CN114280372A

CN114280372A - Reactor multi-scale fault diagnosis and monitoring system

Info

Publication number: CN114280372A
Application number: CN202111592551.8A
Authority: CN
Inventors: 全玉生; 全曦
Original assignee: Individual
Current assignee: Individual
Priority date: 2021-12-24
Filing date: 2021-12-24
Publication date: 2022-04-05

Abstract

The invention provides a reactor multi-scale fault diagnosis and monitoring system based on holographic electrical quantities (steady state, transient state voltage and current). The method takes steady-state and transient fault recording data of voltage and current of each terminal of the reactor as a signal source, and takes the variable quantity of equivalent electrical parameters (impedance Z and admittance Y) of the reactor as a first criterion; taking the distribution probability of the equivalent electrical parameter variable quantity of the reactor in a certain period of time as a second criterion; taking the polarity effect of the time-shifting change rate of the equivalent electrical parameter of the reactor as a third criterion; constructing an overall Abnormal rate (Absolute Ratio AR) matrix based on a first criterion, a second criterion and a third criterion, diagnosing and monitoring multi-scale faults or insulation defects of the reactor according to numerical values of elements of the overall Abnormal rate matrix AR, and identifying positions and severity of faults or defects of the reactor, such as overheating faults, electric arcs, partial discharges, winding deformation, turn-to-turn short circuits and the like.

Description

Reactor multi-scale fault diagnosis and monitoring system

Technical Field

The invention belongs to the field of diagnosis and monitoring of reactor faults or insulation defects. The invention provides a reactor multi-scale fault diagnosis and monitoring system based on holographic electrical quantities (steady state, transient state voltage and current). The invention takes steady state and transient state fault recording data of voltage and current of each terminal of a reactor as a signal source, and takes the variable quantity of equivalent electrical parameters (impedance Z and admittance Y) of the reactor as a first criterion; taking the distribution probability of the equivalent electrical parameter variable quantity of the reactor in a certain period of time as a second criterion; taking the polarity effect of the time-shifting change rate of the equivalent electrical parameter of the reactor as a third criterion; constructing an overall Abnormal rate (Absolute Ratio AR) matrix based on a first criterion, a second criterion and a third criterion, diagnosing and monitoring multi-scale faults or insulation defects of the reactor according to numerical values of elements of the overall Abnormal rate matrix AR, and identifying positions and severity of faults or defects of the reactor, such as overheating faults, electric arcs, partial discharges, winding deformation, turn-to-turn short circuits and the like.

Technical Field

The reactor is an important compensation device of the power system, plays roles in suppressing overvoltage, eliminating secondary arc current, adjusting reactive power and the like of the power system, and the operation condition of the reactor is closely related to the safety, stability and economic operation of the power system. The research and development of the reactor fault diagnosis and monitoring system has important theoretical significance and engineering practical value.

The DGA method can trace back to 1928, is the most effective diagnosis method for latent faults of the oil-filled reactor, and is generally concerned by domestic and foreign experts, scholars and engineering technicians. IEC statistical data shows that the DGA method effectively solves the problem of diagnosing about 50% of latent faults of oil-filled reactors in various countries in the world. However, the DGA method also has some drawbacks, (1) is not effective for faster developing faults; (2) gas production has aftereffects and cumulative effects, and gas components and codes can still exist for a period of time even if transient faults disappear; (3) failure to diagnose the core leg and winding where the fault is located; (4) common winding deformation faults cannot be diagnosed.

The searched references show that no effective method for comprehensively realizing fault diagnosis and monitoring of the reactor only by using steady-state or transient fault recording data as a signal source exists so far. Major accidents caused by reactor faults occur for many times at home and abroad, and the urgency and the severity of diagnosis and monitoring of the reactor faults or insulation defects can be seen. Aiming at the problems existing in the diagnosis and monitoring of the fault or the insulation defect of the reactor and the current research situation, a method and a system for effectively diagnosing and monitoring the fault or the insulation defect of the reactor are still a difficult problem to be solved urgently.

Disclosure of Invention

The invention provides a reactor multi-scale fault diagnosis and monitoring system, which is a fault diagnosis and monitoring system consisting of hardware and software and can carry out multi-scale fault diagnosis and monitoring on a reactor. The diagnosis and monitoring method and the system are reliable, practical, economical, convenient and fast, have strong operability, and have no relevant report in the current retrieved reference documents.

1. The main content of the invention

(1) The invention takes the variable quantity of the equivalent electrical parameters (impedance Z and admittance Y) of the reactor as the first criterion for fault diagnosis and monitoring. Equivalent electrical parameters of faults or insulation defects such as electric arc, overheating, partial discharge, winding deformation and the like correspond to a certain operation working condition; the variation of the electrical parameters in normal and fault states is an important basis for identifying the fault or insulation defect of the reactor.

(2) The invention takes the distribution probability of the equivalent electrical parameter variable quantity of the reactor in a certain period of time as the second criterion of fault diagnosis. Any engineering practical system has certain nonlinearity and uncertainty in nature, and faults or insulation defects of the reactor have certain randomness. The distribution probability of the equivalent electrical parameter variable quantity of the reactor is an important basis for identifying the fault or insulation defect of the reactor.

(3) The invention takes the polarity effect of the time-shifting change rate of the equivalent electrical parameter of the reactor as the third criterion of fault diagnosis. The insulation defect of the reactor is not artificially set, and the equivalent electrode system which generates faults such as electric arc, partial discharge and the like or insulation defects is mostly in a rod-plate structure and generally has a certain polarity effect. The polarity effect of the equivalent electrical parameter time-shift change rate is an important basis for identifying the fault and the insulation defect of the reactor.

(4) Book (I)The invention constructs an overall Abnormal rate (Abnormal Ratio AR) matrix based on the 'variation of the equivalent electrical parameter', 'distribution probability of the variation of the equivalent electrical parameter' and 'polar effect of time-shift variation rate of the equivalent electrical parameter' ((1), (2), (3)). And according to the numerical values of all elements of the comprehensive abnormal rate matrix AR, diagnosing and monitoring the multi-scale fault or the insulation defect of the reactor.

The rows respectively correspond to the category of the fault or the insulation defect of the reactor, the core column where the fault or the insulation defect is located, the winding where the fault or the insulation defect is located, and the severity of the fault or the insulation defect; the columns respectively correspond to signal source data reliability criteria, reactor overall faults, arc faults or insulation defects, partial discharge faults or insulation defects, winding deformation faults or defects, overheating faults or defects, and turn-to-turn short circuit faults or defects. According to the value of each element of the comprehensive abnormal rate matrix AR, diagnosis and monitoring of multi-scale faults or insulation defects within the range of 1% -100% of equivalent electrical parameter change of the reactor can be conveniently realized.

The first criterion, the second criterion and the third criterion are properly combined, the essential characteristics of the fault or the insulation defect of the reactor can be locked from three important aspects of 'the variation of the equivalent electrical parameter', 'the distribution probability of the variation of the equivalent electrical parameter' and 'the polarity effect of the time-shifting variation rate of the equivalent electrical parameter', and the accuracy and the reliability of diagnosis and monitoring of the fault or the insulation defect of the reactor are ensured. And the comprehensive abnormal rate matrix AR constructed based on the first, second and third criteria ensures the operability and convenience of diagnosis and monitoring of the fault or insulation defect of the reactor.

The reactor multi-scale fault diagnosis and monitoring system has universality, and is not only suitable for various oil-filled reactors, but also suitable for various dry-type reactors.

2. Main technical conditions and functions of the invention

The main technical conditions and functions of the reactor multi-scale fault diagnosis and monitoring system are as follows.

I. Signal source

The steady state or transient state fault recording data (voltage, current and zero sequence voltage, current-holographic electric quantity) of the reactor.

Major functions of the diagnosis and monitoring system

The diagnosis and monitoring of the multi-scale fault or the insulation defect are mainly divided into the following three levels.

(1) And diagnosing the position of the fault or the insulation defect of the reactor and the severity of the fault or the insulation defect. 1) Diagnosing whether the whole reactor has faults or insulation defects and the severity of the faults or the insulation defects; 2) diagnosing low, medium and high energy faults or insulation defects of the reactor body and positions thereof; 3) diagnosing a main reactor, a neutral point small reactor and an ABC core column where the fault or the insulation defect is located; 4) diagnosing the reactor with or without an arc fault, the position of the arc fault and the severity of the arc fault; 5) diagnosing whether the reactor has winding deformation, the position of the winding deformation and the severity of the winding deformation; 6) diagnosing whether the reactor winding has partial discharge, the position of the partial discharge and the severity of the partial discharge; 7) and diagnosing whether the reactor winding has turn-to-turn short circuit, the position of the turn-to-turn short circuit and the severity of the turn-to-turn short circuit.

(2) Multi-scale division of fault or insulation defect severity. Dividing faults or insulation defects into four categories (18 levels) of maintenance, strengthening supervision, maintenance and urgent maintenance, sequencing from top to bottom and from low to high, and having the characteristic of screening from tiny insulation defects to large hidden troubles of faults; with the increase of the identification scale, the diagnosis and monitoring system gradually focuses from the details (tiny) of the insulation defect to the insulation defect with larger harm or the fault with larger hidden danger, and provides quantitative judgment basis for maintenance decision; the maintenance work is promoted to be purposeful, the maintenance is not needed, the supervision is strengthened, the maintenance is needed, and the first maintenance is needed urgently, so that the condition of mild and severe is divided, and the excessive maintenance is avoided; the electric reactor state maintenance is changed to intelligent maintenance, and an important technical guarantee is provided.

(3) Fault or insulation defect diagnosis and monitoring is ordered on multiple scales of reliability and sensitivity. Redundant, variable focus diagnostic results are provided for both reliability and sensitivity. The reliability and the sensitivity are mutually restricted, and the fish and the bear paw can not be obtained at the same time, but can be reasonably adjusted between the two. The overall diagnosis of the fault is made by using a high-reliability criterion, and the missing fault details of the high-reliability criterion are determined by using a high-sensitivity criterion; if the high-reliability criterion can make complete fault diagnosis, the diagnosis result of the high-sensitivity criterion can further prove the details of the high-reliability criterion.

3 Key technical analysis of the invention

(1) The steady state recording data of the reactor is a stable random process, and the statistical characteristic quantities of the variation of the equivalent electrical parameters, the distribution probability of the variation of the equivalent electrical parameters and the polarity effect of the time-shifting variation rate of the equivalent electrical parameters correspond to corresponding mathematical expectation, variance and covariance and do not change along with time within a certain time period, so that the fault or defect of the reactor can be revealed; transient recording data of the reactor is a non-stationary random process and may contain mutation information of fault types or severity, and mutation points and mutation intervals of voltage and current of the transient recording data correspond to critical points of fault types or modes of the reactor and are important basis for distinguishing normal working conditions from abnormal faults or insulation defects, so that the types and the severity of the faults or the insulation defects of the reactor can be disclosed.

(2) The fault development process of the reactor has certain randomness and certain fuzziness, so that the comprehensive abnormal rate matrix AR is adopted to diagnose and monitor the fault of the reactor or the position of an insulation defect (A, B, C columns, a main reactor and a small neutral point reactor). The reactor fault or insulation defect diagnosis and monitoring system is essentially a deep data mining system; the subjective and objective statistical rules of the fault or the insulation defect of the reactor are fused, and the comprehensive monitoring and diagnosis of the reactor are realized. The reactor fault diagnosis and monitoring system is free of embedded sensors, primary wiring of the system is not changed, fault diagnosis and monitoring can be implemented only by sharing partial voltage and current data of self-protection or fault recording, simplicity and practicability are achieved, operability is high, and the technical advantages of the reactor multi-scale fault diagnosis and monitoring system are theoretically guaranteed.

(3) The transient fault recording data of the electric reactor is a non-stable random process, and the time-space characteristics of the maximum and minimum break variables are adopted, so that the physical essence (time-space break points) of sudden faults accumulated from defects of the electric reactor is grasped, and the fault type, position and severity of the electric reactor can be diagnosed more accurately and effectively. The diagnosis method adopting the steady-state fault recording data is an active diagnosis method, can be used for online monitoring, and can call the fault recording data at any time according to the requirement; the diagnosis method adopting the transient fault recording data is a passive diagnosis method, can be solved when meeting, but the transient data contains a large number of high-amplitude abrupt change and rich high harmonic components, and can further reveal the details, the position and the severity of the defect or the fault; the two complement each other to complement each other.

Compared with the existing reactor fault diagnosis and monitoring method, the diagnosis and monitoring method provided by the invention has certain principles and technical advantages. The method is completely different from the existing diagnosis method, the diagnosis and monitoring method provided by the invention is not only a product but also a phenomenon, and takes the 'variation of the equivalent electrical parameter', 'distribution probability of the variation of the equivalent electrical parameter' and 'polar effect of the time-shift variation rate of the equivalent electrical parameter' caused by the fault as the basis of diagnosis and monitoring, thereby capturing the physical essence of the fault or the insulation defect to a certain extent. The effectiveness and reliability of diagnosis and monitoring are theoretically ensured.

The diagnosis and monitoring system provided by the invention only needs the fault recording data of the voltage and the current of each terminal of the reactor to be diagnosed and monitored, has no other pre-embedded sensors, does not change primary equipment and wiring of the system, does not change secondary main equipment and wiring of the system (only relates to one group of output data of a secondary recorder), has universality, transportability, operability and the like, and is obviously superior to the existing diagnosis and monitoring method.

The diagnosis and monitoring system provided by the invention is essentially a deep data mining system, and the function of fault recording data is greatly expanded. Almost every substation of the power system is provided with a fault recorder, the original fault recording data is used for analyzing the system fault after the accident, judging the fault type, severity and the like of the reactor, and is 'Zhuge Liang' after the accident; the multi-scale reactor fault diagnosis and monitoring system provided by the invention can extract and call steady-state fault recording data at any time, diagnose and monitor the reactor fault or insulation defect on line, and realize 'Zhuge Liang' in the process; transient fault recording data can be adopted to diagnose and monitor faults or insulation defects, so that the aim of 'Zhuge Liang' in advance is achieved. The expansion from 'Zhuge liang after the fact', to 'Zhuge liang in the fact' and 'Zhuge liang before the fact' is pioneering, fills the blank of diagnosis and monitoring of holographic electrical quantity faults or insulation defects of the reactor body, and promotes the diagnosis and monitoring of the faults or the insulation defects of the reactor to a new level.

Detailed Description

The reactor multi-scale fault diagnosis and monitoring system only needs a plurality of reactor steady-state or transient fault recording data of 2.1-4.1 second time intervals to calculate the 'variation of equivalent electrical parameters', 'distribution probability of the equivalent electrical parameter variation' and 'polarity effect of equivalent electrical parameter time-shift variation rate' of the reactor to generate a comprehensive abnormal rate matrix AR, and can implement multi-scale diagnosis and monitor faults or insulation defects of the reactor according to values of elements of the comprehensive abnormal rate matrix AR.

Case 1

The connection group of the Tianjin Shuanggang #3 transformer is YNyn0d11, and the diagnosis result of the invention is that for the steady state recording of 20210823: the maximum value of the first column value of the synthetic abnormal rate matrix AR reaches 1011, which is almost impossible; the data was diagnosed as severely abnormal. The diagnosis result is consistent with the diagnosis result (DGA + electrical test) fed back on site, the reactor is a super simplified transformer without middle and low voltage windings, and the diagnosis and monitoring results of the transformer can also indirectly verify the diagnosis and monitoring results of the reactor.

Case 2

Zhejiang Anlan No. 5 reactor, for 21/04/07/2020: 37-21: and 44, carrying out fault diagnosis on the Zhejiang Anlan No. 5 reactor according to five groups of transient fault recording data. The diagnosis results of the present invention: and 3 groups of the five groups of wave recording data exceed the threshold value, and the diagnosis result shows that the B column of the main reactor of the Anlan #5 reactor has overheating defects, the change of the electrical parameters is 2%, and no partial discharge or winding deformation exists. The diagnostic result is consistent with the field-fed diagnostic result (DGA).

Case 3

South-south China south-yang #3 reactor, for 04/07/21/2020: 37-21: the 6 sets of transient fault recording data of 44 were subjected to fault diagnosis of the #3 reactor. The diagnosis results of the present invention: the diagnosis results of 6 groups of wave recording data are the same, and the C column of the Nanyang reactor has smaller winding deformation defects; causing a change in the electrical parameter of 5%; without overheating. No partial discharge. The diagnosis result is consistent with the diagnosis result (electrical test) fed back on site.

Since 2018, the functions of the reactor multi-scale fault diagnosis and monitoring system are subjected to back-to-back examination and verification of Tianjin (1 transformer indirect verification) and 2 (power) reactors in Zhejiang and Henan on-site actual fault recording data: the verification result is 100% consistent with the field diagnosis result, and the accuracy, reliability and operability of the fault diagnosis and monitoring system are preliminarily verified.

Description of the drawings:

and (3) a picture of a reactor multi-scale fault diagnosis and monitoring system in the attached drawing.

Claims

1. The invention takes reactor steady state or transient state fault recording data as a signal source, generates a plurality of identification criteria for diagnosing and monitoring reactor faults or insulation defects, constructs a fault diagnosing and monitoring system consisting of hardware and software, and implements the diagnosis and monitoring of multi-scale reactor faults or insulation defects.

2. The invention takes the variable quantity of the equivalent electrical parameters (impedance Z and admittance Y) of the reactor as a criterion for fault diagnosis and monitoring, realizes the multi-scale diagnosis and monitoring of the fault or the insulation defect of the reactor from the angle of the variable quantity of the equivalent electrical parameters (impedance Z and admittance Y), and identifies the position and the severity of the fault or the insulation defect.

3. The method takes the distribution probability of the equivalent electrical parameter variation of the reactor in a certain period of time as a criterion for fault diagnosis and monitoring, realizes multi-scale diagnosis and monitoring of the fault or insulation defect of the reactor from the angle of the distribution probability of the equivalent electrical parameter variation, and identifies the position and the severity of the fault or insulation defect.

4. The invention takes the polar effect of the equivalent electrical parameter time-shifting change rate of the reactor as a criterion for fault diagnosis and monitoring, realizes multi-scale diagnosis and monitoring of faults or insulation defects of the reactor from the angle of the polar effect of the equivalent electrical parameter time-shifting change rate, and identifies the position and the severity of the faults or the insulation defects.

5. The invention constructs an Abnormal rate (Abnormal Ratio AR) matrix based on the 'electrical parameter variation amount', 'distribution probability of the electrical parameter variation amount' and 'polarity effect of time-shift variation rate of the electrical parameter',

the rows in the AR matrix respectively correspond to the fault category of the reactor, the core column where the fault or the insulation defect is located, the winding where the fault or the insulation defect is located, and the severity of the fault or the insulation defect; the columns respectively correspond to data reliability criteria of a signal source, total faults of the reactor, arc faults or insulation defects, partial discharge faults or insulation defects, winding deformation faults or defects, overheating faults or defects and turn-to-turn short circuit faults or defects. According to the values of all elements of the comprehensive abnormal rate matrix AR, multi-scale diagnosis and monitoring of faults or insulation defects of the reactor are implemented: (1) diagnosing whether the whole reactor has faults or insulation defects and the severity of the faults or the insulation defects; (2) diagnosing low, medium and high energy faults or insulation defects of the reactor and the positions of the faults or the insulation defects; (3) diagnosing a main reactor, an ABC core column and a neutral point small reactor where a fault or an insulation defect is located; (4) diagnosing the reactor with or without an arc fault, the position of the arc fault and the severity of the arc fault; (5) diagnosing whether the reactor has winding deformation, the position of the winding deformation and the severity of the winding deformation; (6) diagnosing whether the reactor winding has partial discharge, the position of the partial discharge and the severity of the partial discharge; (7) diagnosing reactor winding turns or notThe location of inter-short, turn-to-turn short and their severity; the diagnosis and monitoring of multi-scale faults or insulation defects within the range of 1% -100% of the energy change of the reactor are realized.

6. The reliability of signal source data is judged according to the value of the identification criterion based on the data reliability identification criterion (the first column of the comprehensive abnormal rate matrix AR) constructed by the electrical parameter variation, the distribution probability of the electrical parameter variation and the polarity effect of the time-shifting variation rate of the electrical parameter.

7. The reactor multi-scale fault diagnosis and monitoring system has universality, and is not only suitable for various oil-filled reactors, but also suitable for various dry-type reactors.