Dynamic optimization and adaptive iteration transformer state evaluation method and device

Abstract

The invention discloses a dynamic optimization and adaptive iteration-based transformer state evaluation method and device. The method comprises the following steps of: constructing a multi-level equipment health evaluation index system of an index layer, a dimension layer, a component layer and an equipment layer; all key index attributes are extracted and integrated, indexes of the transformer evaluation system and all the index attributes are configured in a user-defined mode, and continuous optimization iteration of the self-adaptive evaluation index system is supported. According to index data characteristic classification, establishing an index score deduction model library of an'index layer 'covering full-amount indexes, and according to attributes configured by index parameters, automatically matching an optimal index score deduction model; constructing a'component-dimension layer 'scoring model; constructing an'equipment-component layer 'scoring model; and finally, constructing an'equipment layer 'model based on a comprehensive evaluation algorithm model of AHP-dynamic weighting fusion, and continuously and accurately analyzing and evaluating the operation health state of the transformer.

Description

Technical Field

[0001] The present invention belongs to the field of substation equipment management and intelligent operation, maintenance and overhaul of substation equipment, and specifically relates to a transformer state evaluation method and device with dynamic optimization and adaptive iteration. Background Art

[0002] Transformer is one of the important power transmission and transformation equipment in the power system. It plays a pivotal role in the power system. Its operation safety and reliability are directly related to the safety of the power system. For a long time, the judgment of the health level and operation status of transformers in China has mainly relied on regular maintenance. Missed inspection and over-maintenance often occur, resulting in waste of materials and low efficiency, and also reducing the safety and reliability of normal operation of substations.

[0003] At present, the status evaluation of transformers and other main substation equipment is mainly achieved by scoring according to the "Guidelines for Status Evaluation of Oil-immersed Transformers (Reactors)" (standard number DL / T 1685-2017) and other guideline documents; however, the data fusion standards centered on equipment health management are not unified, and the accuracy of the analysis algorithm model is not high. The structure and principle of the main equipment of the substation are complex, and its operating status is affected by many factors. It is impossible to determine the operating status by a single indicator, which greatly reduces the effectiveness and accuracy of the status evaluation method. At present, the unified health evaluation model of transformers is difficult to meet the characteristics of various transformers, and needs to be customized for each station, and the portability is very low.

[0004] Therefore, how to establish an evaluation model that is highly portable and can adapt to the characteristics of different transformers, effectively and comprehensively utilize and analyze multi-modal and multi-dimensional transformer status information, accurately and timely give the health status of the equipment, repair equipment abnormalities, and realize dynamic optimization and adaptive iteration of the transformer status evaluation model is a difficult problem that needs to be solved urgently. Summary of the invention

[0005] Purpose of the invention: In order to solve the problems in the background technology, the present invention proposes a transformer health status evaluation method and device with dynamic optimization and adaptive iteration.

[0006] Technical solution: In the first aspect, the present invention discloses a transformer state evaluation method with dynamic optimization and adaptive iteration, comprising:

[0007] Step 1: Collect the monitoring status of the transformer from the data source;

[0008] Step 2: Screen the multi-dimensional state quantity indicators in the monitoring state quantity, and construct a multi-level transformer evaluation index system of "indicator layer-dimensional layer-component layer-equipment layer"; the transformer evaluation index system is provided with a multi-layer scoring model, including:

[0009] The "indicator layer" scoring model is provided with an indicator deduction model library, which is used to automatically match the most appropriate indicator deduction model from the indicator deduction model library according to the indicator data characteristics of the state quantity indicator;

[0010] The “component-dimension level” scoring model is used to obtain the individual scores of each transformer component in each dimension;

[0011] The “equipment-component layer” scoring model is used to obtain the comprehensive score of each transformer component in all dimensions and the health status level of each component;

[0012] The “equipment level” scoring model is used to obtain the comprehensive score of the transformer status and the health status level of the transformer;

[0013] Step 3: According to the indicator attributes of the state quantity indicator, the transformer state is evaluated using the transformer evaluation indicator system.

[0014] Furthermore, the steps for constructing the "indicator layer" model are:

[0015] S1. Extract 6 indicator attributes of the state quantity indicator, and provide configuration options for each indicator attribute; the indicator attributes and the corresponding options include:

[0016] The components include the main body, bushing, tap changer, cooling system, and non-electrical protection;

[0017] The dimensions include main insulation of equipment, main conductive circuit and magnetic circuit, equipment function realization, equipment operating environment, and low-voltage DC auxiliary equipment;

[0018] Importance, which indicates the influence of each state quantity index in its corresponding component, including four levels: I, II, III, and IV, with the influence increasing in sequence;

[0019] Degradation level indicates the degree of degradation of the transformer, including degradation I, degradation II, degradation III, and degradation IV, with the degradation degree increasing in ascending order;

[0020] Data types, including telemetry, online monitoring, signals, change amounts, and change rates;

[0021] The type of limit violation includes upper limit violation and lower limit violation;

[0022] S2. The operation and maintenance personnel select the indicators that need to be substituted into the transformer evaluation system, and customize the indicator attribute options of all indicators;

[0023] S3. Establish an indicator deduction model library, set up indicator deduction model categories according to different indicator data characteristics, and then divide a indicator deduction model category into several indicator deduction model subcategories according to degradation conditions or change cycles;

[0024] S4. According to the indicator data characteristics of the state quantity indicator, the optimal indicator deduction model corresponding to the state quantity indicator is matched.

[0025] Furthermore, the indicator data characteristics include numerical type exceeding the upper limit, numerical type exceeding the lower limit, remote signal type, increase amount, decrease amount, growth rate, and decrease rate. The corresponding indicator deduction model categories include numerical type rising model, numerical type decreasing model, remote signal type model, increase amount model, decrease amount model, growth rate model, and decrease rate model.

[0026] Furthermore, a major category of an indicator deduction model is divided into several minor categories of indicator deduction models according to degradation conditions; the degradation conditions include any combination of one or more of the four degradation levels of degradation I, degradation II, degradation III, and degradation IV. Combined with 7 indicator data characteristics and 4 types of importance, the indicator deduction model library can cover 420 indicator minor categories.

[0027] Furthermore, for the state quantity indicators whose indicator data characteristics are growth, decline, growth rate, and decline rate, the change cycle of the indicator data includes weekly change / change rate, daily change rate / change rate, 4h change / change rate, and 2h change / change rate.

[0028] For weekly variation, daily variation, 4h variation and 2h variation, the calculation formula of the variation period is:

[0029] △C=|C i,2 -C i,1 |

[0030] Among them, △C is the weekly, daily, 4h and 2h changes, unit is μL / L; C i,2 The latest online data of the corresponding characteristic gas, in μL / L; C i,1 Corresponding characteristic gas reference value, unit: μL / L;

[0031] For the weekly change rate, daily change rate, 4h change rate and 2h change rate, the calculation formula for the change period is:

[0032] Latest online data - Periodic variation gas reference value | / Periodic variation gas reference value × 100%

[0033] Furthermore, the state quantity deduction value obtained by the "indicator layer" model is determined by the degradation level and importance of the state quantity indicator. No deduction is made when the state quantity indicator is normal. The specific state quantity deduction values ​​are as follows:

[0034]

[0035] Furthermore, the scoring mechanism of the "component-dimension layer" scoring model is to deduct points from each component according to the state quantity index of each component of the transformer in each dimension, and obtain the individual score of each component in each dimension. The calculation formula for the cumulative deduction is:

[0036]

[0037] Among them, m represents the component, s represents the dimension, x1,x2,…,x n represents all state indicators of dimension s under component m, t1, t2, …, t n Represents state quantity index x1, x2,…, x n The deduction value.

[0038] Furthermore, the scoring mechanism of the "equipment-component layer" scoring model is to accumulate the state quantity indicators of each transformer component in all dimensions and deduct points from each component to obtain a comprehensive score of each component in all dimensions. The calculation formula is:

[0039]

[0040] Among them, x 11 ,...,x 1n ,x 21 ,...,x 2n ,x 31 ,...,x 3n ,x 41 ,...,x 4n ,x 51 ,...,x 5n Represents all state indicators belonging to component m in all dimensions, Indicates the deduction value of all state quantity indicators in all dimensions of component m.

[0041] Furthermore, the scoring mechanism of the "equipment layer" scoring model is a comprehensive evaluation algorithm model based on AHP-dynamic weighting fusion, which calculates the weight coefficient of each component in the comprehensive score, assigns weights to the scores of each component, and evaluates the status of the transformer based on the indicator attributes and the results output by other hierarchical scoring models to obtain the transformer status score and the transformer health status level.

[0042] Furthermore, referring to the State Grid guidelines, the health status of the matching components includes normal status, caution status, abnormal status and severe status. The rating standards for the health status of each component are as follows:

[0043]

[0044] Furthermore, for matching the transformer health status level, the overall evaluation of the transformer should integrate the evaluation results of its components: when all components are evaluated as normal, the overall evaluation is normal; when the status of any component is a warning state, an abnormal state or a serious state, the overall evaluation should be the most serious state.

[0045] Furthermore, the data sources include professional inspection data, experimental data, daily inspection data, and system data.

[0046] In a second aspect, the present invention further discloses a transformer health status evaluation device with dynamic optimization and adaptive iteration, comprising:

[0047] A collection module, used for collecting monitoring status quantities of the transformer from a data source;

[0048] A construction module is used to screen the multi-dimensional state quantity indicators in the monitoring state quantity and construct a multi-level transformer evaluation indicator system of "indicator layer-dimensional layer-component layer-equipment layer"; a multi-layer scoring model is set in the transformer evaluation indicator system, including an "indicator layer" scoring model, a "component-dimensional layer" scoring model, a "equipment-component layer" scoring model, and an "equipment layer" scoring model;

[0049] The evaluation module is used to evaluate the transformer state according to the indicator attributes of the state quantity indicator and using the transformer evaluation indicator system.

[0050] Furthermore, the construction module is specifically used to: screen out 6 indicator attributes, provide configuration options for each indicator attribute, and the operation and maintenance personnel select the indicators that need to be substituted into the transformer evaluation system and customize the configuration;

[0051] The indicator attributes and the corresponding options provided include:

[0052] The components include the main body, bushing, tap changer, cooling system, and non-electrical protection;

[0053] The dimensions include main insulation of equipment, main conductive circuit and magnetic circuit, equipment function realization, equipment operating environment, and low-voltage DC auxiliary equipment;

[0054] Importance, which indicates the influence of each state quantity index in its corresponding component, including four levels: I, II, III, and IV, with the influence increasing in sequence;

[0055] Degradation level indicates the degree of degradation of the transformer, including degradation I, degradation II, degradation III, and degradation IV, with the degradation degree increasing in ascending order;

[0056] Data types, including telemetry, online monitoring, signals, change amounts, and change rates;

[0057] The type of limit violation includes upper limit violation and lower limit violation;

[0058] The "indicator layer" model is provided with an indicator deduction model library, and the indicator deduction model categories are set according to the characteristics of different indicator data, and then one indicator deduction model category is divided into several indicator deduction model subcategories according to the degradation situation or the change cycle;.

[0059] Furthermore, the indicator data characteristics include numerical type exceeding the upper limit, numerical type exceeding the lower limit, remote signal type, increase amount, decrease amount, growth rate, and decrease rate. The corresponding indicator deduction model categories include numerical type rising model, numerical type decreasing model, remote signal type model, increase amount model, decrease amount model, growth rate model, and decrease rate model.

[0060] Furthermore, a major category of an indicator deduction model is divided into several minor categories of indicator deduction models according to degradation conditions; the degradation conditions include any combination of one or more of the four degradation levels of degradation I, degradation II, degradation III, and degradation IV. Combined with 7 indicator data characteristics and 4 types of importance, the indicator deduction model library can cover 420 indicator minor categories.

[0061] Furthermore, for the state quantity indicators whose indicator data characteristics are growth, decline, growth rate, and decline rate, the change cycle of the indicator data includes weekly change / change rate, daily change rate / change rate, 4h change / change rate, and 2h change / change rate.

[0062] For weekly variation, daily variation, 4h variation and 2h variation, the calculation formula of the variation period is:

[0063] △C=|C i,2 -C i,1 |

[0064] Among them, △C is the weekly, daily, 4h and 2h changes, unit is μL / L; C i,2 The latest online data of the corresponding characteristic gas, in μL / L; C i,1 Corresponding characteristic gas reference value, unit: μL / L;

[0065] For the weekly change rate, daily change rate, 4h change rate and 2h change rate, the calculation formula for the change period is:

[0066] |Latest online data-cycle variation gas reference value| / cycle variation gas reference value×100%

[0067] Furthermore, the state quantity deduction value obtained by the "indicator layer" model is jointly determined by the degradation level and importance of the state quantity indicator, and no deduction is made when the state quantity indicator is normal.

[0068] Furthermore, the scoring mechanism of the "component-dimension layer" scoring model in the construction module is to deduct points from each component according to the state quantity index of each component of the transformer in each dimension, and obtain the individual score of each component in each dimension. The calculation formula for the cumulative deduction is:

[0069]

[0070] Among them, m represents the component, s represents the dimension, x1,x2,…,x n represents all state indicators of dimension s under component m, t1, t2, …, t n Represents state quantity index x1, x2,…, x n The deduction value.

[0071] Furthermore, the scoring mechanism of the "equipment-component layer" scoring model in the construction module is to accumulate the state quantity indicators of each transformer component in all dimensions and deduct points from each component to obtain a comprehensive score of each component in all dimensions. The calculation formula is:

[0072]

[0073] Among them, x 11 ,...,x 1n ,x 21 ,...,x 2n ,x 31 ,...,x 3n ,x 41 ,...,x 4n ,x 51 ,...,x 5n Represents all state indicators belonging to component m in all dimensions, Indicates the deduction value of all state quantity indicators in all dimensions of component m.

[0074] Furthermore, the "equipment layer" scoring model in the construction module is based on the comprehensive evaluation algorithm model of AHP-dynamic weighting fusion, calculates the weight coefficient of each component in the comprehensive score, assigns weights to the scores of each component, and evaluates the status of the transformer based on the indicator attributes and the results output by other hierarchical scoring models to obtain the transformer status score and the transformer health status level.

[0075] Furthermore, the “equipment layer” scoring model refers to the State Grid guidelines to match the health status of components, including normal status, caution status, abnormal status and severe status.

[0076] Furthermore, for matching the transformer health status level, the overall evaluation of the transformer should integrate the evaluation results of its components: when all components are evaluated as normal, the overall evaluation is normal; when the status of any component is a warning state, an abnormal state or a serious state, the overall evaluation should be the most serious state.

[0077] Furthermore, the data sources of the acquisition module include professional inspection data, experimental data, daily inspection data, and system data.

[0078] In a third aspect, the present invention discloses a computer-readable storage medium having a computer program stored thereon, which implements the steps of the aforementioned method when executed by a processor.

[0079] Beneficial effects:

[0080] The method of the present invention has significant innovative advantages. First, through this method, the evaluation index attributes are customized and the optimal deduction model is adaptively matched. It can be flexibly applied to the individual characteristics and key requirements of each transformer, and an adaptive health status evaluation model can be intelligently generated, which has strong portability.

[0081] Second, the comprehensive utilization and analysis of multi-modal and multi-dimensional transformer status information can accurately analyze and evaluate the operating health status of the transformer under the influence of many factors.

[0082] Third, through the comprehensive evaluation algorithm model of AHP-dynamic empowerment fusion, the indicator set that needs to be substituted into the evaluation system is customized and the weights are dynamically adjusted. The transformer state evaluation model adapts to the subsequent increase, deletion, and modification of the access state quantity, realizes dynamic optimization and adaptive iteration of the evaluation model, and reduces the manual maintenance cost of continuous iteration of the algorithm during engineering application. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 A flowchart of a transformer state evaluation method of dynamic optimization and adaptive iteration according to the present invention;

[0084] Figure 2 A flowchart for customizing the attributes of each indicator of a transformer according to the present invention;

[0085] Figure 3 Configure flow charts for degradation levels of telemetry and online monitoring indicators;

[0086] Figure 4 Configure flow charts for degradation levels of change amount and change rate indicators. DETAILED DESCRIPTION

[0087] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments. Figure 1Shown is a flow chart of a specific implementation of a transformer state evaluation method with dynamic optimization and adaptive iteration according to the present invention.

[0088] In this embodiment, electrical quantities such as voltage, current, load, power, etc. of the transformer, operating oil temperature, winding temperature, oil level, etc., online monitoring data such as oil chromatography, core grounding current, partial discharge anomaly, sound print anomaly, etc., visible light image defects such as oil leakage, foreign matter, damage, meter readings, infrared temperature measurement features of boxes, joints, etc., light gas alarm, heavy gas alarm, pressure relief valve alarm and other non-electrical quantity alarm signals are collected from professional inspection data, experimental data, daily inspection data, system data, etc., to construct a multi-dimensional transformer status indicator.

[0089] In this embodiment, the multi-dimensional transformer state quantity indicators are also preprocessed. First, data exploration is performed to judge the data quality. Then, the multi-dimensional indicators are screened through correlation analysis and indicator importance measurement. Finally, the data is preprocessed by singular value elimination and missing value interpolation methods.

[0090] According to the State Grid standards, the transformer is divided into five components: body, bushing, tap changer, cooling system, and non-electrical protection. Each component is further divided into five dimensions: main insulation of the equipment, main conductive circuit and magnetic circuit, equipment function realization, equipment operating environment, and low-voltage DC auxiliary equipment. The collected monitoring indicators are matched to the dimensions of the corresponding components.

[0091] Optionally, in this embodiment, the number and types of dimensions are redefined according to individual characteristics such as transformer grade, operating status, and key focus, or the dimension layer is removed when no dimension is distinguished.

[0092] A multi-level transformer evaluation system is constructed, wherein the multiple levels are "indicator layer-dimensional layer-component layer-equipment layer"; scoring models of different levels are constructed in the multi-level evaluation system, including an "indicator layer" scoring model, a "component-dimensional layer" scoring model, a "equipment-component layer" scoring model, and an "equipment layer" scoring model.

[0093] Construct an "indicator layer" scoring model. First, extract the status attributes of transformer status evaluation into the components, dimensions, importance, data type, degradation level, and over-limit type. Set the options provided by the status attributes according to the specific situation of the equipment. Figure 2 Shown is a flow chart of customizing the attributes of each indicator of the transformer according to the present invention.

[0094] Importance refers to the influence of each indicator in its corresponding component, which is divided into four levels: I, II, III, and IV.

[0095] Data types include telemetry, online monitoring, signal, change amount, and change rate.

[0096] The degradation grade refers to the degree of degradation of each indicator, which is divided into four grades: degradation I, degradation II, degradation III, and degradation IV.

[0097] The over-limit type is divided into two categories: over-upper limit and over-lower limit. Over-upper limit means that the larger the value, the more serious the degradation, such as C2H2 gas exceeding the standard, and over-lower limit means that the smaller the value, the more serious the degradation, such as SF6 gas pressure leakage.

[0098] The operation and maintenance personnel customize the selection of evaluation indicators that need to be substituted into the transformer evaluation system, and customize the configuration in the status attribute options of each evaluation indicator. The adaptive matching of the optimal model is achieved through the customized configuration of evaluation indicators.

[0099] In this embodiment, the C2H2 gas content indicator attributes are set to: main body, equipment main insulation, measurement, importance level III, over-limit, four-level degradation level, and the degradation level is configured as degradation I, threshold 0.5; degradation II, threshold 1; degradation III, threshold 3, degradation IV, threshold 5, such as Figure 3 As shown. The C2H2 gas change indicator attributes are set to: main body, equipment main insulation, measurement, importance level III, over-limit, four-level degradation level, degradation level configuration is degradation I, weekly increment threshold 0.3; degradation II, weekly increment threshold 0.6; degradation III, weekly increment threshold 1.2, 4h increment threshold 1; degradation IV, weekly increment threshold 2, 2h increment threshold 1.5, as shown. Figure 4 As shown in the figure, the transformer status evaluation model is allowed to flexibly adapt to the continuous iterative upgrade of the subsequent access status quantity addition, deletion and modification in the station, and realize the self-upgrade of the evaluation index system.

[0100] The deduction mechanism of the “indicator layer” scoring model is determined by the indicator degradation degree and importance, that is, the state quantity deduction value is equal to the degradation deduction value of the state quantity multiplied by the importance coefficient, as shown in Table 1. No deduction is made when the state quantity is normal.

[0101] Table 1 Deduction value of state quantity index

[0102]

[0103] The evaluation indicators are classified according to the characteristics of the indicator data, and an indicator deduction model library is established. The indicator data characteristics include the upper limit of the numerical type, the lower limit of the numerical type, the remote signal type, the increase, the decrease, the growth rate, and the decrease rate. The indicator deduction model library is provided with a large category of indicator deduction models corresponding to the indicator data characteristics, including a numerical type increase model, a numerical type decrease model, a remote signal type model, an increase model, a decrease model, a growth rate model, and a decrease rate model.

[0104] The optimal indicator deduction model is matched according to the characteristics of the indicator data. Each major category is further divided into multiple subcategories according to the combination of degradation conditions and change cycles, and an indicator deduction model library covering all indicators is established. According to the different limit levels of each indicator, the degree of degradation may be set to 1, 2, 3, 4, and may be any combination of degradation I, degradation II, degradation III, and degradation IV. n degradation levels are applicable In this case, there is

[0105]

[0106] A total of 420 indicator subcategories can be covered in the indicator deduction model library.

[0107] For indicators whose data characteristics are increase, decrease, growth rate, and decrease rate, the change cycle of indicator data includes weekly change / change rate, daily change rate / change rate, 4h change / change rate, and 2h change / change rate.

[0108] The weekly change, daily change, 4h change and 2h change are calculated in the following way:

[0109] △C=|C i,2 -C i,1 |

[0110] Among them, △C is the weekly, daily, 4h and 2h changes, unit is μL / L; C i,2 The latest online data of the corresponding characteristic gas, in μL / L; C i,1 Corresponding characteristic gas reference value, unit: μL / L.

[0111] For the weekly change rate, daily change rate, 4h change rate and 2h change rate, the same reference value as the periodic change amount is used, and the calculation formula for the change period is:

[0112] |Latest online data-cycle variation gas reference value| / cycle variation gas reference value×100%

[0113] Calculation of reference value of weekly variation C i,1 Take the arithmetic mean of the online data between 336 hours (not included) and 168 hours (inclusive) before the timestamp hour value of this data. Calculation of daily variation reference value C i,1 Take the arithmetic mean of the online data between 48 hours (not included) and 24 hours (inclusive) before the timestamp hour value of this data.

[0114] Calculation of reference value of change every 4 hours and every 2 hours: When the acquisition cycle is 4 hours, C i,1 Take the arithmetic mean of the 4 online data 4 hours ago; when the data collection cycle is 2 hours, C i,1Take the arithmetic mean of the 4 online data 2 hours ago, and also calculate the change △C every 4 hours 4h When the data collection cycle is 1 hour, the 4-hour change is the arithmetic mean of the 4 online data 4 hours ago; the 2-hour change is the arithmetic mean of the 4 online data 2 hours ago.

[0115] In this embodiment, the 4-hour growth and weekly growth of the C2H2 content index are calculated. The latest data of the C2H2 content is 0.7. The 4 online data of the C2H2 content 4 hours ago are 0.32, 0.41, 0.52, and 0.6. The 4-hour growth is:

[0116] △C=0.7-(0.32+0.41+0.52+0.6)÷4=0.2375

[0117] The weekly growth rate of C2H2 content is calculated as:

[0118] △C=0.7-(0.02+0.06+0.12+0.07+0.06+0.03+0.09+0.01+0.07

[0119] +0.13+0.16+0.15+0.23+0.21+0.19+0.2+0.21+0.24+0.15+0.18

[0120] +0.16+0.17+0.13+0.11+0.11+0.09+0.12+0.07+0.07+0.08+0.09

[0121] +0.1+0.12+0.11+0.14+0.18+0.19+0.2+0.21+0.21+0.23+0.22)÷42≈0.5645

[0122] According to the attributes of the indicator parameter configuration, the optimal indicator deduction model is automatically matched. In this embodiment, the C2H2 gas content is matched to the numerical rising model, the C2H2 gas change is matched to the growth model, and the C2H2 gas change rate is matched to the growth rate model.

[0123] According to the state quantity indicator attribute setting, the C2H2 gas content is 0.7, which belongs to the degradation degree between I and II, and the importance level is III. The deduction of the C2H2 gas content indicator is calculated using the numerical ascending model:

[0124] t C2H2气体含量 =(0.7-0.5)÷(1-0.5)×(3×4-3×2)+(3×2)=8.4

[0125] According to the state quantity indicator attribute setting, the C2H2 gas growth is 0.5645, which belongs to the degradation degree I and the importance level III. The deduction of the C2H2 gas change indicator is calculated using the growth rate model:

[0126] t C2H2增长量 =(0.5645-0.3)÷(0.6-0.3)×(3×4-3×2)+(3×2)=11.29

[0127] Construct a "component-dimension layer" scoring model. Divide into five components: body, bushing, tap changer, cooling system, and non-electrical protection, each component corresponds to a dimension. Accumulate the parameter deductions of each component in the five dimensions to obtain the scores of the five components in the five dimensions:

[0128] T m,s =100-(t x1 +t x2 +...+t xn )

[0129] Among them, m represents the component, s represents the dimension, x1,x2,…,x n represents all state indicators of dimension s under component m, t1, t2, …, t n Represents state quantity index x1, x2,…, x n If a component does not have a matching indicator attribute on a certain dimension, no points will be deducted.

[0130] Construct an "equipment-component layer" scoring model. The scores of each transformer component are deducted in five dimensions, and finally the comprehensive scores of the five components, namely, the transformer body, bushing, tap changer, cooling system, and non-electrical protection, are obtained:

[0131]

[0132] Among them, x 11 ,...,x 1n ,x 21 ,...,x 2n ,x 31 ,...,x 3n ,x 41 ,...,x 4n ,x 51 ,...,x 5n Represents all state indicators belonging to component m in all dimensions, Indicates the deduction value of all state quantity indicators in all dimensions of component m.

[0133] For example, the indicator deduction values ​​of the transformer body in five dimensions are: The deduction values ​​for the remaining indicators are all 0, and the final score calculation formula is as follows:

[0134] T 本体 =100-(8.4+11.29)=80.31 points

[0135] According to the matching component health status rating standard table, the transformer body is judged as: attention status.

[0136] Refer to the State Grid guidelines to match the health status levels of components. The health status rating standards for each component are shown in Table 2.

[0137] Table 2 Component health status rating standards

[0138]

[0139] The other four components, namely bushing, tap changer, cooler and non-electrical protection, are calculated, scored and component graded according to the above method.

[0140] Based on the comprehensive evaluation algorithm model integrated with "AHP-dynamic weighting", the AHP comprehensive evaluation algorithm improved by the index degradation coverage vector is used to intelligently dynamically calculate the weight ratio of each component's health to the equipment's health under different equipment failure conditions. The component weight is dynamically calculated based on the number of indicators of each important level of the component and the degradation level of the indicator.

[0141] In this embodiment, the scores of each component are:

[0142] T 本体 =80.31, T 套管 =100, T 冷却系统 =100, T 分接开关 =100, T 非电量保护 =100

[0143] The number of indicators covered by the five components at each importance level is as follows:

[0144] T 本体 =(9,7,5,8), T 套管 =(2,2,3,1), T 冷却系统 =(1,3,2,1), T 分接开关 =(3,0,2,1), T 非电量保护 =(0,0,5,2)

[0145] The number of indicators covering each degradation level of the 5 components is:

[0146] T 本体 =(0,1,1,0), T 套管 =(0,0,0,0), T 冷却系统 =(0,1,0,0), T分接开关 =(0,0,0,0), T 非电量保护 =(0,0,0,0) The weight coefficients of the five components in the comprehensive score are calculated as follows:

[0147] W (本体,套管,冷却系统,分接开关,非电量保护) =(0.53, 0.1, 0.17, 0.08, 0.11)

[0148] Finally, the transformer comprehensive score is:

[0149] T = 80.31*0.53+100*0.1+100*0.17+100*0.08+100*0.11 = 89.56 points

[0150] The overall evaluation of the transformer should integrate the evaluation results of its components. When all components are evaluated as normal, the overall evaluation is normal; when any component is in a caution state, abnormal state or severe state, the overall evaluation should be the most severe state. In this embodiment, the main body component is judged to be in a caution state, and the other components are all in a normal state, then the overall evaluation of the transformer is the most severe component state, that is, the caution state.

[0151] After obtaining the final transformer score, the health status level of the transformer is matched through fuzzy comprehensive evaluation to form an accurate evaluation of the health status of the transformer, and corresponding measures are recommended according to the transformer status level:

[0152] Normal state: normal operation;

[0153] Attention status: Still running, strengthen monitoring during operation;

[0154] Abnormal status: monitor operation and arrange power outage for maintenance in a timely manner;

[0155] Serious condition: Arrange power outage and maintenance as soon as possible.

[0156] In this embodiment, the corresponding measures recommended are finally pushed according to the transformer attention status level: the change trend of the single state quantity is developing towards the direction of approaching the standard limit, but has not exceeded the standard limit, and it can still continue to operate, and the monitoring during operation should be strengthened.

[0157] The method of the present invention is flexibly applicable to the individual characteristics and key requirements of each substation, customizes the optimal evaluation system, and intelligently generates an adaptive health status evaluation model. The dynamic optimization mechanism realizes the flexible optimization of the evaluation index layer model, dynamically adjusts the weights, and adaptively iterates the evaluation model to ensure continuous and accurate analysis and evaluation of the operating health status of the transformer. It supports the self-upgrade of the evaluation system and model, greatly enhances the portability of inter-station deployment, and eliminates the difficulty of manual maintenance of the algorithm of the transformer health evaluation system within the station.

[0158] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A transformer state evaluation method with dynamic optimization and adaptive iteration, characterized in that: include: Step 1: Collect the monitoring status of the transformer from the data source; Step 2: Screen the multi-dimensional state quantity indicators in the monitoring state quantity, and construct a multi-level transformer evaluation index system of "indicator layer-dimensional layer-component layer-equipment layer"; the transformer evaluation index system is provided with a multi-layer scoring model, including: The "indicator layer" scoring model is provided with an indicator deduction model library, which is used to automatically match the most appropriate indicator deduction model from the indicator deduction model library according to the indicator data characteristics of the state quantity indicator; "Component-dimension level" scoring model, which is used to obtain individual scores for each transformer component in each dimension; The "equipment-component level" scoring model is used to obtain the comprehensive score of each transformer component in all dimensions and the health status level of each component; "Equipment level" scoring model, used to obtain the comprehensive score of transformer status and the health status level of the transformer; Step 3: According to the indicator attributes of the state quantity indicator, the transformer state is evaluated using the transformer evaluation indicator system.

2. The transformer status evaluation method according to claim 1, characterized in that: The construction steps of the "indicator layer" model are: S1. Extract 6 indicator attributes of the state quantity indicator, and provide configuration options for each indicator attribute; the indicator attributes and the corresponding options include: The components include the main body, bushing, tap changer, cooling system, and non-electrical protection; The dimensions include main insulation of equipment, main conductive circuit and magnetic circuit, equipment function realization, equipment operating environment, and low-voltage DC auxiliary equipment; Importance, which indicates the influence of each state quantity index in its corresponding component, including four levels: I, II, III, and IV, with the influence increasing in sequence; Degradation level indicates the degree of degradation of the transformer, including degradation I, degradation II, degradation III, and degradation IV, with the degradation degree increasing in ascending order; Data types, including telemetry, online monitoring, signals, change amounts, and change rates; The type of limit violation includes upper limit violation and lower limit violation; S2. The operation and maintenance personnel select the indicators that need to be substituted into the transformer evaluation system, and customize the indicator attribute options of all indicators; S3. Establish an indicator deduction model library, set up indicator deduction model categories according to different indicator data characteristics, and then divide a indicator deduction model category into several indicator deduction model subcategories according to degradation conditions or change cycles; S4. According to the indicator data characteristics of the state quantity indicator, the optimal indicator deduction model corresponding to the state quantity indicator is matched.

3. The transformer status evaluation method according to claim 2, characterized in that: The indicator data characteristics include the upper limit of the numerical type, the lower limit of the numerical type, the remote signal type, the increase, the decrease, the growth rate, and the decrease rate. The corresponding indicator deduction model categories include the numerical type increase model, the numerical type decrease model, the remote signal type model, the increase model, the decrease model, the growth rate model, and the decrease rate model.

4. The transformer status evaluation method according to claim 3, characterized in that: The indicator deduction model category is divided into several indicator deduction model subcategories according to the degradation conditions; the degradation conditions include any combination of one or more of the four degradation levels of degradation I, degradation II, degradation III, and degradation IV. Combined with 7 indicator data characteristics and 4 types of importance, the indicator deduction model library can cover 420 indicator subcategories.

5. The transformer status evaluation method according to claim 4, characterized in that: For state indicators whose data characteristics are growth, decline, growth rate, and decline rate, the change cycles of indicator data include weekly change / change rate, daily change rate / change rate, 4h change / change rate, and 2h change / change rate. For weekly variation, daily variation, 4h variation and 2h variation, the calculation formula of the variation period is: △C=|C i,2 -C i,1 Among them, △C is the weekly, daily, 4h and 2h changes, unit is μL / L; C i,2 The latest online data of the corresponding characteristic gas, in μL / L; C i,1 Corresponding characteristic gas reference value, unit: μL / L; For the weekly change rate, daily change rate, 4h change rate and 2h change rate, the calculation formula for the change period is: |Latest online data-cycle variation gas reference value| / cycle variation gas reference value×100%.

6. The transformer status evaluation method according to claim 5, characterized in that: The state quantity deduction value obtained by the "indicator layer" model is determined by the degradation level and importance of the state quantity indicator. No deduction is made when the state quantity indicator is normal.

7. The transformer status evaluation method according to claim 1, characterized in that: The scoring mechanism of the "component-dimension layer" scoring model is to deduct points from each component according to the state quantity index of each component of the transformer in each dimension, and obtain the individual score of each component in each dimension. The calculation formula for the cumulative deduction is: Among them, m represents the component, s represents the dimension, x1,x2,…,x n represents all state indicators of dimension s under component m, t1, t2, …, t n Represents state quantity index x1, x2,…, x n The deduction value.

8. The transformer status evaluation method according to claim 1, characterized in that: The scoring mechanism of the "equipment-component layer" scoring model is to accumulate the state quantity indicators of each transformer component in all dimensions and deduct points from each component to obtain a comprehensive score of each component in all dimensions. The calculation formula is: Among them, x 11 ,...,x 1n ,x 21 ,...,x 2n ,x 31 ,...,x 3n ,x 41 ,...,x 4n ,x 51 ,...,x 5n Represents all state indicators belonging to component m in all dimensions, Indicates the deduction value of all state quantity indicators in all dimensions of component m.

9. The transformer status evaluation method according to claim 1, characterized in that: The scoring mechanism of the "equipment layer" scoring model is a comprehensive evaluation algorithm model based on AHP-dynamic weighting fusion, which calculates the weight coefficient of each component in the comprehensive score, assigns weights to the scores of each component, and evaluates the status of the transformer based on the transformer status score and the health status level of the transformer obtained according to the indicator attributes and the results output by other hierarchical scoring models.

10. The transformer status evaluation method according to claim 9, characterized in that: Refer to the State Grid guidelines to match the health status of components, including normal status, caution status, abnormal status and critical status.

11. The transformer status evaluation method according to claim 10, characterized in that: The matching transformer health status level, the overall evaluation of the transformer should integrate the evaluation results of its components: when all components are evaluated as normal, the overall evaluation is normal; when the status of any component is warning status, abnormal status or serious status, the overall evaluation should be the most serious status among them.

12. The transformer status evaluation method according to claim 1, characterized in that: The data sources include professional inspection data, experimental data, daily inspection data, and system data.

13. A transformer health status evaluation device with dynamic optimization and adaptive iteration, characterized in that: include: A collection module, used for collecting monitoring status quantities of the transformer from a data source; A construction module is used to screen the multi-dimensional state quantity indicators in the monitoring state quantity and construct a multi-level transformer evaluation indicator system of "indicator layer-dimensional layer-component layer-equipment layer"; a multi-layer scoring model is set in the transformer evaluation indicator system, including an "indicator layer" scoring model, a "component-dimensional layer" scoring model, a "equipment-component layer" scoring model, and an "equipment layer" scoring model; The evaluation module is used to evaluate the transformer state according to the indicator attributes of the state quantity indicator and using the transformer evaluation indicator system.

14. The transformer state evaluation device according to claim 13, characterized in that: The construction module is specifically used to: screen out 6 indicator attributes, provide configuration options for each indicator attribute, and the operation and maintenance personnel select the indicators that need to be substituted into the transformer evaluation system and customize the configuration; The indicator attributes and the corresponding options provided include: The components include the main body, bushing, tap changer, cooling system, and non-electrical protection; The dimensions include main insulation of equipment, main conductive circuit and magnetic circuit, equipment function realization, equipment operating environment, and low-voltage DC auxiliary equipment; Importance, which indicates the influence of each state quantity index in its corresponding component, including four levels: I, II, III, and IV, with the influence increasing in sequence; Degradation level indicates the degree of degradation of the transformer, including degradation I, degradation II, degradation III, and degradation IV, with the degradation degree increasing in ascending order; Data types, including telemetry, online monitoring, signals, change amounts, and change rates; The type of limit violation includes upper limit violation and lower limit violation; The "indicator layer" model is provided with an indicator deduction model library, and the indicator deduction model categories are set according to the characteristics of different indicator data, and then one indicator deduction model category is divided into several indicator deduction model subcategories according to the degradation situation or change cycle.

15. The transformer state evaluation device according to claim 14, characterized in that: The indicator data characteristics include the upper limit of the numerical type, the lower limit of the numerical type, the remote signal type, the increase, the decrease, the growth rate, and the decrease rate. The corresponding indicator deduction model categories include the numerical type increase model, the numerical type decrease model, the remote signal type model, the increase model, the decrease model, the growth rate model, and the decrease rate model.

16. The transformer state evaluation device according to claim 15, characterized in that: The indicator deduction model category is divided into several indicator deduction model subcategories according to the degradation conditions; the degradation conditions include any combination of one or more of the four degradation levels of degradation I, degradation II, degradation III, and degradation IV. Combined with 7 indicator data characteristics and 4 types of importance, the indicator deduction model library can cover 420 indicator subcategories.

17. The transformer state evaluation device according to claim 16, characterized in that: For state indicators whose data characteristics are growth, decline, growth rate, and decline rate, the change cycles of indicator data include weekly change / change rate, daily change rate / change rate, 4h change / change rate, and 2h change / change rate. For weekly variation, daily variation, 4h variation and 2h variation, the calculation formula of the variation period is: △C=|C i,2 -C i,1 Among them, △C is the weekly, daily, 4h and 2h changes, unit is μL / L; C i,2 The latest online data of the corresponding characteristic gas, in μL / L; C i,1 Corresponding characteristic gas reference value, unit: μL / L; For the weekly change rate, daily change rate, 4h change rate and 2h change rate, the calculation formula for the change period is: |Latest online data-cycle variation gas reference value| / cycle variation gas reference value×100%.

18. The transformer state evaluation device according to claim 17, characterized in that: The state quantity deduction value obtained by the "indicator layer" model is determined by the degradation level and importance of the state quantity indicator. No deduction is made when the state quantity indicator is normal.

19. The transformer state evaluation device according to claim 13, characterized in that: The "component-dimension layer" scoring model in the construction module deducts points from each component according to the state quantity index of each component of the transformer in each dimension, and obtains the individual score of each component in each dimension. The calculation formula for the cumulative deduction is: Among them, m represents the component, s represents the dimension, x1,x2,…,x n represents all state indicators of dimension s under component m, t1, t2, …, t n Represents state quantity index x1, x2,…, x n The deduction value.

20. The transformer state evaluation device according to claim 13, characterized in that: The "equipment-component layer" scoring model in the construction module is to deduct points from each component by accumulating the state quantity indicators of each transformer component in all dimensions, and obtain the comprehensive score of each component in all dimensions. The calculation formula is: Among them, x 11 ,...,x 1n ,x 21 ,...,x 2n ,x 31 ,...,x 3n ,x 41 ,...,x 4n ,x 51 ,...,x 5n Represents all state indicators belonging to component m in all dimensions, Indicates the deduction value of all state quantity indicators in all dimensions of component m.

21. The transformer state evaluation device according to claim 13, characterized in that: The "equipment layer" scoring model in the building module is based on the comprehensive evaluation algorithm model of AHP-dynamic weighting fusion, calculates the weight coefficient of each component in the comprehensive score, assigns weights to the scores of each component, and evaluates the status of the transformer based on the transformer status score and the health status level of the transformer obtained according to the indicator attributes and the results output by other hierarchical scoring models.

22. The transformer state evaluation device according to claim 21, characterized in that: The "equipment layer" scoring model refers to the State Grid guidelines and matches the health status of components, including normal status, caution status, abnormal status and severe status.

23. The transformer state evaluation device according to claim 22, characterized in that: The matching transformer health status level, the overall evaluation of the transformer should integrate the evaluation results of its components: when all components are evaluated as normal, the overall evaluation is normal; when the status of any component is warning status, abnormal status or serious status, the overall evaluation should be the most serious status among them.

24. The transformer state evaluation device according to claim 13, characterized in that: The data sources of the acquisition module include professional inspection data, experimental data, daily inspection data, and system data.

25. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented.

Images