of China Mobile Communications Ltd. To perform remote monitoring and data analysis. Using ADSL, dial-up, or fixed network, the network system can transmit information to the data control center, and can be realized using a short message method.

2.2

Data acquisition

Well-functioning BS architecture. The BS architecture-based data collection design method ensures the independence of the data and scalable characteristics. However, the degree of realization of this feature is more complex and difficult.

This paper based on the B / S architecture of the data collection design method to ensure that the data between the independent and scalable characteristics, the degree of realization of this feature is more complex and more difficult. The diversity and dynamics of network transmission based on B/S architecture require that the data collected by the data collection tool has strong reliability and real-time, so the quality of data collection should be ensured as much as possible in the design process of the system.

Therefore, when researching and designing data acquisition tools for power equipment operation data analysis systems, the following design requirements should be met: 1)

Larger scale data acquisition black box collection of data complexity, and data real-time and transmission accuracy requirements are high, in addition to the massive transmission of data on the server throughput and stability characteristics are also the requirements.

2)

The black box is basically installed in the outdoor electric power environment, and the electric power data acquisition equipment operating in this environment is easy to be damaged, and then corresponding measures should be taken in the design to ensure the reliability of data communication between the host computer and the black box, and ultimately to realize the dynamic real-time tracking of electric power field data.

In order to effectively solve the above problems, this paper proposes a multi-threaded power data acquisition and processing technology based on multi-threaded software support to realize efficient real-time data acquisition from the dynamically changing network address, a larger number of acquisition devices.

2.2.1

Data cleansing

As the current data of electric power equipment is becoming more and more huge, there are incomplete, complex and redundant data in the massive data, a large amount of data redundancy in everything very much affects the speed of the system to analyze the data, and a large amount of data may lead to errors in the analysis results of the computer, so the system will then carry out data analysis before the pre-processing of some of the original data, through the selection and deletion, leaving the Useful data, which can improve the overall quality of the data, and the entire subsequent power equipment operation data analysis and calculation of the efficiency will be improved, including the accuracy and performance of data analysis will be improved.

2.2.2

Normalization of data

The data normalization process is divided into three kinds of standard transformation, polar deviation normalization transformation and square root standard method, which are implemented as described below: 1)

Standard transformation

The vector of dimension p is represented by $X = (X_{l}, X_{2}, \dots, X_{p})$ and the observation matrix is shown in equation (1): (1) $X = [\begin{matrix} x_{11} & x_{12} & \dots & x_{1 p} \\ x_{21} & x_{22} & \dots & x_{2 p} \\ \dots & \dots & \dots & \dots \\ x_{n 1} & x_{n 2} & \dots & x_{n p} \end{matrix}]$

The normalized transformation is expressed as: (2) $X^{*} = [\begin{matrix} x_{11}^{*} & x_{12}^{*} & \dots & x_{1 p}^{*} \\ x_{21}^{*} & x_{22}^{*} & \dots & x_{2 p}^{*} \\ \dots & \dots & \dots & \dots \\ x_{n 1}^{*} & x_{n 2}^{*} & \dots & x_{n p}^{*} \end{matrix}]$ (3) $x_{i j}^{*} = \frac{x_{i j} - \bar{x_{i}}}{\sqrt{s_{i j}}}, i = 1, 2, \dots, n; j = 1, 2, \dots, p$ (4) $\sqrt{s_{i j}} = \sqrt{\frac{1}{p - 1} \sum_{j - 1}^{p} {(x_{i j} - \bar{x_{i}})}^{2}}$

where ${\bar{x}}_{i}$ represents the mean value obtained from the observations of variable X_i. Variable X_i has a variance of s_ij and a standard deviation of $\sqrt{s_{i j}}$ . After the normalization transformation, the mean of each row of the original matrix is 0 and the standard deviation is 1. 2)

Polar deviation normalization transformation

The observed value matrix X expressed in equation (1) is transformed by the polar-variance normalization transformation operation, which can be obtained: (5) $X^{*} = [\begin{matrix} x_{11}^{R} & x_{12}^{R} & \dots & x_{1 p}^{R} \\ x_{21}^{R} & x_{22}^{R} & \dots & x_{2 p}^{R} \\ \dots & \dots & \dots & \dots \\ x_{n 1}^{R} & x_{n 2}^{R} & \dots & x_{n p}^{R} \end{matrix}]$ (6) $x_{i j}^{R} = \frac{x_{i j} - \min \underset{1 \leq k \leq p}{x_{r k}}}{\max \underset{1 \leq k \leq p}{x_{r k}} - \min \underset{1 \leq k \leq p}{x_{r k}}}, i = 1, 2, \dots, n; j = 1, 2, \dots, p$

where $\min \underset{1 \leq k \leq p}{x_{r k}}$ denotes the minimum value in the observation of variable x_i and $\max \underset{1 \leq k \leq p}{x_{r k}} - \min \underset{1 \leq k \leq p}{x_{r k}}$ denotes the extreme deviation value of observation x_i. The values of all the elements of the matrix range between 0 and 1 after the polar-variance normalization transformation operation. 3)

Square root standard method

The square root standard method is applied to the observation matrix X in Equation (1). (7) $x^{*} = [\begin{matrix} x_{11}^{s} & x_{12}^{s} & \dots & x_{1 p}^{s} \\ x_{21}^{s} & x_{22}^{s} & \dots & x_{2 p}^{s} \\ \dots & \dots & \dots & \dots \\ x_{n 1}^{s} & x_{n 2}^{s} & \dots & x_{n p}^{s} \end{matrix}]$

where $x_{i j}^{2} = \frac{x_{i j}}{\sqrt{x_{11}^{2} + x_{12}^{2} + \dots + x_{i p}^{2}}}$ , i = 1, 2, …, n; j = 1, 2, …, p. After the above operations, the values of the elements of matrix X^* are fixed between 0 and 1.

2.3

Improvement of Decision Tree Classification Algorithm in Fault Diagnosis

There are many ways to categorize data nowadays, and one of the most effective methods is the decision tree. The primary objective of categorizing data is to discover a more precise model for each type of data entered later. Usually, a specific category is used to represent an accurate model.

After analyzing the data classification, the process can be divided into two steps:

Step1. The first step is known as the model training phase, in which the main task is to find out the mapping function representation of the model that is suitable for the given training set.

Step2. The second step is known as the classification rule formation phase, in which the function model identified in the first step is used to predict the categories of the data and finally form a classification rule.

Decision tree as a classifier, we can view it as a directed, acyclic tree such that each element in the decision tree corresponds to a node, all trees without leaf nodes are called pure trees, and each non-terminal node of the decision tree represents a test or a decision for that examined data. We then choose a branch based on the results of this test, and in order to perform a classification operation on a particular piece of data, we must first begin our analysis at the root node and keep making judgments all the way down until we reach a leaf node before we can end it. If this analysis reaches a leaf node, it means that a decision has been made.

ID3 algorithm is the most influential decision tree algorithm in the world, and CLS algorithm is the predecessor of ID3 algorithm.The workflow of CLS is as follows: firstly, select the most influential factor among all the factors, and then divide the data set into several subsets according to the selected factors, and then judge each subset and select the factor with the most influence respectively; and then continue to divide according to the selected factors, and so on, until the data set is divided into several subsets. Then continue the next division according to the selected factors, and so on, until the obtained subset contains only a certain type of data.

In ID3 algorithm, the concept of entropy in information theory is introduced into it, and the entropy before and after segmentation is used to calculate the information gain, which is used as a metric to measure the feature discrimination ability.

Principle of ID3 Algorithm : Let there be a set S containing s sample of data which can take m different values for its class labeling attribute corresponding to m different categories C_i, $i \in {1, 2, \dots \dots, m}$ . Let the number of samples contained in category C_i be S_i, then the desired information required to categorize a given sample is given by Equation (8): (8) $I (s_{1} \cdot s_{2}, \dots \dots \dots s_{m}) = - \sum_{i = 1}^{m} p_{i} \log_{2} (p_{i})$

Let attribute A have v different values ${a_{1}, a_{2}, \dots \dots, a_{v}}$ and divide the set S by attribute A, which can be divided into v different subsets ${S_{1}, S_{2}, \dots \dots, S_{v}}$ , where all the data samples in the S set where attribute A takes the value a₁ are included in S. If the selected test attribute is attribute A, let S_ij be the number of samples in subset S, in category C, then the entropy of the current subset divided by A is calculated by equation (9): (9) $E (A) = \sum_{j = 1}^{v} \frac{s_{1 j} + S_{2 j} + \dots \dots + S_{m j}}{s} I (s_{1 j}, \dots \dots ., S_{m j})$

In the above equation $\frac{s_{1 j} + s_{2 j} + \dots .. + s_{m i}}{s}$ is the ratio of the sum of the number of samples in all subsets where attribute A takes the value of a_j to the total number of samples in the set of S. The smaller the value of entropy computed, the higher the purity of the subset delineation, i.e., in the direction of what is possible, and the information for a given subset S: (10) $I (s_{1}, \dots \dots s_{n j}) = - \sum_{k = 1}^{n} p_{q} \log (p_{v})$

The corresponding information gain values can be obtained from the above calculated entropy value and expectation information respectively, and the attribute A is utilized to divide each current branch node. The obtained information gain is calculated by equation (11): (11) $G a i n (A) = I (S_{1}, S_{2}, \dots \dots, S_{m}) - E (A)$

From the above equation, we can see that the smaller the value of entropy $E (A)$ , the larger the value of information gain $G a i n (A)$ obtained by branching on A, which indicates that the selection of test attribute A provides more information for classification, i.e., the more beneficial it is for classifying the training data.Each selection of the ID3 algorithm involves selecting the attribute with the largest value of information gain and using it as the test attribute of the analyzing node to partition the training set again until the final decision tree is generated.

ID3 Algorithm Decision Making Detailed Steps:

Step1. Randomly draw a sub-training set from the training set so that it contains both category S and category H.

Step2. Invoke the decision tree generation algorithm to generate a decision tree for the current sub-training set.

Step3. Judge the rest of the subsets in the training set using the decision tree formed by Step2, and find the wrongly judged examples according to the actual situation.

Step4. If there are misjudged examples, insert them into Step1, and then go to Step2 and re-generate the decision tree, if there are no misjudged examples, it means that the generated decision tree is trustworthy, and the whole computation process ends.

3

Monitoring system application studies

In order to verify whether the on-line monitoring system for the operation status of power equipment designed in this paper is really effective, this part adopts a controlled experiment to compare the system with two other monitoring systems and to judge the specifics of the monitoring system designed in this paper in terms of monitoring accuracy and operation energy consumption.

3.1

Equipment operation monitoring system testing and analysis

3.1.1

Test preparation

When analyzing the performance of the power equipment operation monitoring system designed in this paper, targeted test analysis was conducted on transformer power equipment. In order to be able to more objectively analyze the monitoring effect, respectively, the fractional order system, as well as to improve the association rule system as a test control group. The basic situation of the test transformer power equipment is analyzed, which is mainly used in the power transmission and distribution system to transmit electrical energy and transform voltage. The test transformer adopts advanced electromagnetic induction technology by changing the amplitude and frequency of AC voltage to meet the needs of different power equipment and users. Table 1 shows the specific performance parameter configuration information of the test power transformer.

Table 1.

Test power transformer performance parameter configuration

Parameter name	Value/Description	Parameter name	Value/Description
Rated capacity	15.0MVA	Cooling mode	Oil immersed self-cooling/forced air cooling
Rated voltage	220.0kV/35.0kV	Insulation class	Grade F
Rated current	68.19A/197.75A	Impedance voltage	7.5%
Rated frequency	60.0Hz	Temperature rise limit	Top oil temperature rise ≤55.5K, Winding temperature rose≤65.5K
Rated power	15000.0kW	Short-circuit impedance	High pressure - Medium pressure:Xk=15%, High-low pressure:Xk=36%
Efficiency	≥99%	Phase number	Three phase
Conditions of use	Outdoors	Wiring group	Yn,d11

The test power transformer is installed in a large substation. For the power system environment, the test power transformer has a high rated voltage and rated current because the substation is the hub of the power transmission and distribution system and needs to handle high voltage levels of power. Affected by fluctuations in power demand, the load of the power system varies relatively greatly. In order to ensure the stable operation of the power system, real-time monitoring of the power transformer is required to ensure the stable transmission and supply of electric energy.

3.1.2

Test results and analysis

The output currents of the test power transformer over a 24-h period were monitored using the three systems described above, respectively. The comparison graph of the monitoring results of different methods is shown in Fig. 2. Combined with the test results shown in Fig. 2, it can be seen that in the monitoring data of the three different systems, the corresponding parameters show more obvious differences from the actual values. Among them, the fractional order system, the specific parameter values did not deviate from the actual situation, but the larger output current fluctuation monitoring results have a more obvious time difference compared with the actual distribution; the improved association rule system, the relatively smooth output current monitoring results are more accurate, but the larger output current fluctuation monitoring results are significantly lower than the actual value; in contrast, the test results of the system designed in this paper In contrast, the test results of the system designed in this paper not only show good synchronization with the actual state of the test transformer in terms of timeliness, but also keep a high degree of fit with the actual value of the specific output current state parameter monitoring results, with good performance in terms of accuracy. Combined with the above test results and comparative analysis, it can be concluded that the power system equipment operation monitoring system designed in this paper can realize the effective monitoring of power equipment status.

The system monitoring data during the operation of the power equipment was further selected to analyze the accuracy of the monitoring data and the energy consumption during the monitoring process of the three different systems. Figure 3 shows the comparison of the monitoring accuracy of the three systems. Combining the extracted monitoring data during the operation of power equipment and Fig. 3, it can be seen that the monitoring accuracy of the system designed in this paper is significantly higher than that of the fractional order system and the improved association rule system. In the first few seconds of the preliminary period, although the system designed in this paper is lower than the accuracy of the other two, the accuracy of the system rises very quickly and finally reaches a stable value of 98.5% accuracy. While the accuracy of the fractional order system is also rising, the accuracy rises much less rapidly than that of the system designed in this paper. The accuracy of the improved association rule system is extremely unstable, sometimes rising and sometimes falling.

In the experiment to verify the operational energy consumption, the method of real-time monitoring of energy consumption data is adopted, and the three systems simultaneously monitor the data of the same power equipment for three hours, and finally calculate the energy consumed by each of the three systems. Fig. 4 shows the comparison of the energy consumption of the three systems in operation. As can be seen from Fig. 4, the fractional order system consumes about 36% of the operating energy, the improved association rule system consumes about 12% of the operating energy, and the monitoring system designed in this paper consumes about 6% of the operating energy, the lower the operating energy consumption, the higher the efficiency of the operation, and the higher the accuracy of the monitoring. Comparing the three systems, the system designed in this paper has high application value in terms of improving data monitoring accuracy and reducing operational energy consumption.

3.2

Experimentation and Analysis of Equipment Fault Diagnosis Algorithm

3.2.1

Experimental environment

In order to verify the accuracy of the decision tree classification algorithm selected for the monitoring system designed in this paper in equipment fault diagnosis, this part selects the relevant fault type data to test the algorithm, and at the same time, in order to improve the accuracy of the experimental results, the algorithm of fractional-order system and the algorithm of the improved association rule system are selected for the comparative test of the experiment to compare the accuracy of the algorithm.

The experimental object chosen is the power operation equipment that operates under the power grid system of a domestic power company. By analyzing the historical operation data of the power equipment of the power system of this company, we can obtain power operation data for various periods. Taking the substation equipment as an example, six different fault types were selected, and Table 2 shows the specific fault types and the corresponding feature vector descriptions, and different numbers of sample test data were selected for each fault according to its operational characteristics and machine learning difficulty.

Table 2.

Description of fault types and feature vectors

Fault number	Fault type	Eigenvector description	Sample size
01	Abnormal position of circuit breaker	Breaker load	150
02	The circuit breaker trips and breaks	Device trip signal	200
03	Single-phase grounding	Secondary loop current	250
04	Distance protection	Circuit breaker phase A current	200
05	The secondary control circuit is disconnected	Resistive leakage current	250
06	Oil level anomaly	Transformer oil position	300

3.2.2

Analysis of experimental results

The experimental comparison criterion adopted in this experiment is the diagnostic accuracy of the equipment fault diagnosis methods. Figure 5 shows the experimental results of the accuracy of the algorithms obtained after calculating the actual analysis results of the three algorithms. Through the numerical comparison can be seen, the diagnostic accuracy of the two conventional equipment fault diagnosis algorithms is low, the average value of less than 70%, and the accuracy rate change is not stable, with the change of the type of faults will appear larger changes in the situation. The decision tree classification algorithm chosen in this paper is more accurate than the two conventional fault diagnosis algorithms, with an average accuracy of around 90%. And the decision tree classification algorithm selected in this paper will not fluctuate with the change of fault type, the accuracy rate is more stable. It can be seen that providing equipment fault diagnosis data to the decision tree classification algorithm for training, can get a higher accuracy rate, therefore, the decision tree classification algorithm applied to the monitoring system designed in this paper, can effectively monitor the operating posture of electric power equipment, and discover the possibility of equipment failure in time.

4

Conclusion

This paper designs an online monitoring system for the operation status of power equipment, and verifies its advantages through controlled experiments. Based on the transformer power equipment to carry out targeted monitoring system testing and analysis. The monitoring system designed in this paper not only in terms of timeliness and the actual state of the test transformer to maintain a good synchronization, and the specific output current status parameters monitoring results with the actual value to maintain a high degree of fit, in terms of precision has a good performance. The monitoring data is stable at 98.5% accuracy. The system can effectively monitor the condition of power equipment. Through the method of real-time monitoring of energy consumption data, the total energy consumption of the monitoring system designed in this paper is only 6% for 3 hours, which is much smaller than that of the other two systems of 36% and 12%, and has the advantage of low energy consumption. Based on the ID3 algorithm, the average value of the fault diagnosis accuracy of electric equipment is about 90%, which is much higher than the algorithms chosen by the other two systems, and has excellent diagnostic capability.

Based on the research in this paper, it is verified that the monitoring system designed in this paper has the dual advantages of high monitoring accuracy and low operational energy consumption, and the system has a good development prospect in the application of monitoring the operational status of power equipment.

Idioma:: Inglés

Calendario de la edición:: 1 veces al año
Temas de la revista:: Ciencias de la vida, Ciencias de la vida, otros, Matemáticas, Matemáticas aplicadas, Matemáticas generales, Física, Física, otros

RSS Feed de revista

Design study of on-line monitoring system for power equipment operation status

Zichen Wu

Yicheng Wang

Bin Gu

Yunxiang Zhang

Hao Hu

Publicado en línea: 26 mar 2025

Recibido: 16 nov 2024

Aceptado: 28 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0802

Palabras clavePower equipment, Online monitoring system, Decision tree classification algorithm, Fault diagnosis, Power transformer

© 2025 Zichen Wu et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Palabras clave
Power equipment, Online monitoring system, Decision tree classification algorithm, Fault diagnosis, Power transformer