A method and system for detecting faults of a DC bias-resistant transformer with a wide metering range

Abstract

The present invention discloses a method and system for detecting faults of a wide-metering range DC bias-resistant transformer, which relates to the technical field of power transformer fault detection. The method collects mechanical data, current data and judgment standard data of each wide-metering range DC bias-resistant transformer under working conditions, constructs constant data sets, variable data sets and algorithm data sets, and constructs reference data sets at the same time; based on the constant data sets, variable data sets and reference data sets, generates a working abnormality data set; based on the working abnormality data set, algorithm data set and reference data set, generates a fault troubleshooting data set; based on the working abnormality data set, fault troubleshooting data set and reference data set, calculates and generates a maintenance cycle; based on the working abnormality data set, fault troubleshooting data set and maintenance cycle, judges and generates a corresponding analysis report, and performs targeted fault repair, replacement or maintenance according to the analysis report. The present invention realizes efficient and accurate detection and prediction of DC bias-resistant transformer faults.

Description

Technical Field

[0001] The invention relates to the technical field of power transformer fault detection, and in particular to a method and system for detecting a transformer fault with a wide metering range and resistance to direct current bias magnetic field. Background Art

[0002] The new power system refers to a power system that takes new energy as the main supplier, ensures energy and power security as the basic premise, meets the power demand of economic and social development as the primary goal, takes a strong smart grid as the hub platform, and is supported by the interaction of source, grid, load and storage and multi-energy complementarity. It has the basic characteristics of clean and low-carbon, safe and controllable, flexible and efficient, intelligent and friendly, and open and interactive. Because new energy has strong randomness, volatility and intermittent characteristics, after large-scale new energy is connected to the grid, the system will face multiple challenges in different time scales such as accurate measurement, stable control, and operation. For example, during the windy winter nights and sunny summer days, wind turbines and solar cells basically operate at full load, while the power generation capacity is weak or basically zero at other times, so the operating current of the transformer fluctuates greatly, far exceeding the metering range of 1%~120% of the rated current of the traditional current transformer; moreover, there are a large number of power users in the power system with irregular electricity consumption, such as seasonal electricity consumption, phased electricity consumption, daily electricity consumption, intermittent electricity consumption and uncertain electricity consumption, which makes the peak-to-valley difference of the power load in the low-voltage substation continue to increase, and the working current variation range of the current transformer in the low-voltage substation is getting wider and wider; at the same time, the uncertainty of the development of power consumption in the substations of residential and industrial users and the unpredictability of load changes in different periods, as well as the fast speed and large range of changes in power load, the peak-to-valley difference can reach more than 30 times, make the current transformer often operate in a low-load state or an overload state, resulting in inaccurate transformer measurement.

[0003] In order to meet the needs of accurate electric energy measurement under wide dynamic range and nonlinear load, research on wide-range transformer technology has been carried out. The implementation method is mainly to change the series-parallel connection mode of the primary winding or select the secondary winding tap by using multiple sets of ratios in the design of the current transformer to select the ratio of the current transformer. However, when using the above-mentioned wide-range current transformer, a DC component will appear in the fluctuation process of the transformer operating current, causing the current transformer to generate DC bias magnetism, and the transformer winding current will generate a large number of harmonics, making it difficult for the current transformer to achieve accurate measurement under the influence of DC bias magnetism, affecting the accuracy of electric energy measurement. The DC magnetic flux generated by the DC component acts on the transformer core, which will cause the core to be magnetically saturated, so that the AC signal cannot be transmitted equivalently, so that the transformer loses its measurement function and then the electricity metering becomes abnormal. For this reason, a new type of current transformer that is resistant to DC bias magnetism has been proposed.

[0004] In actual use, the anti-DC bias transformer can meet the current measurement range of 120 times during operation. However, once the current measurement range is too large, multiple faults may occur, such as secondary circuit open circuit, overheating, core abnormality, output reduction, insulation damage, poor contact and loose winding bolts. When the secondary current path is accidentally interrupted, high voltage will be generated inside the transformer, which may damage the insulation and affect the normal operation of the protection device. In addition, due to poor contact of the primary wiring, severe oxidation of the secondary wiring board surface, short circuit between turns in the anti-DC bias transformer, etc., it is easy to cause overheating, smoking or glue flow in the anti-DC bias transformer. As the working time of the anti-DC bias transformer increases, the long-term operation of the anti-DC bias transformer may cause insulation performance degradation or even breakdown due to environmental factors or aging. Structural and quality defects may also cause poor contact between the screw and the insert screw hole, resulting in an open circuit. Therefore, in order to ensure the reliability and safety of the DC bias transformer, regular maintenance and inspection are essential. When any signs of fault are found, measures should be taken immediately to deal with it. If necessary, the DC bias transformer should be taken out of operation and repaired or replaced.

[0005] At present, the fault data analysis, detection and prediction of traditional DC bias-resistant transformers rely on the quality and integrity of the data. If the collected data is erroneous or incomplete, the accuracy of the final detection and prediction results will be affected. Moreover, when faced with a large amount of real-time monitoring data, the failure of DC bias-resistant transformers may be caused by a variety of factors, such as unqualified manufacturing processes, insulation problems, structural design defects, etc. Complex fault conditions can easily lead to difficulties in accurately analyzing and detecting all types of fault causes, resulting in weak generalization capabilities and slow processing speeds. Summary of the invention

[0006] In order to solve the shortcomings of the above-mentioned prior art, the present invention provides a method and system for detecting faults of a DC bias-resistant transformer with a wide metering range. The method collects multi-dimensional related data of the DC bias-resistant transformer in a working state in real time, calculates and generates a working abnormality data group, a fault troubleshooting data group and a maintenance cycle in sequence according to the data, and generates a corresponding analysis report based on the generated results to perform fault detection and prediction, thereby solving the problems that the fault analysis of traditional DC bias-resistant transformers relies on high-quality data, has limited generalization ability in the face of complex fault situation analysis and has a slow processing speed.

[0007] In a first aspect, the present invention provides a method for detecting faults of a transformer with resistance to direct current bias magnetic field and a wide metering range.

[0008] A method for detecting faults of a transformer with DC bias resistance in a wide metering range, comprising:

[0009] Collect mechanical data, current data and judgment standard data of each wide measurement range of DC bias-resistant transformer under working conditions, build constant data set, variable data set and algorithm data set; at the same time, build a reference data set based on the pre-built database;

[0010] Based on the constant data set, the variable data set and the reference data set, the abnormal data and the type of the transformer are calculated and identified to generate a working abnormal data group;

[0011] Based on the abnormal working data group, the algorithm data set and the reference data set, the fault data and the type of the transformer are calculated and identified to generate a fault troubleshooting data group;

[0012] Calculate and generate maintenance cycles based on the abnormal working data set, the troubleshooting data set and the reference data set;

[0013] Based on the abnormal working data group, troubleshooting data group and maintenance cycle, the corresponding analysis report is generated, and targeted fault repair, replacement or maintenance is carried out according to the analysis report.

[0014] In a further technical solution, the constant data set includes the terminal temperature data of the anti-DC bias transformer in the working state, the data on whether there are cracks or flash marks on the appearance of the loop wiring, the oil level data, the casing data, the core residual magnetism data and the wiring direction data, which are numbered as follows: 、 、 、 、 and ;

[0015] The variable data set includes real-time current data of the DC bias-resistant transformer under working condition, numbered as , , ,... ;

[0016] The algorithm data set includes temperature standard data for judging whether the terminal is overheated, image standard data for judging whether there are cracks or flash marks on the appearance of the loop wiring, capacity standard data for judging whether the oil level is abnormal, cleanliness standard data for judging whether the bushing is contaminated, current standard data for judging whether the iron core is saturated, and operation standard data for judging the wiring direction of the DC bias transformer, which are numbered as follows: 、 、 、... ;

[0017] The reference data set includes a deep learning algorithm model and a maintenance standard data set, which are numbered as and Among them, based on the maintenance standard data set, the maintenance standard algorithm model is obtained through machine learning algorithm training.

[0018] A further technical solution is to calculate and identify abnormal data and its type of the transformer based on the constant data set, the variable data set and the reference data set, and generate a working abnormal data group, including:

[0019] ;

[0020] ;

[0021] In the above formula, Indicates the work abnormality data group, Indicates the integration of all data in the constant data set, It indicates the extreme value data of the maximum current that the DC bias transformer can withstand under working condition. express and Abnormal data of the DC bias-resistant transformer identified by the deep learning algorithm model;

[0022] The abnormal data of the identified DC bias-resistant transformer are numbered. 、 、 、... , constitutes a work abnormality data group; among them, Indicates the type of abnormal data of the DC bias-resistant transformer, including terminal abnormality, circuit abnormality, oil level abnormality, bushing abnormality, core abnormality, wiring abnormality, or current abnormality.

[0023] A further technical solution is to calculate and identify the fault data and its type of the transformer based on the abnormal working data set, the algorithm data set and the reference data set, and generate a fault troubleshooting data set, including:

[0024] ;

[0025] ;

[0026] In the above formula, Indicates the troubleshooting data group, It means that all the data in the algorithm data set are input into the deep learning algorithm model, and the comprehensive prediction model is constructed by combining the troubleshooting data set. It represents the fault data obtained by substituting all the data in the abnormal working data group into the comprehensive prediction model for analysis;

[0027] The fault data of the identified DC bias-resistant transformer are numbered as follows: 、 、 、... , generate a troubleshooting data set; where, Indicates the fault type that causes the DC bias transformer to work abnormally, including terminal fault, circuit fault, oil level fault, bushing fault, core fault, wiring fault or current fault.

[0028] A further technical solution is to calculate and generate a maintenance cycle based on the abnormal working data set, the troubleshooting data set and the reference data set, including:

[0029] ;

[0030] In the above formula, Indicates the maintenance cycle, It means that all the data of the abnormal working data group and the fault troubleshooting data group are input into the maintenance standard algorithm model to obtain the maintenance cycle required to ensure the normal operation of the DC bias magnetic transformer.

[0031] A further technical solution is that when there is data in any one of the working abnormality data group and the fault troubleshooting data group, it is determined that there is a fault abnormality in the anti-DC bias magnetic transformer, a fault report is generated based on the identified fault abnormality data, and the faulty parts of the anti-DC bias magnetic transformer are repaired and replaced in a targeted manner based on the fault report;

[0032] When there is no data in the working abnormality data group and the fault troubleshooting data group, but the actual working time of the anti-DC bias magnetic transformer exceeds the maintenance cycle, it is judged that there is a failure risk of the anti-DC bias magnetic transformer, and a prediction report is generated. According to the prediction report, targeted maintenance is performed on the parts of the anti-DC bias magnetic transformer.

[0033] In a second aspect, the present invention provides a DC bias-resistant transformer fault detection system with a wide metering range.

[0034] A DC bias-resistant transformer fault detection system with a wide metering range, comprising:

[0035] The multi-dimensional acquisition module is used to collect mechanical data, current data and judgment standard data under the working state of each DC bias-resistant transformer with a wide measurement range, and to construct a constant data set, a variable data set and an algorithm data set; at the same time, a reference data set is constructed based on a pre-built database;

[0036] The fault analysis module is used to calculate and identify the abnormal data and types of the transformer based on the constant data set, the variable data set and the reference data set, and generate a working abnormality data group; calculate and identify the fault data and types of the transformer based on the working abnormality data group, the algorithm data set and the reference data set, and generate a fault troubleshooting data group; calculate and generate a maintenance cycle based on the working abnormality data group, the fault troubleshooting data group and the reference data set;

[0037] The prediction management module is used to judge and generate corresponding analysis reports based on the work abnormality data group, fault troubleshooting data group and maintenance cycle, and perform targeted fault repair, replacement or maintenance according to the analysis report.

[0038] One or more of the above technical solutions have the following beneficial effects:

[0039] 1. The present invention provides a method and system for detecting faults of a DC bias-resistant transformer with a wide measurement range. The method collects multi-dimensional related data of the DC bias-resistant transformer in a working state in real time to form a constant data set, a variable data set and an algorithm data set, and constructs a reference data set based on a database. According to the constant data set, the variable data set and the reference data set, a working abnormality data set is calculated and generated. , quickly screen out the mechanical parts data and current data of the DC bias transformer when it works abnormally, and mark the data type of the DC bias transformer working abnormally in detail by number, effectively avoiding errors and omissions in the subsequent fault analysis process; then calculate and generate the fault troubleshooting data group based on the abnormal working data group, algorithm data set and reference data set For the comprehensive faults of anti-DC bias magnetic transformers, a variety of fault analysis methods are flexibly used to deeply analyze the causes of the faults, and the multi-dimensional fault data is identified with high accuracy. The faulty parts that cause the abnormal operation of the anti-DC bias magnetic transformer are marked in detail by number, which effectively avoids errors and omissions in the subsequent maintenance process and realizes efficient and accurate anti-DC transformer fault detection.

[0040] 2. The DC bias-resistant transformer fault detection method and system with a wide metering range proposed by the present invention also generates a working abnormal data set based on the calculation. , Troubleshooting Data Group And reference data sets, calculate and generate maintenance cycles Different fault types of anti-DC bias transformers correspond to maintenance cycles of different lengths. Detection based on the maintenance cycle can more reliably ensure the stability of the operation of the anti-DC bias transformer. On this basis, fault detection and prediction are performed according to the work abnormality data group, fault troubleshooting data group and maintenance cycle, and a fault report or prediction report is generated. Based on the fault report, it is possible to monitor in real time whether abnormal data appears during the operation of the anti-DC bias transformer, respond in time, and avoid the development of key parts from minor faults to serious faults; based on the prediction report, the failure risk of key parts due to excessive current metering range can be effectively predicted, and preventive measures can be taken in advance to reduce the equipment replacement cost caused by failures. This method can further effectively improve the generalization ability and processing speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0042] Figure 1 It is a flow chart of the method for detecting faults of a transformer with DC bias resistance in a wide metering range according to an embodiment of the present invention;

[0043] Figure 2 This is a structural block diagram of a DC bias-resistant transformer fault detection system with a wide metering range according to an embodiment of the present invention. DETAILED DESCRIPTION

[0044] It should be noted that the following detailed descriptions are exemplary only, are intended to describe specific embodiments, are intended to provide further explanation of the present invention, and are not intended to limit exemplary embodiments according to the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those of ordinary skill in the art to which the present invention belongs. In addition, it should also be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0045] Embodiment 1

[0046] Considering that the analysis and prediction of traditional DC bias transformer fault data rely on high-quality data, have limited generalization ability and slow processing speed in the face of complex fault conditions, this embodiment provides a DC bias transformer fault detection method with a wide metering range, such as Figure 1 As shown, the method specifically comprises the following steps:

[0047] Step S1, collecting mechanical data, current data and judgment standard data of each wide measurement range of anti-DC bias transformer under working state, constructing constant data set, variable data set and algorithm data set; at the same time, constructing a reference data set based on the pre-constructed database;

[0048] Step S2, based on the constant data set, the variable data set and the reference data set, calculate and identify the abnormal data and the type of the transformer, and generate a working abnormal data group;

[0049] Step S3, based on the abnormal working data set, the algorithm data set and the reference data set, calculate and identify the fault data and its type of the transformer, and generate a fault troubleshooting data set;

[0050] Step S4, calculating and generating a maintenance cycle based on the abnormal working data set, the troubleshooting data set and the reference data set;

[0051] Step S5: Based on the abnormal working data group, the troubleshooting data group and the maintenance cycle, a corresponding analysis report is generated, and targeted fault repair, replacement or maintenance is performed according to the analysis report.

[0052] The DC bias-resistant transformer fault detection method with a wide metering range proposed in this embodiment is introduced in more detail through the following content.

[0053] In step S1, first, mechanical data, current data and judgment standard data of each wide measurement range anti-DC bias transformer in working state are collected to construct a constant data set, a variable data set and an algorithm data set.

[0054] Specifically, in the working state, the relevant data of the DC bias-resistant transformer in each wide measurement range within the set time period are collected, including: (1) for the constant data set: the mechanical data of the DC bias-resistant transformer in the working state are collected and numbered, and the constant data sets are numbered as follows: , , , , and , that is, the constant data set includes the terminal temperature data of the anti-DC bias transformer in the working state, the data on whether there are cracks or flash marks on the circuit wiring appearance, the oil level data, the bushing data, the core residual magnetism data and the wiring direction data. The above collected data are all constant values; (2) For the variable data set: the current data of the anti-DC bias transformer in the working state is collected and numbered, and the variable data set number is , , ,... , nIndicates the number of data, and the variable data set number corresponds to the current data displayed on the DC bias transformer each time; (3) For the algorithm data set: the judgment standard data for the DC bias transformer fault analysis in the big data platform is collected and numbered, and the algorithm data sets are numbered as follows: , , ,... , n Indicates the number of data. The algorithm data set includes but is not limited to the temperature standard for judging whether the terminal is overheated, the image standard for judging whether there are cracks or flash marks on the appearance of the circuit wiring, the capacity standard for judging whether the oil level is abnormal, the cleanliness standard for judging whether the bushing is contaminated, the current standard for judging whether the core is saturated, and the operation standard for judging the wiring direction of the DC bias transformer. By collecting constant data, variable data, and judgment standard data as described above, more comprehensive information can be obtained to avoid detection errors caused by a single perspective, which is conducive to the subsequent timely discovery of abnormal problems, fault causes, or unnoticed correlations.

[0055] At the same time, a reference data set is constructed based on a pre-constructed database (including historical normal operating data, abnormal data, fault data, etc. of the DC bias-resistant transformer). The reference data set is a tool that contains pre-defined standards and pre-trained models. It plays the role of a benchmark or reference in fault detection and is used to assist in identifying and classifying the normal working state and abnormal state of the transformer. Specifically, the reference data set includes a pre-trained deep learning algorithm model and a maintenance standard data set, and the maintenance standard algorithm model can be obtained through algorithm training based on the maintenance standard data set (including maintenance standards and historical maintenance data). The pre-trained deep learning algorithm model and the maintenance standard algorithm model are numbered as and .

[0056] In step S2, based on the constant data set, the variable data set and the reference data set, the abnormal data of the transformer and its type are calculated and identified to generate a working abnormal data group.

[0057] Specifically, according to the constant data set, variable data set and reference data set, the working abnormal data set is calculated and generated by the following calculation formula: The calculation formula is:

[0058] ;

[0059] ;

[0060] In the above formula, Indicates the work abnormality data group, Indicates the integration of all data in the constant data set, It indicates the extreme value data of the maximum current that the DC bias transformer can withstand under working condition. express and Abnormal data of the DC bias-resistant transformer identified by the deep learning algorithm model.

[0061] The above-mentioned pre-trained deep learning algorithm model can be implemented by any of a variety of existing machine learning algorithms. In this embodiment, the above-mentioned deep learning algorithm model is implemented by the isolation forest algorithm. The isolation forest algorithm is a tree-based anomaly detection algorithm. As an automatic anomaly detection algorithm, it can use an unsupervised learning method to detect abnormal data points, and then delete the abnormal data points from the training data, and retrain the model on the data set with the outliers removed, so as to improve the model performance. Therefore, in this embodiment, by inputting the data in the constant data set and the variable data set into the model, it can be used to predict or identify whether the state of the mutual inductor is normal, identify the abnormal data therein, and the identified abnormal data constitutes the working abnormal data group.

[0062] On this basis, according to the abnormal data set , can quickly screen out the mechanical parts data and current data of the abnormal operation of the DC bias magnetic transformer, so as to facilitate further in-depth analysis of the cause of the fault.

[0063] Getting the abnormal data set Afterwards, the abnormal data groups of the identified DC bias-resistant transformers are numbered as 、 、 、... , generate a work abnormality data set, where Indicates the type of abnormal data of the DC bias transformer, including terminal abnormality, circuit abnormality, oil level abnormality, bushing abnormality, core abnormality, wiring abnormality or current abnormality. The data type of the DC bias transformer working abnormality is marked in detail by numbering, effectively avoiding errors and omissions in the subsequent fault analysis process.

[0064] In step S3, based on the abnormal working data set, the algorithm data set and the reference data set, the fault data and the type of the transformer are calculated and identified to generate a fault troubleshooting data set.

[0065] Specifically, according to the abnormal work data group , algorithm data set and reference data set, and generate the troubleshooting data set by the following calculation formula The calculation formula is:

[0066] ;

[0067] ;

[0068] In the above formula, Indicates the troubleshooting data group, It means that all the data in the algorithm data set are input into the pre-trained deep learning algorithm model, and the isolation forest algorithm is used to build a comprehensive prediction model in conjunction with the fault troubleshooting data set, which is used to flexibly integrate multiple fault analysis methods, which is conducive to improving the generalization ability of the prediction model; Indicates that the abnormal data group is changed from 1 to n Substitute all the data into the comprehensive prediction model to analyze the fault data.

[0069] In the above step S3, the algorithm data set is actually used as the data standard, and the pre-trained model of the isolation forest algorithm is used to jointly construct a comprehensive prediction model. The abnormal data detected in step S2 is input into the comprehensive prediction model to perform abnormal verification and fault identification, effectively output fault data, and generate a fault troubleshooting data set. In the above manner, according to the fault troubleshooting data set , flexibly use a variety of fault analysis methods to deeply analyze the cause of the fault, and for the comprehensive fault of the DC bias magnetic transformer, the multi-dimensional identification fault data has high accuracy.

[0070] Getting Troubleshooting Data Group Afterwards, the fault data groups of the identified DC bias-resistant transformers are numbered as 、 、 、... , generate a troubleshooting data set; where, Indicates the fault type that causes the abnormal operation of the anti-DC bias magnetic transformer, including terminal fault, circuit fault, oil level fault, bushing fault, core fault, wiring fault or current fault. The faulty parts that cause the abnormal operation of the anti-DC bias magnetic transformer are marked in detail by numbering, effectively avoiding errors and omissions in the subsequent repair and maintenance process.

[0071] In step S4, a maintenance cycle is calculated and generated based on the abnormal working data set, the fault troubleshooting data set and the reference data set.

[0072] Specifically, according to the abnormal work data group , Troubleshooting Data Group and reference data set to calculate the maintenance cycle The calculation formula is:

[0073] ;

[0074] In the above formula, Indicates the maintenance cycle, Indicates that the abnormal data group , Troubleshooting Data Group All data are input into the maintenance standard algorithm model (built based on the maintenance standard data set) to obtain the maintenance cycle required to ensure the normal operation of the DC bias transformer. Different fault types of anti-DC bias magnetic transformers correspond to maintenance cycles of different lengths, which can more reliably ensure the stability of the operation of the anti-DC bias magnetic transformers.

[0075] The above maintenance standard algorithm model is obtained through algorithm training based on the maintenance standard data set (including maintenance standards and historical maintenance data), and is used to calculate the maintenance cycle of the DC bias transformer. In this embodiment, based on the maintenance standards and historical maintenance data, the model is trained using machine learning techniques (such as decision trees, random forests, neural networks, etc.), and the accuracy of the model is verified by cross-validation and other methods, and the trained model is deployed to the actual maintenance system, and its performance is continuously monitored, and the model is adjusted according to feedback. The abnormal working data and troubleshooting data are input into the maintenance standard algorithm model as maintenance data to predict the maintenance needs and optimal maintenance time of the equipment.

[0076] In step S5, based on the abnormal working data group, the troubleshooting data group and the maintenance cycle, a corresponding analysis report is generated, and targeted fault repair, replacement or maintenance is performed according to the analysis report.

[0077] Specifically, when the abnormal data group and troubleshooting data sets When there is data in any of the data groups, it is determined that there is a fault abnormality in the anti-DC bias magnetic transformer. A fault report is generated based on the identified fault abnormality data, so that it can monitor in real time whether abnormal data appears during the operation of the anti-DC bias magnetic transformer, respond in time, and avoid the development of key parts from minor faults to serious faults. Preferably, on this basis, the faulty parts of the anti-DC bias magnetic transformer can be repaired and replaced in a targeted manner according to the fault report.

[0078] When abnormal data group and troubleshooting data sets There is no data, but the actual working time of the DC bias transformer exceeds the maintenance cycle When the DC bias transformer is detected, it is judged that there is a risk of failure, and a prediction report is generated, which can effectively predict the risk of failure of key parts due to excessive current measurement range, take preventive measures in advance, and reduce the cost of equipment replacement caused by failure. Preferably, on this basis, the parts of the DC bias transformer can be maintained in a targeted manner according to the prediction report.

[0079] Embodiment 2

[0080] This embodiment provides a DC bias-resistant transformer fault detection system with a wide metering range. Figure 2 As shown, it specifically includes the following modules:

[0081] The multi-dimensional acquisition module is used to collect mechanical data, current data and judgment standard data under the working state of each DC bias-resistant transformer with a wide measurement range, and to construct a constant data set, a variable data set and an algorithm data set; at the same time, a reference data set is constructed based on a pre-built database;

[0082] The fault analysis module is used to calculate and identify the abnormal data and types of the transformer based on the constant data set, the variable data set and the reference data set, and generate a working abnormality data group; calculate and identify the fault data and types of the transformer based on the working abnormality data group, the algorithm data set and the reference data set, and generate a fault troubleshooting data group; calculate and generate a maintenance cycle based on the working abnormality data group, the fault troubleshooting data group and the reference data set;

[0083] The prediction management module is used to judge and generate corresponding analysis reports based on the work abnormality data group, fault troubleshooting data group and maintenance cycle, and perform targeted fault repair, replacement or maintenance according to the analysis report.

[0084] Specifically, Figure 2 As shown, the multi-dimensional acquisition module includes a constant acquisition unit, a variable acquisition unit and an algorithm data unit, and the multi-dimensional acquisition module is connected to the fault analysis module through a network. The constant acquisition unit is connected to the DC bias magnetic transformer through a network to collect and construct a constant data set, and the collected constant data set is numbered and transmitted to the fault analysis module; the variable acquisition unit is connected to the online monitoring device through a network to collect variable data sets, and the collected variable data sets are numbered and transmitted to the fault analysis module; the algorithm data unit is connected to the big data platform through a network to collect algorithm data sets, and the collected judgment standard data are numbered and transmitted to the fault analysis module.

[0085] The above-mentioned fault analysis module includes a real-time monitoring unit, an identification and troubleshooting unit and a maintenance analysis unit. The fault analysis module is connected to the prediction management module through a network. Moreover, the fault analysis module collects reference data sets through a network connection database and caches them to the fault analysis module after numbering. Among them, the real-time monitoring unit calculates and generates a work abnormality data set based on the constant data set, the variable data set and the reference data set. , after the data group is numbered, it is transmitted to the identification and troubleshooting unit; the identification and troubleshooting unit detects the abnormal data group according to the abnormality of the work. , algorithm data set and reference data set, calculate and generate the troubleshooting data set , after the data group is numbered, it is transmitted to the maintenance analysis unit; the maintenance analysis unit , Troubleshooting Data Group and reference data set to calculate the maintenance cycle .

[0086] The above prediction management module includes a fault prediction unit and a fault management unit. The fault prediction unit generates a fault prediction result based on the abnormal working data set. , Troubleshooting Data Group and maintenance cycle , judge and generate a corresponding analysis report, and transmit it to the fault management unit through the network; the fault management unit then repairs or maintains the anti-DC bias magnetic transformer according to the corresponding analysis report.

[0087] The steps involved in the above embodiment 2 correspond to those in the method embodiment 1. For the specific implementation method, please refer to the relevant description part of the embodiment 1.

[0088] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0089] The above description is only a preferred embodiment of the present invention. Although the specific implementation mode of the present invention is described in conjunction with the accompanying drawings, it is not a limitation of the protection scope of the present invention. Those skilled in the art should understand that on the basis of the technical solution of the present invention, various modifications or variations that can be made by those skilled in the art without creative work are still within the protection scope of the present invention.

Claims

1. A method for detecting faults of a transformer with DC bias resistance in a wide metering range, characterized in that: include: Collect mechanical data, current data and judgment standard data of each wide measurement range of DC bias-resistant transformer under working conditions, build constant data set, variable data set and algorithm data set; at the same time, build a reference data set based on the pre-built database; The algorithm data set includes temperature standard data for judging whether the terminal is overheated, image standard data for judging whether there are cracks or flash marks on the appearance of the loop wiring, capacity standard data for judging whether the oil level is abnormal, cleanliness standard data for judging whether the bushing is contaminated, current standard data for judging whether the iron core is saturated, and operation standard data for judging the wiring direction of the DC bias transformer, which are numbered as follows: 、 、 、... , n=6; The reference data set includes a deep learning algorithm model and a maintenance standard data set, which are numbered as and ; Among them, based on the maintenance standard data set, the maintenance standard algorithm model is obtained through machine learning algorithm training; Based on the constant data set, the variable data set and the reference data set, the abnormal data and the type of the transformer are calculated and identified to generate a working abnormal data group; Based on the abnormal working data set, algorithm data set and reference data set, calculate and identify the fault data and its type of the transformer, and generate a fault troubleshooting data set, including: ； ； In the above formula, Indicates troubleshooting data. It means that all the data in the algorithm data set are input into the deep learning algorithm model, and the comprehensive prediction model is constructed by combining the troubleshooting data set. It represents the fault data obtained by substituting all the data in the abnormal working data group into the comprehensive prediction model for analysis; The fault data of the identified DC bias-resistant transformer are numbered as follows: 、 、 、... , generate a troubleshooting data set; where, Indicates the fault type that causes the DC bias transformer to work abnormally, including terminal fault, circuit fault, oil level fault, bushing fault, core fault, wiring fault or current fault; Calculate and generate maintenance cycles based on the abnormal working data set, the troubleshooting data set and the reference data set; Based on the abnormal working data group, troubleshooting data group and maintenance cycle, the corresponding analysis report is generated, and targeted fault repair, replacement or maintenance is carried out according to the analysis report.

2. A method for detecting faults of a transformer with wide metering range and resistance to DC bias magnetic field as claimed in claim 1, characterized in that: The constant data set includes the terminal temperature data of the anti-DC bias transformer in the working state, the data on whether the loop wiring has cracks or flash marks, the oil level data, the bushing data, the core residual magnetism data and the wiring direction data, which are numbered as follows: 、 、 、 、 and ; The variable data set includes real-time current data of the DC bias-resistant transformer under working condition, numbered as , , ,... .

3. A method for detecting faults of a transformer with wide metering range and resistant to DC bias magnetic field as claimed in claim 2, characterized in that: Based on the constant data set, variable data set and reference data set, calculate and identify the abnormal data and its type of the transformer, and generate a working abnormal data group, including: ； ； In the above formula, Indicates the work abnormality data group, Indicates the integration of all data in the constant data set, It indicates the extreme value data of the maximum current that the DC bias transformer can withstand under working condition. express and Abnormal data of the DC bias-resistant transformer identified by the deep learning algorithm model; The abnormal data of the identified DC bias-resistant transformer are numbered. 、 、 、... , constitutes a work abnormality data group; among them, Indicates the type of abnormal data of the DC bias-resistant transformer, including terminal abnormality, circuit abnormality, oil level abnormality, bushing abnormality, core abnormality, wiring abnormality, or current abnormality.

4. A method for detecting faults of a transformer with wide measurement range and resistance to DC bias magnetic field as claimed in claim 2, characterized in that: Based on the abnormal working data set, troubleshooting data set and reference data set, the maintenance cycle is calculated and generated, including: ； In the above formula, Indicates the maintenance cycle, It means that all the data of the abnormal working data group and the fault troubleshooting data group are input into the maintenance standard algorithm model to obtain the maintenance cycle required to ensure the normal operation of the DC bias magnetic transformer.

5. A method for detecting faults of a transformer with wide metering range and resistance to DC bias magnetic field as claimed in claim 1, characterized in that: When data exists in any of the working abnormality data group and the fault troubleshooting data group, it is determined that the DC bias magnetic transformer has a fault abnormality, and a fault report is generated based on the identified fault abnormality data, and the faulty parts of the DC bias magnetic transformer are repaired and replaced in a targeted manner based on the fault report; When there is no data in the working abnormality data group and the fault troubleshooting data group, but the actual working time of the anti-DC bias magnetic transformer exceeds the maintenance cycle, it is judged that there is a failure risk of the anti-DC bias magnetic transformer, and a prediction report is generated. According to the prediction report, targeted maintenance is performed on the parts of the anti-DC bias magnetic transformer.

6. A DC bias-resistant transformer fault detection system with a wide metering range, characterized in that: include: The multi-dimensional acquisition module is used to collect mechanical data, current data and judgment standard data under the working state of each DC bias-resistant transformer with a wide measurement range, and to construct a constant data set, a variable data set and an algorithm data set; at the same time, a reference data set is constructed based on a pre-built database; A fault analysis module is used to calculate and identify abnormal data and its type of the transformer based on the constant data set, the variable data set and the reference data set, and generate a working abnormal data group; Based on the abnormal working data group, the algorithm data set and the reference data set, the fault data and the type of the transformer are calculated and identified to generate a fault troubleshooting data group; Calculate and generate maintenance cycles based on the abnormal working data set, the troubleshooting data set and the reference data set; The prediction management module is used to judge and generate corresponding analysis reports based on the abnormal working data group, fault troubleshooting data group and maintenance cycle, and perform targeted fault repair, replacement or maintenance according to the analysis report; The algorithm data set includes temperature standard data for judging whether the terminal is overheated, image standard data for judging whether there are cracks or flash marks on the appearance of the loop wiring, capacity standard data for judging whether the oil level is abnormal, cleanliness standard data for judging whether the bushing is contaminated, current standard data for judging whether the iron core is saturated, and operation standard data for judging the wiring direction of the DC bias transformer, which are numbered as follows: 、 、 、... , n=6; The reference data set includes a deep learning algorithm model and a maintenance standard data set, which are numbered as and ; Among them, based on the maintenance standard data set, the maintenance standard algorithm model is obtained through machine learning algorithm training; Based on the abnormal working data set, algorithm data set and reference data set, calculate and identify the fault data and its type of the transformer, and generate a fault troubleshooting data set, including: ； ； In the above formula, Indicates troubleshooting data. It means that all the data in the algorithm data set are input into the deep learning algorithm model, and the comprehensive prediction model is constructed by combining the troubleshooting data set. It represents the fault data obtained by substituting all the data in the abnormal working data group into the comprehensive prediction model for analysis; The fault data of the identified DC bias-resistant transformer are numbered as follows: 、 、 、... , generate a troubleshooting data set; where, Indicates the fault type that causes the DC bias transformer to work abnormally, including terminal fault, circuit fault, oil level fault, bushing fault, core fault, wiring fault or current fault.

7. A wide-metering-range DC bias-resistant transformer fault detection system as claimed in claim 6, characterized in that: The constant data set includes the terminal temperature data of the anti-DC bias transformer in the working state, the data on whether the loop wiring has cracks or flash marks, the oil level data, the bushing data, the core residual magnetism data and the wiring direction data, which are numbered as follows: 、 、 、 、 and ; The variable data set includes real-time current data of the DC bias-resistant transformer under working condition, numbered as , , ,... ; The algorithm data set includes temperature standard data for judging whether the terminal is overheated, image standard data for judging whether there are cracks or flash marks on the appearance of the loop wiring, capacity standard data for judging whether the oil level is abnormal, cleanliness standard data for judging whether the bushing is contaminated, current standard data for judging whether the iron core is saturated, and operation standard data for judging the wiring direction of the DC bias transformer, which are numbered as follows: 、 、 、... ; The reference data set includes a deep learning algorithm model and a maintenance standard data set, which are numbered as and Among them, based on the maintenance standard data set, the maintenance standard algorithm model is obtained through machine learning algorithm training.

8. A DC bias-resistant transformer fault detection system with a wide metering range as claimed in claim 7, characterized in that: Based on the constant data set, variable data set and reference data set, calculate and identify the abnormal data and its type of the transformer, and generate a working abnormal data group, including: ； ； In the above formula, Indicates the work abnormality data group, Indicates the integration of all data in the constant data set, It indicates the extreme value data of the maximum current that the DC bias transformer can withstand under working condition. express and Abnormal data of the DC bias-resistant transformer identified by the deep learning algorithm model; The abnormal data of the identified DC bias-resistant transformer are numbered. 、 、 、... , constitutes a work abnormality data group; among them, Indicates the type of abnormal data of the DC bias-resistant transformer, including terminal abnormality, circuit abnormality, oil level abnormality, bushing abnormality, core abnormality, wiring abnormality, or current abnormality.

9. A wide-metering-range DC bias-resistant transformer fault detection system as claimed in claim 7, characterized in that: Based on the abnormal working data set, troubleshooting data set and reference data set, the maintenance cycle is calculated and generated, including: ； In the above formula, Indicates the maintenance cycle, It means that all the data of the abnormal working data group and the fault troubleshooting data group are input into the maintenance standard algorithm model to obtain the maintenance cycle required to ensure the normal operation of the DC bias magnetic transformer.

10. A DC bias-resistant transformer fault detection system with a wide metering range as claimed in claim 6, characterized in that: When data exists in any of the working abnormality data group and the fault troubleshooting data group, it is determined that the DC bias magnetic transformer has a fault abnormality, and a fault report is generated based on the identified fault abnormality data, and the faulty parts of the DC bias magnetic transformer are repaired and replaced in a targeted manner based on the fault report; When there is no data in the working abnormality data group and the fault troubleshooting data group, but the actual working time of the anti-DC bias magnetic transformer exceeds the maintenance cycle, it is judged that there is a failure risk of the anti-DC bias magnetic transformer, and a prediction report is generated. According to the prediction report, targeted maintenance is performed on the parts of the anti-DC bias magnetic transformer.

Images