Comprehensive evaluation method of health status of substation equipment based on multi-dimensional parameters

Abstract

The present invention discloses a comprehensive evaluation method for the health status of substation equipment based on multi-dimensional parameters, including: first, conducting data exploration and judging data quality; second, combining historical data of the same working condition to predict the confidence of equipment failure and realize horizontal and vertical information linkage analysis; each parameter is given two indicators of degradation degree and variable weight for measurement; third, an improved AHP comprehensive evaluation method is used to construct a comprehensive scoring model for multi-dimensional parameters of substation equipment; finally, the health status of substation equipment is digitally described through membership functions to obtain the health status level of substation equipment, thereby forming an accurate evaluation of the health status of substation equipment during operation. The present invention realizes a comprehensive evaluation of the health status of substation equipment, so that the overall status of substation equipment can be grasped more clearly and the trend of status change can be more clearly defined, so as to find the signs of defects and failures early, guide operation and maintenance in a targeted manner, reduce the time for handling problems, and improve the safety and stability of substation equipment operation.

Description

Technical Field

[0001] The present invention relates to the field of substation equipment operating status diagnosis and substation equipment intelligent operation, maintenance and overhaul, and in particular to a comprehensive health status assessment method for substation equipment based on multi-dimensional parameter linkage analysis. Background Art

[0002] Substation equipment, especially transformers, plays a pivotal role in the power system, and its operational safety and reliability are directly related to the safety of the power system. For a long time, the domestic judgment of the health level and operating status of substation equipment has mainly relied on regular maintenance, which often leads to missed inspections and excessive maintenance, resulting in waste of materials and low efficiency, and also reduces the safety and reliability of normal operation of substations.

[0003] Existing research on substation status assessment mainly focuses on the application of neural networks, fuzzy comprehensive evaluation and other algorithms in the assessment system. However, with the continuous deepening of the digital transformation of substations, unmanned monitoring such as robots, high-definition video, infrared temperature measurement, and online monitoring devices have become the development trend of smart grids. The increase in the types of substation monitoring systems can monitor the operation status of substation equipment in an all-round way. However, while improving the monitoring efficiency, new problems have also emerged. The ability to extract substation equipment monitoring data is insufficient. Most of the existing assessment systems are still based on the application of electrical quantity data such as load current and voltage ratio. The multi-dimensional data of various new monitoring and inspection systems are not fully utilized. The data source is single. In the face of huge and complex data, there is a lack of screening of various indicators and evaluation of data quality, resulting in data redundancy; secondly, with the complexity of substation equipment monitoring methods, multi-dimensional substation equipment monitoring data is difficult to correlate, such as equipment temperature information, equipment defect images, etc. A single source of information is prone to misjudgment and early warning delays, so it is particularly urgent to open up data linkage between various monitoring systems.

[0004] The currently deployed substation equipment monitoring system is only a tool for collecting and displaying data at the station end. It lacks deep information mining and advanced applications, and cannot provide substantial reference value for substation equipment failure prevention and maintenance, resulting in information waste. Therefore, how to integrate the existing multi-dimensional massive data of substation equipment for linkage analysis and comprehensive evaluation of equipment status needs to be solved urgently. Summary of the invention

[0005] Purpose of the invention: The purpose of the present invention is to provide a comprehensive evaluation method for the health status of substation equipment based on multi-dimensional parameter linkage analysis. This method can achieve a comprehensive evaluation of the health status of substation equipment, making the overall status of substation equipment clearer and the status change trend clearer.

[0006] Technical solution: The present invention comprises the following steps:

[0007] (1) Conduct data exploration and judge data quality; screen and preprocess multi-dimensional parameter data through correlation analysis and indicator importance measurement, so as to build a substation equipment evaluation indicator system;

[0008] (2) Combine historical data of the same working condition to predict the confidence level of equipment failure and realize horizontal and vertical information linkage analysis; Combine association rules and historical occurrence rates to build deduction models for different characteristic parameters, and evaluate the initial deduction values ​​of each parameter defect at three levels: general, serious, and critical. The initial deduction values ​​are divided into a binary initial deduction model and a numerical initial deduction model according to the characteristics of the parameter data;

[0009] (3) Each parameter is measured by two indicators: degradation degree and variable weight. The initial deduction model of step (2) only considers the results under the degradation degree. Then, the initial deduction value of each parameter is assigned a variable weight coefficient to obtain the final deduction value. The sum of the defect parameter deductions of each component is calculated separately according to the components, and finally the component-level scores of the body, bushing, tap changer, cooler, and non-electrical protection are obtained;

[0010] (4) The improved AHP comprehensive evaluation method is used to construct a multi-dimensional parameter comprehensive scoring model for substation equipment. The hierarchical analysis algorithm is used to calculate the weight coefficient of each component in the comprehensive score, and the weight coefficient is calculated to finally obtain the comprehensive score of the substation equipment;

[0011] (5) The health status of the substation equipment is digitally described through the membership function to obtain the health status level of the substation equipment, thereby forming an accurate assessment of the health status of the substation equipment during operation. Early warning information is obtained based on the comprehensive assessment results and health status level to accurately guide maintenance.

[0012] Furthermore, the sources of the multi-dimensional parameter data in step (1) specifically include obtaining SCADA alarm signals, telemetry data, oil chromatography online monitoring, iron core online monitoring, operation sound monitoring and partial discharge monitoring online monitoring information, integrating characteristic signals generated by infrared thermal imaging and image recognition information of high-definition video joint patrol.

[0013] Furthermore, after the multi-dimensional parameter data is obtained through various monitoring systems deployed in the substation, after data quality analysis and data preprocessing, a substation equipment evaluation index system is formed.

[0014] Furthermore, the data quality analysis includes validity discrimination, completeness analysis and normative analysis, and the data preprocessing includes data cleaning, correlation analysis and indicator importance measurement.

[0015] Furthermore, the numerical parameter in step (2) adopts a half-step-down initial deduction model, and the binary state quantity adopts a categorical initial deduction model.

[0016] Furthermore, the variable weight coefficient weighting in step (3) refers to the weighting of the importance level of each parameter in the corresponding component. The variable weight coefficient is divided into four levels: I, II, III, and IV to determine the variable weight coefficient of each parameter. The overall score of the component shall not be less than 0 and shall be scored on a percentage basis. If the sum of the deduction values ​​is greater than 100, the total deduction value shall be deemed to be 100.

[0017] Furthermore, in the step (5), in view of the unclear correspondence between the final rating value of the transformer and the fault state, the membership degree and relative degradation degree evaluation results in fuzzy theory are utilized, and the membership function utilizes semi-trapezoidal and triangular distributions to establish a membership function corresponding to the relative degradation degree of the transformer and the fault state level.

[0018] Furthermore, assuming that the final score of the transformer is t, the relative degradation degree calculation formula is as follows:

[0019]

[0020] Furthermore, the health status level of the substation equipment in step (5) is divided into four levels: normal, caution, abnormal, and severe.

[0021] Furthermore, the membership function in step (5) is a function corresponding to the fuzzy boundary intervals of the four transformer status levels and the relative degradation degree, and the membership functions in the four statuses are respectively: according to The final health status of the transformer can be determined.

[0022] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: the present invention improves the one-sided evaluation of substation equipment by a single parameter and strengthens the effective use of intelligent monitoring data information; it realizes a comprehensive evaluation of the health status of substation equipment, making the overall status of substation equipment more clearly understood and the status change trend more clear, thereby early detection of defects and signs of failure, targeted guidance of operation and maintenance, reducing the time for handling problems, and improving the safety and stability of substation equipment operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flow chart of the present invention;

[0024] Figure 2 It is the AHP hierarchical structure diagram;

[0025] Figure 3 This is a descending semi-trapezoidal scoring model diagram;

[0026] Figure 4 It is the membership function model diagram. DETAILED DESCRIPTION

[0027] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0028] The present invention provides a comprehensive evaluation method for the health status of substation equipment based on multi-dimensional parameter linkage analysis. Figure 1 As shown in the figure, this method can make full use of the multi-dimensional data of substation equipment, deeply explore the correlation between various monitoring indicators, and realize the health status assessment and early warning of transformer operation on the basis of collecting all data. Digital empowerment can fully control the operation and maintenance personnel and equipment, provide targeted guidance on operation and maintenance, reduce the time for handling problems, and improve the safety and stability of substation equipment operation. Taking the health assessment and early warning of oil-immersed transformers as an example, it specifically includes the following steps:

[0029] (1) Establish the status system. The equipment parts and status quantities to be monitored of oil-immersed transformers mainly include:

[0030] a. Main body: main body appearance, operating sound, operating oil temperature, hydrogen content, acetylene content, total hydrocarbon content, partial discharge, core (clamp) grounding current, transformer overload, short-circuit current, etc.;

[0031] b. Casing: casing appearance, oil level gauge, lead wire and joint, pressure equalizing ring, end screen, etc.;

[0032] c. Tap changer: tap position, mechanism box, online oil filter, etc.;

[0033] d. Cooling system: appearance, fan, control box, etc.;

[0034] e. Non-electrical protection devices: gas relay, thermometer, pressure relief valve, oil flow relay, pressure sudden relay, etc.;

[0035] f. Oil conservator: oil level meter, appearance, respirator, oil level meter, appearance, etc. of the main oil conservator;

[0036] g. High-voltage switchgear: appearance, instruments, switch opening and closing positions, contacts, wires and cables, etc.

[0037] (2) Establishing association rules: By accessing the information in step A, using the historical data of the oil-immersed transformer operation (the information of each online monitoring system at the same time) to establish association rules, we get:

[0038] a. When multiple monitoring systems reflect the status of the same component, the x system infers that it may be fault A or fault B, and the y system infers that it may be fault B or fault C, then it is highly likely that fault B has occurred.

[0039] b. When a certain defect occurs, it is often accompanied by the occurrence of another defect. For example, when fault A occurs, fault D often occurs together. When fault A occurs, fault D is very likely to occur.

[0040] (3) Determine the variable weight coefficient of the state quantity. When the value of a state quantity deviates seriously from the normal value, it often means that the performance of a certain aspect of the transformer has seriously declined. However, when evaluating according to the constant weight coefficient, if its weight accounts for a small proportion of the overall value, the overall score of the transformer may still be at a normal level, and thus cannot accurately reflect the actual operating status of the transformer. Therefore, it is necessary to introduce a variable weight coefficient that changes due to the change of the state quantity;

[0041] The variable-weight state quantity defects are divided into three categories: general, serious, and critical. General defects refer to defects that are abnormal but will not cause accidents. Such defects are generally required to be recorded, and attention should be paid to the development of the defects. They should be repaired during power outages and planned for test repairs to eliminate the defects. Severe defects refer to those with a high degree of damage that seriously affects the effect of the components but can still operate. They should be handled as soon as possible. Critical defects refer to defects that must be handled urgently, which can easily cause component paralysis or unusability, and even cause larger accidents.

[0042] The variable weight coefficient is ω i (ω 一般 ,ω 严重 ,ω 危急 ) array, i refers to a monitoring system. According to the number of three types of state quantity defects contained in each monitoring system and the probability of occurrence of the three types of state quantity defects estimated based on historical data, the variable weight coefficient of the state quantity under each monitoring system is calculated as shown in formula (1):

[0043]

[0044] Where n1 is the number of normal cases in a monitoring system, n2 is the number of serious cases in a monitoring system, and n3 is the number of critical cases in a monitoring system. Find maxω from the constraint function 危急 , fixed maxω 危急 Find maxω 严重 , and finally calculate maxω 一般 .

[0045] (4) Establish the basis for judging the state quantity and scoring model under each monitoring system. According to the transformer diagnosis standards, specifications and expert database, the classification basis of each state quantity under each monitoring system is determined, and the numerical state quantity and binary state quantity are scored. The deduction value calculation model is used to complete the final scoring.

[0046] Numerical state quantities, such as total hydrocarbon value, upper oil temperature, etc., for the oil chromatography and infrared temperature measurement systems, the common descending semi-trapezoidal scoring model is used, such as Figure 3 As shown, the deduction value calculation is as follows:

[0047]

[0048] Substituting the variable weight coefficient of the corresponding monitoring system into the equation, we can obtain: d′1(y)=d1(v)ω i , v∈i.

[0049] (5) The binary state quantity has only two results: yes and no. For the visible light image system state quantity, considering that each state quantity can accurately represent the operating state of a certain aspect of the transformer, that is, it is very important for the overall health status of the transformer, its constant weight coefficient can be averaged;

[0050] Substituting the variable weight coefficient of the corresponding monitoring system, the final deduction value of each state quantity is as follows:

[0051] d′2(v)=d2(v)ω i ,v∈i (3)

[0052] The deduction value corresponding to each state quantity is calculated according to the deduction value model, including oil chromatography online monitoring data, infrared temperature measurement data, image recognition data, voiceprint monitoring data, core grounding monitoring data, partial discharge monitoring data and SCADA monitoring system data.

[0053] (6) Calculate the sum of the defect parameter deductions for each component separately, and finally obtain the component-level scores of the main body, bushing, tap changer, cooler, and non-electrical protection.

[0054] T(i)=100-di1-di2-…-din (4)

[0055] Among them, i is a transformer component among the body, bushing, tap changer, cooler, and non-electrical protection, and di1, di2, ..., din are the deduction points of each state quantity belonging to component i.

[0056] (7) Establish a hierarchical model, such as Figure 2 As shown in the figure, the AHP hierarchical analysis method is used to divide a problem into multiple levels, and each level is further divided into multiple factors. The relative importance of each factor is determined by mutual comparison, and then the subjective opinions of experts are combined to determine the weight coefficient of each factor.

[0057] (8) Construct the “target layer-first criterion layer” pairwise comparison matrix:

[0058] a. Methods for constructing comparison matrix

[0059] b. Construct the “target layer - first criterion layer” pairwise comparison matrix

[0060] c. Calculate the maximum eigenvalue of the matrix and its corresponding vector

[0061] d. Calculate the eigenvector corresponding to the first criterion layer λ

[0062] Construct a pairwise comparison matrix of "first criterion layer and second criterion layer":

[0063] a. Methods for constructing comparison matrix

[0064] b. Construct a pairwise comparison matrix of "first criterion layer and second criterion layer"

[0065] c. Calculate the maximum eigenvalue of the matrix and its corresponding vector

[0066] d. Calculate the eigenvector (weight vector) corresponding to the second criterion layer λ

[0067] (9) By adding the weights of the contained state quantities, the weight vector of the decision layer (7 monitoring systems) is obtained. Finally, the weight vector of the decision layer is obtained as follows: respectively.

[0068] (10) Based on the scores of the seven monitoring systems and their attached weights, the final transformer status score is obtained by weighted average:

[0069]

[0070] (11) Transformer status classification assessment: In view of the unclear correspondence between the final rating value of the transformer and the fault status, the membership degree and relative degradation degree in fuzzy theory are mainly used to process the assessment results.

[0071] The membership function represents the boundary of the fuzzy set, and the membership degree is determined by the membership function, and its range is any continuous value between 0 and 1. When the membership degree of an indicator is "0", it indicates that the indicator does not belong to this fuzzy set; when its membership degree is "1", it indicates that the indicator completely belongs to this fuzzy set; and when its membership value is between 0 and 1, it indicates the degree to which the indicator belongs to this fuzzy set.

[0072] For the final transformer score t, the relative degradation degree is:

[0073]

[0074] Since the shapes of the semi-trapezoidal and triangular distribution membership functions are relatively simple, and the results obtained are not much different from other types of membership functions, the semi-trapezoidal and triangular distributions are used to establish the membership function corresponding to the final rating value (relative degradation degree) of the transformer and the fault state level, such as Figure 4 As shown:

[0075] The membership functions corresponding to the fuzzy boundary interval of transformer status level and relative degradation degree are as follows:

[0076] Membership function in normal state:

[0077]

[0078] Membership function in general state:

[0079]

[0080] Membership function in severe state:

[0081]

[0082] Membership function in critical state:

[0083]

[0084] From the above, the evaluation result matrix of the transformer can be obtained as shown in Table 1:

[0085] Table 1 Transformer evaluation results matrix

[0086]

[0087] according to The final health status of the transformer can be determined. The transformer status information is scored on a 100-point scale, breaking the previous system of only two statuses: qualified or exceeded. This meets the requirements for more detailed grading of transformers for condition-based maintenance and realizes transformer grade assessment.

Claims

1. A comprehensive evaluation method for the health status of substation equipment based on multi-dimensional parameters, characterized in that: The steps include: (1) Conduct data exploration and judge data quality; screen and preprocess multi-dimensional parameter data through correlation analysis and indicator importance measurement, so as to build a substation equipment evaluation indicator system; (2) Combine historical data of the same working conditions to predict the confidence level of equipment failure and realize horizontal and vertical information linkage analysis; combine association rules and historical occurrence rates to construct deduction models for different characteristic parameters, and evaluate the initial deduction values ​​of each parameter defect at three levels: general, serious, and critical. The initial deduction values ​​are divided into a binary initial deduction model and a numerical initial deduction model according to the characteristics of the parameter data; (3) Each parameter is measured by two indicators: degradation degree and variable weight. The initial deduction model of step (2) only considers the results under the degradation degree. Then, the initial deduction value of each parameter is assigned a variable weight coefficient to obtain the final deduction value. The sum of the defect parameter deductions of each component is calculated separately, and finally the component-level scores of the body, bushing, tap changer, cooler, and non-electrical protection are obtained. (4) The improved AHP comprehensive evaluation method is used to construct a multi-dimensional parameter comprehensive scoring model for substation equipment. The hierarchical analysis algorithm is used to calculate the weight coefficient of each component in the comprehensive score. The weight coefficient is assigned and the comprehensive score of the substation equipment is finally obtained. (5) The health status of substation equipment is digitally described through membership functions to obtain the health status level of substation equipment, thereby forming an accurate assessment of the health status of substation equipment during operation. Early warning information is obtained based on the comprehensive assessment results and health status level to accurately guide maintenance; In the step (5), in view of the unclear correspondence between the final rating value of the transformer and the fault state, the membership degree and relative degradation degree evaluation results in fuzzy theory are used, and the membership function uses semi-trapezoidal and triangular distributions to establish a membership function corresponding to the relative degradation degree of the transformer and the fault state level.

2. The comprehensive evaluation method for health status of substation equipment based on multi-dimensional parameters according to claim 1 is characterized in that: The sources of the multi-dimensional parameter data in step (1) specifically include obtaining SCADA alarm signals, telemetry data, oil chromatography online monitoring, iron core online monitoring, operation sound monitoring and partial discharge monitoring online monitoring information, integrating characteristic signals generated by infrared thermal imaging and image recognition information of high-definition video joint patrol.

3. The comprehensive evaluation method for health status of substation equipment based on multi-dimensional parameters according to claim 2 is characterized in that: The multi-dimensional parameter data is obtained through various monitoring systems deployed in the substation, and then undergoes data quality analysis and data preprocessing to form a substation equipment evaluation index system.

4. The comprehensive evaluation method for health status of substation equipment based on multi-dimensional parameters according to claim 3 is characterized in that: The data quality analysis includes validity discrimination, completeness analysis and normative analysis, and the data preprocessing includes data cleaning, correlation analysis and indicator importance measurement.

5. The comprehensive evaluation method for health status of substation equipment based on multi-dimensional parameters according to claim 1 is characterized in that: The numerical parameters in step (2) adopt a half-step-down initial deduction model, and the binary state quantity adopts a categorical initial deduction model.

6. The comprehensive evaluation method for health status of substation equipment based on multi-dimensional parameters according to claim 1 is characterized in that: The variable weight coefficient weighting in step (3) refers to the weighting of the importance level of each parameter in the component to which it belongs. The variable weight coefficient is divided into four levels: I, II, III, and IV to determine the variable weight coefficient of each parameter. The overall score of the component shall not be less than 0 and shall be scored on a percentage basis. If the sum of the deduction values ​​is greater than 100, the total deduction value shall be deemed to be 100.

7. The comprehensive evaluation method for health status of substation equipment based on multi-dimensional parameters according to claim 1 is characterized in that: Assume that the final score of the transformer is t , the relative degradation calculation formula is as follows: 。 8. The comprehensive evaluation method for health status of substation equipment based on multi-dimensional parameters according to claim 1 is characterized in that: The health status level of the substation equipment in step (5) is divided into four levels: normal, caution, abnormal, and severe.

9. The comprehensive evaluation method for health status of substation equipment based on multi-dimensional parameters according to claim 1 is characterized in that: The membership function in step (5) is a function corresponding to the fuzzy boundary intervals of the four transformer status levels and the relative degradation degree. The membership functions for the four statuses are respectively: ,according to Determines the final health status of the transformer.

Images