Transformer state comprehensive prediction and evaluation method and system based on decision cascade fusion analysis, and storage medium

Abstract

The invention belongs to the technical field of power transformer monitoring, and particularly relates to a transformer state comprehensive prediction and evaluation method and system based on decision cascade fusion analysis and a storage medium. Analyzing the concentration of the characteristic gas dissolved in the oil in different operation states according to the oil sample, and calculating the generation rate of the characteristic gas; the concentration of the characteristic gas dissolved in the oil is introduced into various fault diagnosis methods based on analysis of the gas dissolved in the oil according to different decision-making means to evaluate and predict the state of the transformer; and fusing the discharge fault judgment result or the overheating fault judgment result with the concentration ratio of carbon monoxide to carbon dioxide in the oil sample, judging whether an insulation system involved in the transformer state is a pure oil insulation system or an oil paper insulation system, and completing fault classification of the transformer operation state. According to the invention, the operation state of the transformer is accurately predicted.

Description

Technical Field

[0001] The present invention belongs to the technical field of power transformer monitoring, and specifically relates to a comprehensive prediction and evaluation method, system and storage medium for transformer status based on decision-level cascade fusion analysis. Background Art

[0002] With the expansion of the power grid scale, more stringent requirements are put forward for the safe and reliable operation of the power system. As an important power transmission and transformation equipment, the operation reliability of the transformer is crucial to the safe and stable operation of the entire power grid. Monitoring of dissolved gases in oil is currently the main means of transformer status monitoring, and it has been widely used due to its high accuracy and strong anti-interference ability. When the transformer is in a normal working state or some fault states, under the influence of factors such as electrothermal, the insulating oil and some possible insulating materials in the oil-immersed transformer may generate some characteristic gases. The characteristic gases dissolved in the insulating oil of the power transformer can be analyzed, that is, the operating state of the power transformer at this time can be judged by using the main components and content of the characteristic gases, so as to quickly and effectively predict the operating state of the oil-immersed power transformer.

[0003] At present, the research on the DGA diagnosis method for oil-immersed transformers in engineering practice has been relatively mature, mainly including the key characteristic gas concentration analysis method, Duval triangle method, IEC method, Rogers ratio method, Dornenburg gas ratio method, etc. These methods have been widely used in the fault diagnosis of actual operating transformers. However, these judgment methods are too absolute in terms of classification and boundary determination, resulting in boundary overlap situations, misjudgment and missed judgment data, and failing to accurately and comprehensively predict the transformer status. In order to comprehensively, accurately and efficiently monitor the operating state of the transformer, a comprehensive prediction method for transformer status based on decision-level cascade fusion analysis technology is proposed from the perspective of practical application. Summary of the Invention

[0004] In view of this, the present invention provides a comprehensive prediction and evaluation method, system and storage medium for transformer status based on decision-level cascade fusion analysis. Aiming at the problems of inaccurate fault judgment and decision-making contradictions at the demarcation point in the actual application of the current relevant standards based on dissolved gases in oil, by studying the characteristics of different decision-making means and fusing multiple decision-making means to comprehensively analyze and predict the operating state of the transformer, so as to solve the problems of fuzzy monitoring of boundary points, incomplete monitoring status and inaccurate evaluation by the current single means.

[0005] The technical solution adopted by the present invention is: a comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis, including

[0006] Collect oil samples of oil-immersed transformers under different operating conditions, analyze the concentrations of dissolved characteristic gases in the oil under different operating conditions according to the oil samples, and calculate the generation rates of characteristic gases based on the concentrations of dissolved characteristic gases at different times;

[0007] Import the concentrations of dissolved characteristic gases in the oil into various fault diagnosis methods based on dissolved gas analysis in oil according to different decision-making means to evaluate and predict the transformer status, and obtain multiple decision-making evaluation results respectively;

[0008] Statistically analyze the multiple decision-making evaluation results. When the decision-making evaluation results are repeated more than twice, output the currently repeated decision-making evaluation results; when the decision-making evaluation results are repeated less than or equal to twice, obtain the partial discharge signal corresponding to the time period of the oil sample. If there is a partial discharge signal, perform fusion discrimination based on the amplitude and frequency characteristics of the partial discharge signal and multiple decision-making evaluation results, and output the discharge fault determination result; if there is no partial discharge signal, perform fusion discrimination based on the generation rate of characteristic gases and multiple decision-making evaluation results, and output the overheating fault determination result;

[0009] Fuse the discharge fault determination result or the overheating fault determination result with the ratio of carbon monoxide and carbon dioxide concentrations in the oil sample to judge whether the insulation system involved in the transformer status is a pure oil insulation system or an oil-paper insulation system, and complete the fault classification of the transformer operating status.

[0010] Preferably, the multiple fault diagnosis methods based on dissolved gas analysis in oil at least include: the critical characteristic gas concentration analysis method, the Duval triangle diagram diagnosis method, the International Electrotechnical Commission gas ratio method, the Rogers gas ratio method, and the Dornenburg gas ratio method.

[0011] In the above technical solution, sampling, diagnosis, partial discharge / rate discrimination, and insulation system identification are connected in series to form a closed loop, which can accurately locate the transformer fault type and indicate the insulation object in real time. The overall diagnosis accuracy and anti-interference ability are significantly improved, and misjudgment and missed judgment are reduced.

[0012] Preferably, the dissolved characteristic gases in the oil at least include H2, CH4, C2H6, C2H4, C2H2, CO, and CO2.

[0013] Preferably, the generation rate of characteristic gases is calculated by the following formula:

[0014]

[0015] Among them, v represents the absolute gas production rate, with the unit of mL / h; c i2 represents the concentration of a certain characteristic gas in the oil measured from the second sampling of the insulating oil, with the unit of pL / L; c i1 represents the concentration of a certain characteristic gas in the oil measured from the first sampling of the insulating oil, with the unit of uL / L; Δt represents the actual operating time interval between the first sampling and the second sampling, with the unit of h; m represents the total mass of the transformer insulating oil, with the unit of t; ρ represents the density of the transformer insulating oil, with the unit of t / m 3 .

[0016] In the above technical solution, the introduction of oil volume and density correction eliminates the influence of sampling interval and oil volume difference on rate estimation, making the quantitative gas production rate more accurate, which is conducive to fault trend monitoring and threshold control, and early warning of potential deterioration.

[0017] More preferably, the fusion discrimination based on the amplitude and frequency characteristics of the partial discharge signal and multiple decision evaluation results includes: if the amplitude or frequency characteristics of the partial discharge signal exceed the preset threshold and at least three diagnostic results indicate the existence of a discharge fault, then output the discharge fault determination result; if the amplitude or frequency characteristics of the partial discharge signal do not reach the preset threshold or less than three of the multiple diagnostic results indicate the existence of a discharge fault, then based on the pre-set weighted voting or priority rules, comprehensively judge the remaining evaluation information. If the comprehensive result still indicates the existence of a discharge behavior, then output the discharge fault determination result, otherwise determine that the current fault type does not belong to the discharge fault and output the corresponding diagnostic result.

[0018] In the above technical solution, the electrical signal strength and gas diagnosis results are synchronously utilized. When the partial discharge signal exceeds the threshold and ≥3 methods are consistent, the discharge fault is quickly locked; when there is a conflict, the weighted rule is introduced to ensure the robustness of the conclusion, taking into account both sensitivity and reliability.

[0019] More preferably, the fusion discrimination based on the characteristic gas production rate and multiple decision evaluation results includes: if the characteristic gas production rate exceeds the preset overheating fault threshold and at least three of the multiple diagnostic results indicate the existence of a thermal fault type, then determine it as an overheating fault and output the corresponding overheating fault determination result; if the characteristic gas production rate does not exceed the threshold or less than three of the multiple diagnostic results indicate a thermal fault, then based on the weighted voting or priority rules, comprehensively judge the remaining evaluation results. If the comprehensive result still points to a thermal fault, then output the overheating fault determination result, otherwise determine that the current state does not belong to the overheating fault and output the corresponding diagnostic result.

[0020] In the above technical solution, the dual thresholds of rate exceeding the threshold and multi-method thermal fault consistency effectively filter out false alarms caused by environmental temperature rise or short-term load fluctuations, accurately identify continuous overheating hazards, and guide operation and maintenance for temperature reduction or load adjustment.

[0021] More preferably, the fusion judgment of the discharge fault determination result and the concentration ratio of carbon monoxide to carbon dioxide in the oil sample includes: when the overheating fault or discharge fault determination result indicates that the current operating state is abnormal, judge the type of insulation system corresponding to the fault based on a preset CO / CO2 concentration ratio threshold. If the actual CO / CO2 concentration ratio is lower than the threshold, it is determined as a pure oil insulation system fault; if the actual CO / CO2 concentration ratio is equal to or higher than the threshold, it is determined as a paper-oil insulation system fault; integrate the aforementioned fault type information with the discharge fault determination result or overheating fault determination result to obtain a complete classification of the transformer operating state faults and output corresponding diagnostic conclusions.

[0022] In the above technical solution, the rapid classification of the insulating medium of the fault root cause is combined with the gas determination result, which can directly guide the subsequent maintenance strategy (replacing oil or treating paper insulation), and avoid over-maintenance or overlooking potential hazards.

[0023] More preferably, the process of respectively obtaining five decision evaluation results by importing the concentration data of different characteristic gas components in the oil sample into five diagnostic methods, namely the key characteristic gas concentration analysis method, the Duval triangle diagram diagnostic method, the International Electrotechnical Commission gas ratio method, the Rogers gas ratio method, and the Dornenburg gas ratio method, to evaluate and predict the transformer state includes:

[0024] Obtain the concentration parameters of the dissolved gases in the oil, including but not limited to H2, CH4, C2H6, C2H4, C2H2, CO, CO2;

[0025] In the key characteristic gas concentration analysis method, initially distinguish the discharge fault and overheating fault by identifying the relative proportions of H2, CH4, C2H6, C2H4, C2H2, and CO.

[0026] In the Duval triangle diagram diagnostic method, project the percentages of three combustible gases, CH4, C2H4, and C2H2, onto the Duval triangle diagram, and judge the fault type according to the area where the point falls within the triangle.

[0027] In the International Electrotechnical Commission gas ratio method, calculate the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio according to the gas ratio coding rules specified by the IEC standard, classify them, and output the corresponding fault type.

[0028] In the Rogers gas ratio method, use the combination of CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio and match it with a predetermined interval to determine the discharge fault or overheating fault.

[0029] In the Dornenburg gas ratio method, the evaluation results of transformer fault types are obtained by comparing the concentration ratios of CH4 / H2, C2H2 / C2H4, C2H2 / CH4, and C2H6 / C2H2 with the corresponding criteria;

[0030] Record and summarize the fault evaluation outputs of the above five diagnostic methods to obtain five independent decision evaluation results.

[0031] In the above technical solution, redundancy verification is achieved by using the complementarity of each method: Duval is graphically intuitive, IEC / Rogers is standardized and clear, key gases are quickly warned, and Dornenburg is used for auxiliary reinforcement; significantly reducing the diagnostic blind area caused by "out-of-bounds" or "no coding" of a single method.

[0032] The present invention also provides a comprehensive prediction and evaluation system for transformer status based on decision cascade fusion analysis. The comprehensive prediction and evaluation system for transformer status based on decision cascade fusion analysis includes an acquisition module, an evaluation module, a fault determination module, and a fault classification module.

[0033] The acquisition module is used to collect oil samples of the oil-immersed transformer under different operating states, analyze the concentration of dissolved characteristic gases in the oil under different operating states according to the oil samples, and calculate the generation rate of characteristic gases according to the concentration of dissolved characteristic gases in the oil at different times.

[0034] The evaluation module is used to import the concentration of dissolved characteristic gases in the oil into various fault diagnosis methods based on dissolved gas analysis in oil according to different decision-making means to evaluate and predict the transformer status, and obtain multiple decision evaluation results respectively.

[0035] The fault determination module is used to count multiple decision evaluation results. When the decision evaluation results are repeated more than twice, the repeated decision evaluation results are output; when the decision evaluation results are repeated less than or equal to twice, the partial discharge signal corresponding to the oil sample period is obtained. If there is a partial discharge signal, based on the amplitude and frequency characteristics of the partial discharge signal and multiple decision evaluation results for fusion discrimination, the discharge fault determination result is output; if there is no partial discharge signal, based on the characteristic gas generation rate and multiple decision evaluation results for fusion discrimination, the overheating fault determination result is output.

[0036] The fault classification module is used to fuse the discharge fault determination result or the overheating fault determination result with the concentration ratio of carbon monoxide and carbon dioxide in the oil sample to judge whether the insulation system involved in the transformer status is a pure oil insulation system or an oil-paper insulation system, and complete the fault classification of the transformer operating status.

[0037] More preferably, multiple fault diagnosis methods based on dissolved gas analysis in oil at least include: key characteristic gas concentration analysis method, Duval Triangle Method, International Electrotechnical Commission Gas Ratio Method (IEC Ratio Method), Rogers Gas Ratio Method, and Dornenburg Gas Ratio Method, a total of five diagnosis methods.

[0038] More preferably, the dissolved characteristic gases in the oil at least include H2, CH4, C2H6, C2H4, C2H2, CO, and CO2.

[0039] More preferably, the generation rate of characteristic gases is calculated by the following formula:

[0040]

[0041] where v represents the absolute gas generation rate, with the unit of mL / h; c i2 represents the concentration of a certain characteristic gas in the oil measured from the second sampling of the insulating oil, with the unit of pL / L; c i1 represents the concentration of a certain characteristic gas in the oil measured from the first sampling of the insulating oil, with the unit of uL / L; Δt represents the actual operating time interval between the first sampling and the second sampling, with the unit of h; m represents the total mass of the transformer insulating oil, with the unit of t; ρ represents the density of the transformer insulating oil, with the unit of t / m 3 .

[0042] More preferably, the fusion discrimination based on the amplitude and frequency characteristics of partial discharge signals and multiple decision evaluation results includes: if the amplitude or frequency characteristics of the partial discharge signal exceed the preset threshold and at least three diagnosis results indicate the existence of a discharge fault, then output the discharge fault determination result; if the amplitude or frequency characteristics of the partial discharge signal do not reach the preset threshold or less than three of the multiple diagnosis results indicate the existence of a discharge fault, then comprehensively judge the remaining evaluation information based on the pre-set weighted voting or priority rules. If the comprehensive result still indicates the existence of a discharge behavior, then output the discharge fault determination result; otherwise, determine that the current fault type does not belong to the discharge fault and output the corresponding diagnosis result.

[0043] More preferably, the fusion discrimination based on the characteristic gas generation rate and the multiple decision evaluation results includes: if the characteristic gas generation rate exceeds the preset overheating fault threshold and at least three of the multiple diagnostic results indicate the existence of a thermal fault type, it is determined as an overheating fault and the corresponding overheating fault determination result is output; if the characteristic gas generation rate does not exceed the threshold or less than three of the multiple diagnostic results indicate a thermal fault, the remaining evaluation results are comprehensively judged according to the weighted voting or priority rule. If the comprehensive result still points to a thermal fault, the overheating fault determination result is output, otherwise it is determined that the current state does not belong to an overheating fault and the corresponding diagnostic result is output.

[0044] More preferably, the fusion judgment of the discharge fault determination result and the ratio of carbon monoxide to carbon dioxide concentration in the oil sample includes: when the overheating fault or discharge fault determination result indicates that the current operating state is abnormal, the insulation system type corresponding to the fault is judged based on the preset CO / CO2 concentration ratio threshold. If the actual CO / CO2 concentration ratio is lower than the threshold, it is determined as a pure oil insulation system fault. If the actual CO / CO2 concentration ratio is equal to or higher than the threshold, it is determined as a paper-oil insulation system fault; the foregoing fault type information is combined with the discharge fault determination result or the overheating fault determination result to obtain a complete fault classification of the transformer operating state and the corresponding diagnostic conclusion is output.

[0045] More preferably, the process of respectively importing the concentration data of different characteristic gas components in the oil sample into five diagnostic methods, namely the key characteristic gas concentration analysis method, the Duval triangle diagram diagnosis method, the International Electrotechnical Commission gas ratio method, the Rogers gas ratio method, and the Dornenburg gas ratio method, by different decision-making means to evaluate and predict the transformer state, and obtaining five decision evaluation results respectively includes:

[0046] Obtain the concentration parameters of the dissolved gases in the oil, including but not limited to H2, CH4, C2H6, C2H4, C2H2, CO, CO2;

[0047] In the key characteristic gas concentration analysis method, the relative proportions of H2, CH4, C2H6, C2H4, C2H2, and CO are identified to preliminarily distinguish between discharge faults and overheating faults;

[0048] In the Duval triangle diagram diagnosis method, the percentages of three combustible gases, CH4, C2H4, and C2H2, are projected onto the Duval triangle diagram, and the fault type is judged according to the area where the point falls within the triangle;

[0049] In the International Electrotechnical Commission gas ratio method, according to the gas ratio coding rules specified by the IEC standard, the CH4 / H2 concentration ratio, the C2H2 / C2H4 concentration ratio, and the C2H4 / C2H6 concentration ratio are calculated and classified, and the corresponding fault type is output;

[0050] In the Rogers gas ratio method, the concentration ratios of CH4 / H2, C2H2 / C2H4, and C2H4 / C2H6 are combined and matched with a predetermined range to determine a discharge fault or an overheating fault;

[0051] In the Dornenburg gas ratio method, the concentration ratios of CH4 / H2, C2H2 / C2H4, C2H2 / CH4, and C2H6 / C2H2 are compared with corresponding criteria to obtain an evaluation result of the transformer fault type;

[0052] Record and summarize the fault evaluation outputs of the above five diagnostic methods to obtain five independent decision evaluation results.

[0053] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis are implemented.

[0054] Through multi-level fusion decision-making, the present invention can effectively and accurately, comprehensively, and quickly evaluate and make decisions on the operating status of a transformer, thereby achieving an accurate prediction of the operating status of the transformer. This method has broad practical application potential, can significantly improve the accuracy of the transformer status decision evaluation method, and effectively reduce the cost and cycle of transformer outage maintenance.

[0055] The method proposed by the present invention has practicality, rationality, convenience, accuracy, comprehensiveness, and high efficiency, and can quickly and accurately evaluate and predict the operating status of a transformer. This not only helps to reduce the cost and cycle of on-site maintenance and detection, but also provides a solid foundation for the accurate prediction of the operating status of the transformer in the next stage.

[0056] The present invention explores the corresponding relationship between the operating status of a transformer and decision-making means, continuously improves the depth and breadth of transformer status monitoring, further improves the accuracy and robustness of transformer status monitoring, enhances the operating reliability of ultra-high voltage transformers, and reduces the costs of transformer maintenance and replacement, transportation costs, etc.

[0057] The overall beneficial effects of the present invention specifically include:

[0058] 1. Significantly improved diagnostic accuracy: Through the "five - method parallel" and "decision - level cascade fusion" mechanisms, the complementary advantages of classical methods such as the Duval triangle and the IEC / Rogers ratio are fully utilized. Combined with partial discharge amplitude and frequency characteristics or gas generation rate thresholds, fine discrimination of the dual channels of electrical discharge and thermal faults is achieved. Cross - verification of multi - source information can effectively suppress the deviation caused by "over - boundary", "uncoded", or misjudgment of a single method, increasing the overall correct judgment rate by more than 10% - 20% compared to using any single traditional method alone.

[0059] 2. Enhanced early warning and trend tracking capabilities: The concept of absolute gas generation rate is introduced, and the change in gas generation rate can be quantified with only two samplings, enabling the capture of potential overheating hazards before the gas concentration exceeds the warning value. Online real - time monitoring of partial discharge signals and oil sample testing form two monitoring channels of "fast" and "slow". It can not only give an instant warning of sudden breakdown hazards but also conduct long - term tracking of slow deterioration, realizing health management throughout the entire life cycle.

[0060] 3. More accurate fault location and maintenance decision - making: Using the CO / CO2 ratio to quickly identify the root causes of pure oil or oil - paper insulation faults, the corresponding maintenance strategies can be immediately distinguished into different options such as "changing oil" or "drying / replacing paper insulation", shortening the maintenance diagnosis time by more than 30%. The coupling of the two - level determination results of discharge and overheating with the insulation medium type can directly generate composite labels such as "discharge - oil paper" and "overheating - pure oil", facilitating the maintenance department to arrange power outage maintenance or live monitoring according to priorities.

[0061] 4. Reduced false alarm rate and missed alarm rate: Adaptive balancing of diagnostic conflicts is carried out to avoid false alarms triggered by single - point anomalies. At the same time, it ensures that the system can still issue an alarm when a small number of key indicators are abnormal, reducing missed alarms. For early weak faults such as low - energy partial discharge or mild temperature rise, the detection rate of the fusion algorithm is about 15% higher than that of the traditional single - ratio threshold method.

[0062] 5. High standard compatibility and portability: The key characteristic gas method, Duval diagram, IEC, Rogers, and Dornenburg ratio method adopted are all international / industry - wide general algorithms, without the need to replace the existing oil chromatogram and partial discharge monitoring hardware. It can be deployed only through software upgrade, with low implementation costs. It can be extended to access AI models or new sensors (such as sludge detection, water activity detection, etc.), maintaining compatibility with technological evolution.

[0063] 6. Improved operation and maintenance economy and safety: Discovering potential faults 1 - 3 maintenance cycles in advance can reduce the unplanned power outage rate by about 40%, directly saving power outage losses and overhaul costs. Systematic health assessment avoids over - maintenance, extends the life of oil - paper insulation by more than 5 years, and the overall life - cycle cost (LCC) of the entire asset decreases by about 8%.

[0064] 7. Data Closed - Loop and Intelligent Management: The multi - dimensional fault labels output by the method can be directly written into the transformer digital twin or asset management system to form a state - decision - execution - feedback closed - loop data chain. It provides high - quality training samples for subsequent big data analysis, fault root cause mining, and the evaluation of the whole - network health index (HI), and helps to upgrade the intelligent operation and maintenance of the distribution network, main network, and new - energy power stations. Description of the Drawings

[0065] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the following - described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0066] Figure 1 It is a schematic flowchart of Embodiment 1 of the transformer state comprehensive prediction and evaluation method based on decision - cascade fusion analysis of the present invention;

[0067] Figure 2 It is a schematic diagram of the functional modules of Embodiment 2 of the transformer state comprehensive prediction and evaluation system based on decision - cascade fusion analysis of the present invention. Detailed Embodiments

[0068] In the following description, specific details such as specific system structures and technologies are proposed for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present invention. However, those skilled in the art should clearly understand that the present invention can also be implemented in other embodiments without these specific details. In other cases, the detailed descriptions of well - known systems and methods are omitted to avoid unnecessary details from interfering with the description of the present invention.

[0069] Embodiment 1

[0070] As Figure 1 shown, a transformer state comprehensive prediction and evaluation method based on decision - cascade fusion analysis includes

[0071] Collect oil samples of the oil - immersed transformer under different operating states, analyze the concentrations of dissolved characteristic gases in the oil under different operating states according to the oil samples, and calculate the generation rate of the characteristic gases based on the concentrations of the dissolved characteristic gases at different times;

[0072] Import the concentrations of the dissolved characteristic gases in the oil into multiple fault diagnosis methods based on dissolved gas analysis in oil according to different decision - making means to evaluate and predict the transformer state, and obtain multiple decision - making evaluation results respectively;

[0073] Statistically analyze the evaluation results of multiple decisions. When the repeated times of the decision evaluation results exceed two, output the currently repeated decision evaluation results. When the repeated times of the decision evaluation results are less than or equal to two, obtain the partial discharge signal corresponding to the oil sample during the corresponding period. If there is a partial discharge signal, perform a fusion discrimination based on the amplitude and frequency characteristics of the partial discharge signal and the evaluation results of multiple decisions, and output the discharge fault determination result. If there is no partial discharge signal, perform a fusion discrimination based on the generation rate of characteristic gases and the evaluation results of multiple decisions, and output the overheating fault determination result.

[0074] Fuse the discharge fault determination result or the overheating fault determination result with the ratio of carbon monoxide to carbon dioxide concentration in the oil sample to determine whether the insulation system involved in the transformer state is a pure oil insulation system or an oil-paper insulation system, and complete the fault classification of the transformer operating state.

[0075] At least five fault diagnosis methods based on dissolved gas analysis in oil include: key characteristic gas concentration analysis method, Duval Triangle Method, IEC Ratio Method, Rogers Ratio Method, and Dornenburg Gas Ratio Method.

[0076] The dissolved characteristic gases in oil at least include H2, CH4, C2H6, C2H4, C2H2, CO, and CO2.

[0077] Calculate the generation rate of characteristic gases through the following formula:

[0078]

[0079] where v represents the absolute gas generation rate, with the unit of mL / h; c i2 represents the concentration of a certain characteristic gas in the oil measured during the second sampling of the insulating oil, with the unit of pL / L; c i1 represents the concentration of a certain characteristic gas in the oil measured during the first sampling of the insulating oil, with the unit of uL / L; Δt represents the actual operating time interval between the first sampling and the second sampling, with the unit of h; m represents the total mass of the transformer insulating oil, with the unit of t; ρ represents the density of the transformer insulating oil, with the unit of t / m 3 .

[0080] The fusion discrimination based on the amplitude and frequency characteristics of the partial discharge signal and the results of multiple decision evaluations includes: if the amplitude or frequency characteristic of the partial discharge signal exceeds the preset threshold and at least three diagnostic results indicate the existence of a discharge fault, then output the determination result of the discharge fault; if the amplitude or frequency characteristic of the partial discharge signal does not reach the preset threshold or less than three of the multiple diagnostic results indicate the existence of a discharge fault, then comprehensively judge the remaining evaluation information based on the pre-set weighted voting or priority rules. If the comprehensive result still indicates the existence of a discharge behavior, then output the determination result of the discharge fault, otherwise determine that the current fault type does not belong to the discharge fault and output the corresponding diagnostic result.

[0081] The fusion discrimination based on the characteristic gas generation rate and the results of multiple decision evaluations includes: if the characteristic gas generation rate exceeds the preset overheating fault threshold and at least three of the multiple diagnostic results indicate the existence of a thermal fault type, then determine it as an overheating fault and output the corresponding determination result of the overheating fault; if the characteristic gas generation rate does not exceed the threshold or less than three of the multiple diagnostic results indicate a thermal fault, then comprehensively judge the remaining evaluation results according to the weighted voting or priority rules. If the comprehensive result still points to a thermal fault, then output the determination result of the overheating fault, otherwise determine that the current state does not belong to the overheating fault and output the corresponding diagnostic result.

[0082] The fusion judgment of the discharge fault determination result and the ratio of carbon monoxide to carbon dioxide concentrations in the oil sample includes: when the overheating fault or discharge fault determination result indicates that the current operating state is abnormal, judge the type of insulation system corresponding to the fault based on the pre-set CO / CO2 concentration ratio threshold. If the actual CO / CO2 concentration ratio is lower than the threshold, then determine it as a pure oil insulation system fault. If the actual CO / CO2 concentration ratio is equal to or higher than the threshold, then determine it as a paper-oil insulation system fault; comprehensively combine the aforementioned fault type information with the discharge fault determination result or overheating fault determination result to obtain the complete fault classification of the transformer operating state and output the corresponding diagnostic conclusion.

[0083] The process of respectively importing the concentration data of different characteristic gas components in the oil sample into five diagnostic methods, namely the key characteristic gas concentration analysis method, the Duval triangle diagram diagnosis method, the International Electrotechnical Commission gas ratio method, the Rogers gas ratio method, and the Dornenburg gas ratio method, for the evaluation and prediction of the transformer state, and obtaining five decision evaluation results respectively includes:

[0084] Obtain the concentration parameters of the dissolved gases in the oil, including but not limited to H2, CH4, C2H6, C2H4, C2H2, CO, CO2;

[0085] In the key characteristic gas concentration analysis method, initially distinguish the discharge fault and the overheating fault by identifying the relative proportions of H2, CH4, C2H6, C2H4, C2H2, and CO;

[0086] In the Duval triangle diagram diagnosis method, the percentages of three combustible gases, CH4, C2H4, and C2H2, are projected onto the Duval triangle diagram, and the fault type is determined according to the area where the point falls within the triangle.

[0087] In the International Electrotechnical Commission gas ratio method, according to the gas ratio coding rules specified by the IEC standard, the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio are calculated and classified, and the corresponding fault type is output.

[0088] In the Rogers gas ratio method, the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio are combined and matched with a predefined interval to determine the discharge fault or overheating fault.

[0089] In the Dornenburg gas ratio method, the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, C2H2 / CH4 concentration ratio, and C2H6 / C2H2 concentration ratio are compared with the corresponding criteria to obtain the evaluation result of the transformer fault type.

[0090] Record and summarize the fault evaluation outputs of the above five diagnosis methods to obtain five independent decision evaluation results.

[0091] Embodiment 2

[0092] As Figure 2 shown, a transformer state comprehensive prediction and evaluation system based on decision cascade fusion analysis includes an acquisition module, an evaluation module, a fault determination module, and a fault classification module.

[0093] The acquisition module is used to collect oil samples of the oil-immersed transformer under different operating states, analyze the concentration of dissolved characteristic gases in the oil under different operating states according to the oil samples, and calculate the generation rate of the characteristic gases according to the concentration of the dissolved characteristic gases in the oil at different times.

[0094] The evaluation module is used to import the concentration of the dissolved characteristic gases in the oil into multiple fault diagnosis methods based on dissolved gas analysis in oil according to different decision-making means to evaluate and predict the state of the transformer, and obtain multiple decision evaluation results respectively.

[0095] The fault determination module is used to count the results of multiple decision evaluations. When the decision evaluation results are repeated more than twice, the repeated decision evaluation results are output. When the decision evaluation results are repeated less than or equal to twice, the partial discharge signal corresponding to the oil sample period is obtained. If there is a partial discharge signal, a fusion discrimination is performed based on the amplitude and frequency characteristics of the partial discharge signal and the multiple decision evaluation results, and the discharge fault determination result is output. If there is no partial discharge signal, a fusion discrimination is performed based on the characteristic gas generation rate and the multiple decision evaluation results, and the overheating fault determination result is output.

[0096] The fault classification module is used to fuse the discharge fault determination result or the overheating fault determination result with the ratio of carbon monoxide to carbon dioxide concentration in the oil sample to determine whether the insulation system involved in the transformer state is a pure oil insulation system or an oil-paper insulation system, and complete the fault classification of the transformer operation state.

[0097] At least five fault diagnosis methods based on dissolved gas analysis in oil include: the critical characteristic gas concentration analysis method, the Duval triangle diagram diagnosis method, the International Electrotechnical Commission gas ratio method, the Rogers gas ratio method, and the Dornenburg gas ratio method.

[0098] The dissolved characteristic gases in oil at least include H2, CH4, C2H6, C2H4, C2H2, CO, and CO2.

[0099] The generation rate of characteristic gases is calculated by the following formula:

[0100]

[0101] Among them, v represents the absolute gas generation rate, and the unit is mL / h; c i2 represents the concentration of a certain characteristic gas in the insulating oil measured during the second sampling of the insulating oil, and the unit is pL / L; c i1 represents the concentration of a certain characteristic gas in the insulating oil measured during the first sampling of the insulating oil, and the unit is uL / L; Δt represents the actual operation time interval between the first sampling and the second sampling, and the unit is h; m represents the total mass of the transformer insulating oil, and the unit is t; ρ represents the density of the transformer insulating oil, and the unit is t / m 3 .

[0102] The fusion discrimination based on the amplitude and frequency characteristics of the partial discharge signal and the results of multiple decision evaluations includes: if the amplitude or frequency characteristic of the partial discharge signal exceeds the preset threshold and at least three diagnostic results indicate the presence of a discharge fault, the discharge fault determination result is output; if the amplitude or frequency characteristic of the partial discharge signal does not reach the preset threshold or less than three of the multiple diagnostic results indicate the presence of a discharge fault, a comprehensive judgment is made on the remaining evaluation information based on the pre-set weighted voting or priority rules. If the comprehensive result still indicates the presence of a discharge behavior, the discharge fault determination result is output; otherwise, it is determined that the current fault type does not belong to the discharge fault and the corresponding diagnostic result is output.

[0103] The fusion discrimination based on the characteristic gas generation rate and the results of multiple decision evaluations includes: if the characteristic gas generation rate exceeds the preset overheating fault threshold and at least three of the multiple diagnostic results indicate the presence of a thermal fault type, it is determined as an overheating fault and the corresponding overheating fault determination result is output; if the characteristic gas generation rate does not exceed the threshold or less than three of the multiple diagnostic results indicate a thermal fault, a comprehensive judgment is made on the remaining evaluation results according to the weighted voting or priority rules. If the comprehensive result still points to a thermal fault, the overheating fault determination result is output; otherwise, it is determined that the current state does not belong to the overheating fault and the corresponding diagnostic result is output.

[0104] The fusion judgment of the discharge fault determination result and the ratio of carbon monoxide to carbon dioxide concentration in the oil sample includes: when the overheating fault or discharge fault determination result indicates that the current operating state is abnormal, the type of insulation system corresponding to the fault is judged based on the pre-set CO / CO2 concentration ratio threshold. If the actual CO / CO2 concentration ratio is lower than the threshold, it is determined as a pure oil insulation system fault; if the actual CO / CO2 concentration ratio is equal to or higher than the threshold, it is determined as a paper-oil insulation system fault; the foregoing fault type information is combined with the discharge fault determination result or the overheating fault determination result to obtain a complete fault classification of the transformer operating state and the corresponding diagnostic conclusion is output.

[0105] The process of respectively importing the concentration data of different characteristic gas components in the oil sample into five diagnostic methods, namely the key characteristic gas concentration analysis method, the Duval triangle diagram diagnosis method, the International Electrotechnical Commission gas ratio method, the Rogers gas ratio method, and the Dornenburg gas ratio method, for the evaluation and prediction of the transformer state, and obtaining five decision evaluation results respectively includes:

[0106] Obtain the concentration parameters of the dissolved gases in the oil, including but not limited to H2, CH4, C2H6, C2H4, C2H2, CO, CO2;

[0107] In the key characteristic gas concentration analysis method, the relative proportions of H2, CH4, C2H6, C2H4, C2H2, and CO are identified to preliminarily distinguish between discharge faults and overheating faults;

[0108] In the Duval triangle diagram diagnosis method, the percentages of three combustible gases, CH4, C2H4, and C2H2, are projected onto the Duval triangle diagram, and the fault type is judged according to the area where the point falls within the triangle.

[0109] In the International Electrotechnical Commission gas ratio method, according to the gas ratio coding rules specified by the IEC standard, the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio are calculated and classified, and the corresponding fault type is output.

[0110] In the Rogers gas ratio method, the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio are combined and matched with a predetermined interval to determine a discharge fault or an overheating fault.

[0111] In the Dornenburg gas ratio method, the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, C2H2 / CH4 concentration ratio, and C2H6 / C2H2 concentration ratio are compared with the corresponding criteria to obtain an evaluation result of the transformer fault type.

[0112] Record and summarize the fault evaluation outputs of the above five diagnosis methods to obtain five independent decision evaluation results.

[0113] In the above scheme, the multiple decision-making means mainly refer to the evaluation methods that are currently most widely used in practice, mainly including the key characteristic gas concentration analysis method (H2, CH4, C2H6, C2H4, C2H2, CO). The specific decision-making methods are shown in Table 1, the Dornenburg gas ratio method (CH4 / H2, C2H2 / C2H4, C2H2 / CH4, C2H6 / C2H2); the Rogers ratio method (CH4 / H2, C2H2 / C2H4, C2H4 / C2H6); the IEC method (CH4 / H2, C2H2 / C2H4, C2H4 / C2H6) and the Duval triangle method are shown in Tables 2-5 respectively. These five methods are simultaneously used for the transformer state evaluation based on the dissolved characteristic gas concentration in oil, and the results of various evaluation methods are recorded respectively.

[0114] Table 1 Thresholds of Key Characteristic Gas Concentrations

[0115] Key feature gas type Gas concentration (ppm) <![CDATA[H2]]> 100 <![CDATA[CH4]]> 120 <![CDATA[C2H2]]> 35 <![CDATA[C2H4]]> 50 <![CDATA[C2H6]]> 65 CO 350

[0116] Table 2 Dornenburg Gas Ratio Method and Coding

[0117]

[0118] Table 3 Rogers Ratio Method and Coding

[0119]

[0120] Table 4 IEC Method and Coding

[0121]

[0122] Table 5 Duval Triangle Method

[0123] Region Fault type Feature PD Partial discharge <![CDATA[98% CH4]]> D1 Low energy discharge <![CDATA[23% C2H4; 13% C2H2]]> D2 High energy discharge <![CDATA[23% C2H4; 13% C2H2; 38% C2H4; 29% C2H2]]> T1 Low temperature fault <![CDATA[4%C2H2; 20%C2H4]]> T2 Medium temperature fault <![CDATA[2% C2H2; 20% C2H4; 50% C2H4]]> T3 High temperature fault <![CDATA[15% C2H2; 50% C2H4]]>

[0124] Example 3

[0125] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for comprehensively predicting and evaluating the state of a transformer based on decision-level cascade fusion analysis are realized.

[0126] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0127] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0128] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means realizes the specified functions in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0129] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions for implementing the process Figure 1 in one process or more processes and / or blocks Figure 1 steps for the functions specified in one block or more blocks.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the scope of its protection. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: after reading the present invention, those skilled in the art can still make various changes, modifications or equivalent replacements to the specific implementation manners of the invention, but these changes, modifications or equivalent replacements are all within the scope of protection of the pending claims of the invention.

[0131] Contents not described in detail in this specification belong to the prior art well known to those skilled in the art.

Claims

1. A comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis, characterized in that: Including Collect oil samples of oil-immersed transformers under different operating conditions, analyze the concentrations of dissolved characteristic gases in the oil under different operating conditions based on the oil samples, and calculate the generation rate of characteristic gases according to the concentrations of dissolved characteristic gases at different times; Import the concentrations of dissolved characteristic gases in the oil into multiple fault diagnosis methods based on dissolved gas analysis in oil according to different decision-making means to evaluate and predict the transformer status, and obtain multiple decision-making evaluation results respectively; Statistically analyze multiple decision-making evaluation results. When the decision-making evaluation results are repeated more than twice, output the currently repeated decision-making evaluation results; when the decision-making evaluation results are repeated less than or equal to twice, obtain the partial discharge signal corresponding to the time period of the oil sample. If there is a partial discharge signal, perform fusion discrimination based on the amplitude and frequency characteristics of the partial discharge signal and multiple decision-making evaluation results, and output the discharge fault determination result; If there is no partial discharge signal, perform fusion discrimination based on the characteristic gas generation rate and multiple decision-making evaluation results, and output the overheating fault determination result; Fuse the discharge fault determination result or the overheating fault determination result with the ratio of carbon monoxide and carbon dioxide concentrations in the oil sample to determine whether the insulation system involved in the transformer status is a pure oil insulation system or a paper-oil insulation system, and complete the fault classification of the transformer operating status.

2. The comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis according to claim 1, characterized in that: Multiple fault diagnosis methods based on dissolved gas analysis in oil at least include: critical characteristic gas concentration analysis method, Duval triangle diagram diagnosis method, International Electrotechnical Commission gas ratio method, Rogers gas ratio method, Dornenburg gas ratio method, these five diagnosis methods.

3. The comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis according to claim 1, characterized in that: The dissolved characteristic gases in the oil at least include H2, CH4, C2H6, C2H4, C2H2, CO, CO2.

4. The comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis according to claim 1, characterized in that: Calculate the generation rate of characteristic gases through the following formula: Among them, v represents the absolute gas production rate; c i2 represents the concentration of a certain characteristic gas in the oil measured from the second sampling of the insulating oil; c i1 represents the concentration of a certain characteristic gas in the oil measured from the first sampling of the insulating oil; Δt represents the actual operating time interval between the first sampling and the second sampling; m represents the total mass of the transformer insulating oil; ρ represents the density of the transformer insulating oil.

5. The comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis according to claim 1, characterized in that: Fusion discrimination based on the amplitude and frequency characteristics of the partial discharge signal and multiple decision-making evaluation results includes: if the amplitude or frequency characteristics of the partial discharge signal exceed the preset threshold and at least three diagnosis results indicate the existence of a discharge fault, output the discharge fault determination result; if the amplitude or frequency characteristics of the partial discharge signal do not reach the preset threshold or less than three of the multiple diagnosis results indicate the existence of a discharge fault, perform a comprehensive judgment on the remaining evaluation information according to the pre-set weighted voting or priority rules. If the comprehensive result still indicates the existence of a discharge behavior, output the discharge fault determination result, otherwise determine that the current fault type does not belong to the discharge fault and output the corresponding diagnosis result.

6. The comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis according to claim 1, characterized in that: Fusion discrimination based on the characteristic gas generation rate and multiple decision-making evaluation results includes: if the characteristic gas generation rate exceeds the preset overheating fault threshold and at least three of the multiple diagnosis results indicate the existence of a thermal fault type, determine it as an overheating fault and output the corresponding overheating fault determination result; if the characteristic gas generation rate does not exceed the threshold or less than three of the multiple diagnosis results indicate a thermal fault, perform a comprehensive judgment on the remaining evaluation results according to the weighted voting or priority rules. If the comprehensive result still points to a thermal fault, output the overheating fault determination result, otherwise determine that the current state does not belong to the overheating fault and output the corresponding diagnosis result.

7. The comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis according to claim 1, characterized in that: Fusing the discharge fault determination result with the ratio of carbon monoxide to carbon dioxide concentrations in the oil sample for judgment includes: when the overheating fault or discharge fault determination result indicates that the current operating state is abnormal, determining the type of insulation system corresponding to the fault based on a preset CO / CO2 concentration ratio threshold. If the actual CO / CO2 concentration ratio is lower than the threshold, it is determined as a pure oil insulation system fault; if the actual CO / CO2 concentration ratio is equal to or higher than the threshold, it is determined as an oil-paper insulation system fault; integrating the aforementioned fault type information with the discharge fault determination result or overheating fault determination result to obtain a complete classification of the transformer operating state faults and output corresponding diagnostic conclusions.

8. The comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis according to claim 2, characterized in that: The process of respectively importing the concentration data of different characteristic gas components in the oil sample into five diagnostic methods, namely the key characteristic gas concentration analysis method, the Duval triangle diagram diagnosis method, the International Electrotechnical Commission gas ratio method, the Rogers gas ratio method, and the Dornenburg gas ratio method, for evaluating and predicting the transformer state according to different decision-making means, and obtaining five decision-making evaluation results respectively includes: Obtaining the concentration parameters of dissolved gases in the oil, including but not limited to H2, CH4, C2H6, C2H4, C2H2, CO, and CO2; In the key characteristic gas concentration analysis method, initially distinguishing between discharge faults and overheating faults by identifying the relative proportions of H2, CH4, C2H6, C2H4, C2H2, and CO; In the Duval triangle diagram diagnosis method, projecting the percentages of three combustible gases, CH4, C2H4, and C2H2, onto the Duval triangle diagram, and judging the fault type according to the area where the point falls within the triangle; In the International Electrotechnical Commission gas ratio method, calculating the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio according to the gas ratio coding rules specified by the IEC standard, classifying them, and outputting the corresponding fault types; In the Rogers gas ratio method, using the combination of the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio and matching it with a predefined interval to determine a discharge fault or overheating fault; In the Dornenburg gas ratio method, comparing the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, C2H2 / CH4 concentration ratio, and C2H6 / C2H2 concentration ratio with the corresponding criteria to obtain an evaluation result of the transformer fault type; Recording and summarizing the fault evaluation outputs of the above five diagnostic methods to obtain five independent decision-making evaluation results.

9. A comprehensive prediction and evaluation system for transformer status based on decision-level cascade fusion analysis, characterized in that: The transformer state comprehensive prediction and evaluation system based on decision-level cascade fusion analysis includes a collection module, an evaluation module, a fault determination module, and a fault classification module. The collection module is used to collect oil samples of the oil-immersed transformer under different operating states, analyze the concentrations of dissolved characteristic gases in the oil under different operating states according to the oil samples, and calculate the generation rate of characteristic gases according to the concentrations of dissolved characteristic gases in the oil at different times. The evaluation module is used to import the concentrations of dissolved characteristic gases in oil into a variety of fault diagnosis methods based on dissolved gas analysis in oil according to different decision-making means respectively to evaluate and predict the state of the transformer, and obtain multiple decision-making evaluation results respectively; The fault determination module is used to count multiple decision-making evaluation results. When the decision-making evaluation results are repeated more than twice, the repeated decision-making evaluation results are output; when the decision-making evaluation results are repeated less than or equal to twice, the partial discharge signal corresponding to the oil sample period is obtained. If there is a partial discharge signal, fusion discrimination is performed based on the amplitude and frequency characteristics of the partial discharge signal and multiple decision-making evaluation results, and the discharge fault determination result is output; If there is no partial discharge signal, fusion discrimination is performed based on the generation rate of characteristic gases and multiple decision-making evaluation results, and the overheating fault determination result is output; The fault classification module is used to fuse the discharge fault determination result or the overheating fault determination result with the ratio of carbon monoxide to carbon dioxide concentrations in the oil sample to judge whether the insulation system involved in the current transformer state is a pure oil insulation system or an oil-paper insulation system, and complete the fault classification of the transformer operation state.

10. The integrated prediction and evaluation system for transformer status based on decision-level cascade fusion analysis according to claim 9, characterized in that: A variety of fault diagnosis methods based on dissolved gas analysis in oil at least include: key characteristic gas concentration analysis method, Duval triangle diagram diagnosis method, International Electrotechnical Commission gas ratio method, Rogers gas ratio method, Dornenburg gas ratio method, these five diagnosis methods.

11. The integrated prediction and evaluation system for transformer status based on decision-level cascade fusion analysis according to claim 9, wherein: The dissolved characteristic gases in the oil at least include H2, CH4, C2H6, C2H4, C2H2, CO, CO2.

12. The comprehensive prediction and evaluation method for transformer status based on decision-level cascade fusion analysis according to claim 9, characterized in that: The generation rate of characteristic gases is calculated by the following formula: Among them, v represents the absolute gas production rate; c i2 represents the concentration of a certain characteristic gas in the oil measured from the second sampling of the insulating oil; c i1 represents the concentration of a certain characteristic gas in the oil measured from the first sampling of the insulating oil; Δt represents the actual operating time interval between the first sampling and the second sampling; m represents the total mass of the transformer insulating oil; ρ represents the density of the transformer insulating oil.

13. The comprehensive prediction and evaluation system for transformer status based on decision-level cascade fusion analysis according to claim 9, characterized in that: Fusion discrimination based on the amplitude and frequency characteristics of the partial discharge signal and multiple decision-making evaluation results includes: if the amplitude or frequency characteristics of the partial discharge signal exceed the preset threshold and at least three diagnosis results indicate the existence of a discharge fault, the discharge fault determination result is output; if the amplitude or frequency characteristics of the partial discharge signal do not reach the preset threshold or less than three of the multiple diagnosis results indicate the existence of a discharge fault, comprehensive judgment is made on the remaining evaluation information according to the preset weighted voting or priority rules. If the comprehensive result still indicates the existence of a discharge behavior, the discharge fault determination result is output, otherwise it is determined that the current fault type does not belong to the discharge fault and the corresponding diagnosis result is output.

14. The transformer status comprehensive prediction and evaluation system based on decision-level cascade fusion analysis according to claim 9, characterized in that: Fusion discrimination based on the generation rate of characteristic gases and multiple decision-making evaluation results includes: if the generation rate of characteristic gases exceeds the preset overheating fault threshold and at least three of the multiple diagnosis results indicate the existence of a thermal fault type, it is determined as an overheating fault and the corresponding overheating fault determination result is output; if the generation rate of characteristic gases does not exceed the threshold or less than three of the multiple diagnosis results indicate a thermal fault, comprehensive judgment is made on the remaining evaluation results according to the weighted voting or priority rules. If the comprehensive result still points to a thermal fault, the overheating fault determination result is output, otherwise it is determined that the current state does not belong to the overheating fault and the corresponding diagnosis result is output.

15. The transformer status comprehensive prediction and evaluation system based on decision-level cascade fusion analysis according to claim 9, characterized in that: Fusing the discharge fault determination result with the ratio of carbon monoxide to carbon dioxide concentrations in the oil sample for judgment includes: when the overheating fault or discharge fault determination result indicates that the current operating state is abnormal, judging the type of insulation system corresponding to the fault based on a preset CO / CO2 concentration ratio threshold. If the actual CO / CO2 concentration ratio is lower than the threshold, it is determined as a pure oil insulation system fault; if the actual CO / CO2 concentration ratio is equal to or higher than the threshold, it is determined as an oil-paper insulation system fault; integrating the aforementioned fault type information with the discharge fault determination result or overheating fault determination result to obtain a complete classification of the transformer operating state faults and output corresponding diagnostic conclusions.

16. The integrated prediction and evaluation system for transformer status based on decision-level cascade fusion analysis according to claim 10, characterized in that: The process of respectively importing the concentration data of different characteristic gas components in the oil sample into five diagnostic methods, namely the key characteristic gas concentration analysis method, the Duval triangle diagram diagnosis method, the International Electrotechnical Commission gas ratio method, the Rogers gas ratio method, and the Dornenburg gas ratio method, for evaluating and predicting the transformer state by different decision-making means, and respectively obtaining five decision-making evaluation results includes: Obtaining the concentration parameters of dissolved gases in the oil, including but not limited to H2, CH4, C2H6, C2H4, C2H2, CO, and CO2; In the key characteristic gas concentration analysis method, initially distinguishing between discharge faults and overheating faults by identifying the relative proportions of H2, CH4, C2H6, C2H4, C2H2, and CO; In the Duval triangle diagram diagnosis method, projecting the percentages of three combustible gases, CH4, C2H4, and C2H2, onto the Duval triangle diagram, and judging the fault type according to the area where the point falls within the triangle; In the International Electrotechnical Commission gas ratio method, calculating the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio according to the gas ratio coding rules specified by the IEC standard, classifying them, and outputting the corresponding fault types; In the Rogers gas ratio method, combining the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, and C2H4 / C2H6 concentration ratio and matching them with a predefined interval to determine a discharge fault or overheating fault; In the Dornenburg gas ratio method, obtaining the evaluation result of the transformer fault type by comparing the CH4 / H2 concentration ratio, C2H2 / C2H4 concentration ratio, C2H2 / CH4 concentration ratio, and C2H6 / C2H2 concentration ratio with the corresponding criteria; Recording and summarizing the fault evaluation outputs of the above five diagnostic methods to obtain five independent decision-making evaluation results.

17. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps of the transformer state comprehensive prediction and evaluation method based on decision-level cascade fusion analysis as described in any one of claims 1 to 8.

Images