A cooling response verification control method for abnormal operation of an oil-immersed transformer

Abstract

This invention relates to a cooling response verification and control method for abnormal operation of an oil-immersed transformer. The technical problem this invention aims to solve is to provide a cooling response verification and control method for abnormal operation of an oil-immersed transformer. The technical solution adopted by this invention includes: the method is executed by a local monitoring terminal, temperature controller, or edge controller of the oil-immersed transformer. The oil-immersed transformer is equipped with a cooling device and an online hydrogen monitoring device. The method includes the following steps: collecting operating status data of the oil-immersed transformer; performing time synchronization processing on the operating status data to form a current detection window, a pre-verification window, and a historical stable window. The advantages of this invention are: through the control chain of "abnormal triggering—first cooling adjustment—temperature response verification—hydrogen trend verification—dual-coefficient decoupling—graded load limiting—threshold and weight feedback correction," the accuracy of cooling control, load limiting control, and abnormal early warning can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of transformer operation monitoring and cooling control, and particularly relates to a cooling response verification control method for operation abnormality of an oil-immersed transformer. BACKGROUND

[0002] The oil-immersed transformer is an important device for power transformation and power distribution in the power system, and is widely used in transformer substations, industrial and mining enterprises, new energy booster stations and power distribution networks. During the operation of the oil-immersed transformer, the internal insulating oil and solid insulating materials are in a combined environment of electric field, thermal field and mechanical stress for a long time. When the load continues to rise, the ambient temperature rises, the cooling device performance declines or the internal insulation state deteriorates, the top layer oil temperature rises, local overheating occurs, and the gas content in the oil changes, etc. If the above abnormalities cannot be identified and handled in time, it may further develop into accelerated insulation aging, discharge fault, winding overheating and even transformer shutdown accident. The existing oil-immersed transformer is usually equipped with a temperature monitoring device, a load acquisition device and a cooling control device. The temperature monitoring device is used to acquire data such as top layer oil temperature, winding temperature or ambient temperature; the load acquisition device is used to acquire operation data such as load current, load rate or active power; and the cooling control device is used to control the start-stop or gear switching of cooling devices such as cooling fans and oil pumps according to the oil temperature, load or preset operation strategy. Through the above method, the temperature rise of the transformer can be inhibited to a certain extent, and the transformer can be prevented from being in a high temperature operation state for a long time. At the same time, with the development of online monitoring technology of oil gas, some oil-immersed transformers are also equipped with online hydrogen monitoring devices or online oil gas monitoring devices. Hydrogen H2 is one of the more important fault characteristic gases in transformer oil. When there is a partial discharge, low-energy discharge, local overheating or insulating oil deterioration in the transformer, the hydrogen concentration in the oil and its growth trend may change. Therefore, in the prior art, the latent fault in the transformer is often judged by the hydrogen concentration in the oil, the hydrogen growth rate or the hydrogen change trend.

[0003] However, in actual operation, the transformer temperature change, load change, cooling device response and oil hydrogen growth are not a simple and synchronous relationship. After the transformer load rises, the top layer oil temperature usually lags behind and rises due to thermal inertia; after the cooling device starts or increases the gear, the top layer oil temperature needs a certain time to show a downward trend; and the change of the hydrogen concentration in the oil is affected by the oil temperature, oil circulation, fault type, gas diffusion and sampling period of the monitoring device, and its change speed is usually slower than that of the temperature data and load data. As can be seen, if only the oil temperature data or hydrogen data at a single moment is used for judgment, the abnormal source identification may be inaccurate.

[0004] For example, when the transformer load is short-time increased, the top oil temperature may be increased accordingly, but the temperature rise may belong to the normal thermal load response and does not necessarily represent that there is a fault inside the transformer. If the system only outputs a high-level alarm or a load limiting instruction according to that the oil temperature exceeds the threshold at this time, it may cause false alarm or unnecessary load limiting. For another example, when the hydrogen growth trend in the oil is enhanced, the existing system usually directly judges it as a gas abnormal risk, but the relationship between the hydrogen growth and the temperature rise, load, and cooling state is not further analyzed, and it is difficult to judge whether it is caused by thermal load change or by discharge, local overheating, and other non-thermal factors.

[0005] In addition, the existing cooling control method usually takes the cooling device as a simple temperature reducing execution component, and the control logic is mostly to start cooling when the oil temperature reaches the threshold, or to adjust the cooling gear according to the load rate and the oil temperature change. Such a method focuses on the control effect of the cooling action on the temperature, and less uses the temperature response result and the hydrogen trend change result after the cooling action to make a reverse judgment on the abnormal source. In other words, the cooling device in the existing technology is usually only an object of “temperature reduction execution”, and is not fully used as an active response means of “abnormal source verification”.

[0006] In some actual scenarios, even if the cooling device has started or increased the operation gear, the top oil temperature may still not appear obvious decrease. The reasons for causing this situation may include that the cooling fan or oil pump does not actually act according to the instruction, the cooling device efficiency is decreased, the ambient temperature is too high, the load is continuously increased, or there is a more serious compound degradation risk inside the transformer. If the existing technology directly increases the alarm level according to that the oil temperature does not decrease, it is often difficult to distinguish whether the cooling response is insufficient or the abnormality inside the transformer is aggravated. Similarly, if the top oil temperature has obviously decreased after the cooling adjustment, but the hydrogen growth slope is still not slowed down, or the hydrogen growth deviation continues to expand, it may indicate that the hydrogen abnormality is not caused by the thermal load increase alone, but may be related to the non-thermal abnormality. The existing method usually lacks a subdivision judgment mechanism for the above-mentioned situations.

[0007] Furthermore, the sampling periods of temperature, load, cooling state, and hydrogen concentration data are usually not consistent. The load data, top oil temperature data, and cooling state data are generally collected at a high frequency, while the online hydrogen monitoring device is affected by the detection principle and the oil-gas balance process, and the sampling period is usually relatively long. If the data of different sampling periods are not time-synchronized and effectively maintained, it may cause that the risk judgment is made by using expired hydrogen data, and then the load limiting or alarm is mistakenly triggered.

[0008] Furthermore, existing methods may be affected by changes in cooling system settings, sudden load changes, and temperature transitions when determining historical stability benchmarks. For example, when the cooling system switches from off to low-speed operation or from low-speed to high-speed operation, the top oil temperature and gas trends in the oil will undergo transitional changes. If the data from this transitional phase is directly used as historical stability window data in the calculation of the hydrogen benchmark growth slope, it may distort the judgment of hydrogen growth deviation and affect the subsequent anomaly identification results.

[0009] In summary, while existing monitoring and control methods for oil-immersed transformers can achieve temperature monitoring, hydrogen monitoring, cooling control, and alarm output, they still have the following shortcomings: First, existing methods mostly rely on direct judgment based on oil temperature thresholds, load thresholds, or hydrogen thresholds, making it difficult to distinguish between different sources of anomalies such as normal thermal load response, non-thermal hydrogen anomalies, insufficient cooling response, and combined degradation. Second, existing cooling control is mainly used to reduce oil temperature, lacking the technical design to use controlled cooling actions as a verification process for anomaly sources, and cannot fully utilize the temperature response and hydrogen trend changes after cooling adjustment for reverse judgment. Third, existing methods usually lack a phased processing mechanism for temperature response and hydrogen trend response, making it difficult to adapt to the actual operating characteristics of rapid changes in top-layer oil temperature and slow changes in hydrogen trend. Fourth, existing methods are prone to triggering alarms or load limits based on expired hydrogen data when the hydrogen data sampling period is long or the data is not updated in a timely manner. Fifth, existing methods mostly use fixed rules for load limit ratios, alarm levels, and subsequent judgment thresholds, making it difficult to dynamically adjust based on whether cooling is effective, whether the hydrogen trend is deteriorating, and whether the cooling device is actually executing commands. Summary of the Invention

[0010] To address the above problems, the technical problem to be solved by the present invention is to provide a cooling response verification and control method for abnormal operation of oil-immersed transformers.

[0011] The technical solution adopted by the cooling response verification and control method for abnormal operation of oil-immersed transformer of the present invention is characterized by the following: the method is executed by the local monitoring terminal, temperature controller or edge controller of the oil-immersed transformer, the oil-immersed transformer is equipped with a cooling device and an online hydrogen monitoring device, and the method includes the following steps: S1 collecting the operating status data of the oil-immersed transformer, the operating status data including load data, top oil temperature data, ambient temperature data, cooling device operating status data and hydrogen (H2) concentration data in the oil obtained by the online hydrogen monitoring device; S2 performs time synchronization processing on the operating status data to form a current detection window, a pre-verification window, and a historical stable window. It then calculates initial anomaly judgment parameters based on the current detection window, the pre-verification window, and the historical stable window. The initial anomaly judgment parameters include load change rate, oil temperature rise change rate, hydrogen growth slope, and hydrogen growth deviation. S3 determines whether the cooling response verification trigger condition is met based on the initial anomaly judgment parameters. The cooling response verification trigger condition includes a temperature rise trigger condition and a hydrogen trigger condition. When the oil temperature rise rate exceeds a preset temperature rise threshold, the temperature rise trigger condition is met. When the hydrogen growth slope exceeds a preset hydrogen slope threshold, or the hydrogen growth deviation exceeds a preset hydrogen deviation threshold, the hydrogen trigger condition is met. When either the temperature rise trigger condition or the hydrogen trigger condition is met, the cooling response verification trigger condition is determined to be met. When the cooling response verification trigger condition is met, S4 sends a first cooling adjustment command to the cooling device. When the cooling device is in the off state, low-level cooling is started; when the cooling device is in the running state but not at the highest level, the cooling intensity is increased by one level based on the current running intensity; when the cooling device is already at the highest level, the current cooling intensity is maintained, and the cooling device is marked as being in full-load cooling state. S5 starts the cooling response verification window and collects cooling response data within the cooling response verification window. The cooling response data includes the drop in top oil temperature after cooling adjustment, the rate of drop in top oil temperature, the actual operating status of the cooling device, the change in hydrogen growth slope, and the change in hydrogen growth deviation. S6 generates anomaly decoupling types based on the cooling response data, specifically including: When the temperature drop of the top oil reaches the preset effective temperature drop threshold, and the change in hydrogen growth slope is not greater than the preset stable slope threshold and the change in hydrogen growth deviation is not greater than the preset stable deviation threshold, a heat load anomaly is generated. When the temperature drop of the top oil reaches the preset effective temperature drop threshold, and the change in the hydrogen growth slope is greater than the preset slope deterioration threshold, or the change in the hydrogen growth deviation is greater than the preset deviation deterioration threshold, non-thermal hydrogen abnormality is generated. When the cooling device has executed the first cooling adjustment command, if the drop in top oil temperature does not reach the preset effective temperature drop threshold, and the change in hydrogen growth slope is not greater than the preset slope stability threshold and the change in hydrogen growth deviation is not greater than the preset deviation stability threshold, an insufficient cooling response anomaly is generated. When the cooling device has executed the first cooling adjustment command, if the drop in top oil temperature does not reach the preset effective temperature drop threshold, and the change in hydrogen growth slope is greater than the preset slope deterioration threshold, or the change in hydrogen growth deviation is greater than the preset deviation deterioration threshold, a compound deterioration anomaly is generated. S7 generates a corresponding hierarchical control strategy based on the aforementioned anomaly decoupling type, specifically including: When the abnormal decoupling type is a thermal load type abnormality, maintain enhanced cooling and enter the observation period. When the abnormal decoupling type is a non-thermal hydrogen abnormality, reduce the allowable load limit and output a hydrogen abnormality warning. When the abnormal decoupling type is a cooling response inadequate type abnormality, output a cooling device verification prompt and reduce the allowable load limit. When the abnormal decoupling type is a compound deterioration type abnormality, reduce the allowable load limit and simultaneously output a hydrogen abnormality warning, a cooling device verification prompt, and a maintenance prompt. Based on the changes in top-level oil temperature, hydrogen growth trend, and actual operating status of the cooling device after the implementation of the graded control strategy, S8 makes corresponding corrections to the temperature rise judgment threshold, hydrogen slope threshold, hydrogen deviation threshold, cooling response verification trigger condition, or abnormal decoupling judgment weight for the next detection cycle.

[0012] The cooling device includes a cooling fan, an oil pump, or a combination thereof, for regulating the top oil temperature of the oil-immersed transformer; The online hydrogen monitoring device is selected from the following devices: a single online hydrogen monitoring device, a multi-gas online monitoring device, or an online gas analysis device that includes a hydrogen detection channel in oil, used to obtain the hydrogen concentration in oil and its changing trend.

[0013] The load data includes one or more of load current, load rate, and active power; The operating status data of the cooling device includes one or more of the following: cooling fan start / stop status, cooling fan operating speed, oil pump start / stop status, oil pump operating frequency, and cooling device command feedback status. The online hydrogen monitoring device is either a single online hydrogen monitoring device or an online oil gas monitoring device that includes a hydrogen detection channel; The cooling device command feedback status is used to determine whether the first cooling adjustment command is actually executed by the cooling device. When the first cooling adjustment command is not actually executed by the cooling device, the abnormal decoupling type is preferentially corrected to the insufficient cooling response type abnormality.

[0014] In step S2, the time synchronization process includes: The load data, top oil temperature data, ambient temperature data, and cooling device operating status data are synchronized according to the first sampling period; The H2 concentration data in the oil was synchronized according to the second sampling cycle; When the second sampling period is longer than the first sampling period, the load data, top oil temperature data, ambient temperature data and cooling device operating status data in the first sampling period are aggregated into the corresponding second sampling period and then correlated with the hydrogen (H2) concentration data in the oil. If no new H2 concentration data in the oil is obtained after the preset effective holding time has expired, it is prohibited to trigger the load limit control solely based on the hydrogen growth slope or hydrogen growth deviation, and the cooling response verification trigger condition is limited to the temperature rise trigger condition.

[0015] The rate of change of oil temperature rise is calculated based on the change in temperature difference between the top oil temperature and the ambient temperature; The hydrogen growth slope is calculated based on the changing trend of the hydrogen (H2) concentration in the oil over time within the current detection window. The hydrogen growth deviation is the amount by which the hydrogen growth slope in the current detection window deviates from the hydrogen baseline growth slope in the historical stable window. The historical stability window is a historical time interval that simultaneously meets the following conditions: the load change rate is less than the preset load stability threshold, the oil temperature rise change rate is less than the preset temperature rise stability threshold, the cooling device operating status does not experience a sudden change in gear, and the hydrogen growth slope does not exceed the preset hydrogen stability threshold. When the operating status of the cooling device changes abruptly within the historical stable window, the transitional data before and after the change is discarded, and the transitional data is not used as the basis for calculating the hydrogen reference growth slope.

[0016] The cooling response verification window includes a temperature response sub-window and a hydrogen trend response sub-window set in series. The temperature response sub-window starts timing from the moment the first cooling adjustment command is actually executed by the cooling device, and is used to calculate the drop in top oil temperature and the rate of drop in top oil temperature. The hydrogen trend response sub-window starts timing after the end of the temperature response sub-window and is used to calculate the change in hydrogen growth slope and the change in hydrogen growth deviation. The change in hydrogen growth slope is the change in the hydrogen growth slope within the hydrogen trend response sub-window relative to the hydrogen growth slope within the pre-verification window. The change in hydrogen growth deviation is the change in hydrogen growth deviation within the hydrogen trend response sub-window relative to the hydrogen growth deviation within the pre-verification window.

[0017] The duration of the temperature response sub-window is determined based on the rated capacity of the oil-immersed transformer, the type of cooling device, and the ambient temperature. The duration of the hydrogen trend response sub-window is determined based on the sampling cycle of the online hydrogen monitoring device and the hydrogen growth slope within the pre-verification window. When the sampling period of the online hydrogen monitoring device increases, the hydrogen trend response sub-window is extended; When the hydrogen growth slope in the pre-verification window exceeds the preset rapid growth threshold, the hydrogen trend response sub-window is shortened, and the judgment priority of non-thermal hydrogen anomalies and compound deterioration anomalies is increased.

[0018] In step S6, the cooling effectiveness coefficient and hydrogen deterioration coefficient are further calculated; The effective cooling coefficient is calculated based on the drop in top oil temperature, the rate of drop in top oil temperature, and the actual operating status of the cooling device. The hydrogen deterioration coefficient is calculated based on the change in the hydrogen growth slope and the change in the hydrogen growth deviation. When the cooling effectiveness coefficient reaches the preset cooling effectiveness threshold and the hydrogen deterioration coefficient does not reach the preset hydrogen deterioration threshold, a heat load anomaly is generated. When the cooling effectiveness coefficient reaches the preset cooling effectiveness threshold and the hydrogen deterioration coefficient reaches the preset hydrogen deterioration threshold, non-thermal hydrogen is generated abnormally. When the cooling effectiveness coefficient does not reach the preset cooling effectiveness threshold and the hydrogen deterioration coefficient does not reach the preset hydrogen deterioration threshold, an insufficient cooling response anomaly is generated. When the cooling effectiveness coefficient fails to reach the preset cooling effectiveness threshold and the hydrogen deterioration coefficient reaches the preset hydrogen deterioration threshold, a compound deterioration anomaly is generated.

[0019] In step S7, different abnormal decoupling types correspond to different load limiting ratios; When the abnormal decoupling type is thermal load type abnormality, the allowable load limit shall not be reduced during the observation period, or the allowable load limit shall be reduced according to the first load limit ratio. When the abnormal decoupling type is non-thermal hydrogen abnormality, the allowable load limit is reduced according to the second load limit ratio. When the abnormal decoupling type is insufficient cooling response, the allowable load limit is reduced according to the third load limit ratio. When the abnormal decoupling type is a composite deterioration type abnormality, the allowable load limit is reduced according to the fourth load limit ratio. The first load limit ratio, the second load limit ratio, the third load limit ratio, and the fourth load limit ratio are determined based on the cooling effectiveness coefficient and the hydrogen deterioration coefficient, and the first load limit ratio is less than the second load limit ratio, the second load limit ratio is less than or equal to the third load limit ratio, and the third load limit ratio is less than the fourth load limit ratio.

[0020] In step S8, the abnormal decoupling judgment weights include temperature rise judgment weights, hydrogen judgment weights, and cooling response judgment weights. When the abnormal decoupling type is heat load type abnormality, and the cooling effective coefficient reaches the preset cooling effective threshold and the hydrogen deterioration coefficient does not reach the preset hydrogen deterioration threshold in multiple consecutive detection cycles, the temperature rise judgment threshold of the next detection cycle is increased and the hydrogen judgment weight is reduced. When the abnormal decoupling type is non-thermal hydrogen abnormality, and the cooling effectiveness coefficient reaches the preset cooling effectiveness threshold and the hydrogen deterioration coefficient reaches the preset hydrogen deterioration threshold, the hydrogen slope threshold and hydrogen deviation threshold of the next detection cycle are reduced, and the hydrogen judgment weight is increased. When the abnormal decoupling type is insufficient cooling response, and the actual operating status of the cooling device is inconsistent with the first cooling adjustment command, the weight of the cooling response judgment is increased, and an abnormal cooling device execution prompt is output. When the abnormal decoupling type is a composite deterioration type, the temperature rise judgment threshold, hydrogen slope threshold and hydrogen deviation threshold of the next detection cycle are reduced, and the hydrogen judgment weight and cooling response judgment weight are increased.

[0021] The advantages of the cooling response verification and control method for abnormal operation of oil-immersed transformers in this invention are as follows: through the control chain of "abnormal triggering - first cooling adjustment - temperature response verification - hydrogen trend verification - dual coefficient decoupling - graded load limiting - threshold and weight feedback correction", the method realizes the detailed identification and targeted handling of abnormal operation of oil-immersed transformers, which can improve the accuracy of cooling control, load limiting control and abnormal early warning, and reduce the possibility of false alarms, false load limiting and misjudgment of fault sources. Attached Figure Description

[0022] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0023] Figure 1 The overall flowchart of the cooling response verification control method for abnormal operation of oil-immersed transformers according to the present invention is shown below. Figure 2 This is a schematic diagram of the runtime status data acquisition and local execution architecture in this invention; Figure 3 This is a schematic diagram showing the time relationship between the current detection window, the pre-verification window, the historical stability window, the temperature response sub-window, and the hydrogen trend response sub-window in this invention. Figure 4 This is a flowchart of the cooling response verification trigger and the generation of the first cooling adjustment command in this invention; Figure 5 This is a schematic diagram illustrating the cascade verification of the temperature response sub-window and the hydrogen trend response sub-window in this invention; Figure 6 A logic diagram is generated for calculating the cooling effectiveness coefficient, hydrogen deterioration coefficient, and abnormal decoupling type in this invention. Figure 7 This is a diagram showing the relationship between the abnormal decoupling type, hierarchical control strategy, and correction for the next detection cycle in this invention. Detailed Implementation

[0024] like Figures 1-7As shown, this embodiment provides a cooling response verification and control method for abnormal operation of an oil-immersed transformer. This method is applied to oil-immersed transformers equipped with a cooling device and an online hydrogen monitoring device. The method is executed by a local monitoring terminal located at the transformer site. The local monitoring terminal is communicatively connected to a load acquisition device, a top-level oil temperature sensor, an ambient temperature sensor, a cooling device control unit, and an online hydrogen monitoring device, respectively, and is used to collect operating status data, calculate abnormal judgment parameters, send a first cooling adjustment command, and generate a graded control strategy.

[0025] In this embodiment, the cooling device uses a cooling fan, which has four operating states: off, low speed, medium speed, and high speed. The online hydrogen monitoring device is a single online hydrogen monitoring device used to acquire hydrogen (H2) concentration data in the oil.

[0026] I. Operational Status Data Acquisition The local monitoring terminal collects operating status data of the oil-immersed transformer according to a preset sampling period. This operating status data includes load data, top oil temperature data, ambient temperature data, cooling system operating status data, and hydrogen (H2) concentration data in the oil. Each type of operating status data is marked with a sampling time stamp for subsequent time synchronization and window division.

[0027] In this embodiment, the load data uses load current and load rate. The load current is obtained from the transformer's outgoing current transformer or smart meter, and the load rate is calculated based on the current load current and the transformer's rated current. Top oil temperature data is obtained from a temperature sensor installed on top of the transformer tank. Ambient temperature data is obtained from an ambient temperature sensor located near the transformer. Cooling device operating status data includes the cooling fan's start / stop status, cooling fan operating speed, and cooling device command feedback status. Hydrogen (H2) concentration data in the oil is obtained from an online hydrogen monitoring device.

[0028] In this embodiment, load data, top oil temperature data, ambient temperature data, and cooling device operating status data are collected according to a first sampling period of 1 minute. Hydrogen (H2) concentration data in the oil is collected according to a second sampling period of 10 minutes. Since the online hydrogen monitoring device's detection response is typically slower than the temperature and load data collection process, the second sampling period is longer than the first sampling period.

[0029] During data acquisition, the local monitoring terminal marks abnormal data. When sensor communication is interrupted, sampling time is missing, data exceeds the device's range, or multiple consecutive sampling points maintain abnormal fixed values, the corresponding data is marked as data to be reviewed. Data to be reviewed is not the sole basis for triggering load limit control, but is used to output data review prompts.

[0030] II. Time Synchronization and Window Division The local monitoring terminal performs time synchronization processing on the collected operational status data and forms the current detection window, the pre-verification window, and the historical stable window on the synchronized time series.

[0031] In this embodiment, the current detection window is set to a time interval extending 30 minutes backward from the current judgment time. The current detection window is used to calculate the load change rate, oil temperature rise rate, hydrogen growth slope, and hydrogen growth deviation, and to determine whether the current operating state meets the cooling response verification trigger conditions.

[0032] The pre-verification window is set to the 30-minute time interval before the first cooling adjustment command is issued. The pre-verification window is used to record the oil temperature change trend and hydrogen growth trend before cooling adjustment, and serves as a benchmark for subsequent calculations of the change in hydrogen growth slope and the change in hydrogen growth deviation.

[0033] The historical stability window is selected from historical operating data stored on the local monitoring terminal. The historical stability window is used to determine the baseline hydrogen growth slope of the oil-immersed transformer under relatively stable operating conditions. The historical stability window meets the following conditions: the load change rate is less than a preset load stability threshold; the oil temperature rise change rate is less than a preset temperature rise stability threshold; there is no sudden change in the cooling fan's operating speed; and the hydrogen growth slope does not exceed a preset hydrogen stability threshold.

[0034] Since the first and second sampling periods are different, the local monitoring terminal uses the second sampling period as the synchronization period for judging the hydrogen trend. Specifically, within each 10-minute synchronization period, the average value of the top oil temperature data and the ambient temperature data within that synchronization period is taken, the average value and the maximum value of the load data are taken, the actual feedback status of the cooling fan operation status is taken, and then a correspondence is established with the corresponding H2 concentration data in the oil for that synchronization period.

[0035] If no new H2 concentration data in the oil is obtained after a preset effective holding time, the local monitoring terminal determines that the current hydrogen data is not suitable as a basis for triggering load limiting control alone. In this embodiment, the preset effective holding time is set to 20 minutes. If no new H2 concentration data in the oil is obtained after 20 minutes, triggering load limiting control solely based on the hydrogen growth slope or hydrogen growth deviation is prohibited, and the cooling response verification trigger condition is limited to the temperature rise trigger condition.

[0036] When screening historical stable windows, if there is a sudden change in the operating speed of the cooling fan, the transitional data within 10 minutes before and after the change is discarded and not used as the basis for calculating the hydrogen baseline growth slope. The sudden change in operating speed includes the cooling fan switching from off to low speed, from low to medium speed, from medium to high speed, and from high to lower speed. By discarding the transitional data before and after the cooling speed change, the impact of unstable temperature changes and hydrogen trend changes during the cooling switching process on the historical stable baseline can be avoided, thereby improving the accuracy of hydrogen growth deviation calculation.

[0037] Through the above-mentioned data collection, time synchronization and window division of the operating status, this embodiment can form a unified time series under different data sampling periods, and provide a reliable data foundation for subsequent cooling response verification triggering, first cooling adjustment command execution, temperature response sub-window calculation and hydrogen trend response sub-window calculation.

[0038] III. Calculation of Initial Anomaly Detection Parameters After completing the acquisition of operational status data, time synchronization, and window division, the local monitoring terminal calculates initial anomaly judgment parameters based on the current detection window, the pre-verification window, and the historical stable window. These initial anomaly judgment parameters include the load change rate, oil temperature rise rate, hydrogen growth slope, and hydrogen growth deviation.

[0039] In this embodiment, the current detection window is set to 30 minutes, the pre-verification window is set to 30 minutes before the first cooling adjustment command is issued, and the second sampling period is set to 10 minutes. Therefore, within a current detection window, the local monitoring terminal obtains at least three synchronous data points corresponding to the H2 concentration data in the oil. Each synchronous data point includes the load data, top oil temperature data, ambient temperature data, cooling device operating status data, and H2 concentration data in the oil at the corresponding time.

[0040] First, the local monitoring terminal calculates the oil temperature rise. For any given synchronization moment, the oil temperature rise TD is calculated using the following formula: TD = T1 - T0 Where T1 represents the top oil temperature at the synchronization moment, T0 represents the ambient temperature at the synchronization moment, and TD represents the oil temperature rise at the synchronization moment.

[0041] In this embodiment, the oil temperature rise value TD is used for subsequent judgment, instead of just the absolute value of the top oil temperature T1. Since the top oil temperature is affected by changes in the external ambient temperature, using the difference between the top oil temperature and the ambient temperature can reduce the interference of ambient temperature fluctuations on the temperature rise judgment, making the oil temperature rise rate more reflective of the thermal state changes of the oil-immersed transformer body.

[0042] The load change rate is calculated based on the trend of load data over time within the current detection window. In this embodiment, the load data is the load rate, denoted as L. The local monitoring terminal performs linear fitting between the load rate L at each synchronization moment within the current detection window and the corresponding time to obtain the load change rate RL. RL is used to represent the trend of load increase or decrease within the current detection window. When RL is greater than the preset load change threshold RLth, it indicates that the load within the current detection window shows a significant upward trend.

[0043] The oil temperature rise rate is calculated based on the trend of the oil temperature rise value TD over time within the current detection window. The local monitoring terminal performs linear fitting between the oil temperature rise value TD at each synchronous moment within the current detection window and the corresponding time to obtain the oil temperature rise rate RT. RT is used to represent the trend of oil temperature rise within the current detection window. When RT is greater than the preset temperature rise change threshold RTth, it indicates that the oil temperature rise within the current detection window shows a significant upward trend.

[0044] The hydrogen growth slope is calculated based on the change trend of hydrogen (H2) concentration in the oil over time within the current detection window. Let H be the hydrogen (H2) concentration in the oil at the corresponding synchronization time. The local monitoring terminal performs linear fitting between the hydrogen concentration H at each synchronization time within the current detection window and the corresponding time to obtain the hydrogen growth slope K1 within the current detection window. K1 represents the growth trend of hydrogen (H2) concentration in the oil within the current detection window.

[0045] The historical stability window is used to determine the baseline hydrogen growth slope. Let K0 be the baseline hydrogen growth slope within the historical stability window. K0 is calculated from the hydrogen concentration change trends corresponding to multiple valid synchronous data points within the historical stability window. The historical stability window only uses data intervals where the load change rate, oil temperature rise rate, cooling device operating status, and hydrogen growth slope all meet the stability conditions. Transitional data before and after sudden changes in cooling device settings are not included in the K0 calculation.

[0046] The hydrogen growth deviation is used to represent the degree of enhancement of the current hydrogen growth trend relative to the historical steady state. In this embodiment, the hydrogen growth deviation D1 is calculated according to the following formula: D1 = K1 - K0 Where K1 represents the hydrogen growth slope in the current detection window, K0 represents the hydrogen baseline growth slope in the historical stable window, and D1 represents the hydrogen growth deviation in the current detection window.

[0047] When D1 is greater than 0, it indicates that the current hydrogen growth trend is higher than the historical stable growth trend; when D1 is greater than the preset hydrogen deviation threshold Dth, it indicates that the current hydrogen (H2) growth trend in the oil is abnormally enhanced relative to the stable operating state.

[0048] To avoid negative biases interfering with hydrogen risk assessment, when D1 is less than or equal to 0, the local monitoring terminal records D1 as 0 in the hydrogen risk assessment. In other words, this embodiment focuses on assessing the strengthening of the hydrogen growth trend relative to the steady state, rather than using a decreasing hydrogen growth trend as the trigger for hydrogen risk.

[0049] Through the above calculations, the local monitoring terminal obtains the load change rate RL, oil temperature rise change rate RT, hydrogen growth slope K1, and hydrogen growth deviation D1. These parameters respectively reflect the load change, oil temperature rise change, hydrogen growth trend, and the degree of deviation of the hydrogen growth trend from the historical stable state, providing a basis for judgment on subsequent cooling response verification triggering.

[0050] IV. Cooling Response Verification Trigger In this embodiment, the local monitoring terminal determines whether the cooling response verification triggering conditions are met based on the initial anomaly judgment parameters. The cooling response verification triggering conditions include temperature rise triggering conditions and hydrogen triggering conditions.

[0051] The temperature rise trigger condition is: the rate of change of oil temperature RT exceeds the preset temperature rise threshold RTth, that is, the following conditions are met: RT>RTth When the above conditions are met, the local monitoring terminal determines that there is an abnormal temperature rise trend within the current detection window.

[0052] The hydrogen trigger condition is: the hydrogen growth slope K1 exceeds a preset hydrogen slope threshold Kth, or the hydrogen growth deviation D1 exceeds a preset hydrogen deviation threshold Dth. In other words, the hydrogen trigger condition is satisfied when either of the following conditions is met: K1>Kth or: D1>Dth When either the temperature rise trigger condition or the hydrogen trigger condition is met, the local monitoring terminal determines that the cooling response verification trigger condition is satisfied. At this time, the local monitoring terminal does not directly determine the oil-immersed transformer as being in a fault state, nor does it immediately execute high-level load limiting control; instead, it enters the cooling response verification process.

[0053] The purpose of setting up a cooling response verification process in this embodiment is to enable the oil-immersed transformer to exhibit an observable temperature response through a limited-amplitude cooling adjustment, and then, in conjunction with the change in the hydrogen growth trend after the cooling adjustment, to further determine the source of the anomaly. This process differs from the handling method of directly alarming or limiting the load based solely on oil temperature or hydrogen thresholds.

[0054] After determining that the cooling response verification trigger condition is met, the local monitoring terminal sends a first cooling adjustment command to the cooling device control unit. The first cooling adjustment command is generated according to the current operating status of the cooling device.

[0055] When the cooling device is in the off state, the first cooling adjustment command is to start the low-speed cooling command, so that the cooling device enters the low-speed operation state from the off state.

[0056] When the cooling device is running but not at its highest setting, the first cooling adjustment command is to increase the cooling intensity by one level, causing the cooling device to switch from the current setting to an adjacent higher setting. For example, if the cooling device is running at a low setting, the first cooling adjustment command will cause it to switch to a medium setting; if the cooling device is running at a medium setting, the first cooling adjustment command will cause it to switch to a high setting.

[0057] When the cooling system is already at its highest setting, the local monitoring terminal will no longer send commands to further increase the cooling intensity. Instead, it will maintain the current cooling intensity and mark the cooling system as being in full-load cooling mode. This marking is used to subsequently determine whether the anomaly is due to insufficient cooling response or a combination of degradation.

[0058] The first cooling adjustment command is a limited-range cooling adjustment command. This embodiment does not directly force the cooling device to its maximum output after an anomaly is triggered, but instead performs a first-level adjustment based on the current operating state of the cooling device. The temperature response and hydrogen trend response obtained through the first-level adjustment can serve as the basis for subsequent anomaly decoupling judgment, thereby avoiding overcooling, false alarms, or unnecessary load limiting due to a single data anomaly.

[0059] After sending the first cooling adjustment command, the local monitoring terminal reads the cooling device command feedback status. If the cooling device command feedback status matches the first cooling adjustment command, it is determined that the first cooling adjustment command has been actually executed; if the cooling device command feedback status does not match the first cooling adjustment command, it is determined that the first cooling adjustment command has not been actually executed, and a cooling execution anomaly is recorded. Cooling execution anomalies are subsequently used to identify insufficient cooling response anomalies during anomaly decoupling.

[0060] V. Cooling Response Verification Window After the first cooling adjustment command is actually executed by the cooling device, the local monitoring terminal activates the cooling response verification window. The cooling response verification window includes a temperature response sub-window and a hydrogen trend response sub-window set in series.

[0061] The temperature response sub-window is used to observe the change in top-layer oil temperature after the execution of the first cooling adjustment command. The temperature response sub-window starts timing from the moment the first cooling adjustment command is actually executed by the cooling device. In this embodiment, the temperature response sub-window is set to 20 minutes.

[0062] Within the temperature response sub-window, the local monitoring terminal continuously collects top oil temperature data and ambient temperature data, and calculates the magnitude and rate of decrease in top oil temperature.

[0063] The temperature drop Td of the top oil layer is calculated using the following formula: Td = Ts - Te Where Ts represents the top oil temperature at the beginning of the temperature response sub-window, Te represents the top oil temperature at the end of the temperature response sub-window, and Td represents the decrease in top oil temperature.

[0064] The rate of temperature drop (Vd) of the top oil layer is calculated using the following formula: Vd = Td / Tw Where Tw represents the duration of the temperature response sub-window, and Vd represents the rate of decrease in top oil temperature.

[0065] When the temperature drop of the top layer oil, Td, reaches the preset effective temperature drop threshold, Tdth, it indicates that the first cooling adjustment command has produced an effective cooling response to the top layer oil temperature. This judgment condition is expressed as follows: Td>= Tdth When the temperature drop of the top oil layer, Td, does not reach the preset effective temperature drop threshold, Tdth, it indicates insufficient cooling response. Further analysis of the actual operating status of the cooling system and changes in hydrogen gas trends is needed to determine the source of the anomaly. This judgment condition is expressed as follows: Td <Tdth If the first cooling adjustment command is not actually executed by the cooling device, the local monitoring terminal will still record the temperature change data in the temperature response sub-window. However, in the anomaly decoupling judgment, this situation will be given priority as the basis for judging the insufficient cooling response anomaly, and a cooling device execution anomaly prompt will be output. This can avoid mistaking insufficient temperature drop caused by the cooling device not operating as internal composite deterioration of the oil-immersed transformer.

[0066] The hydrogen trend response sub-window is used to observe the hydrogen growth trend after the execution of the first cooling adjustment command. Since the change in hydrogen (H2) concentration in the oil is slower than the change in top-layer oil temperature, this embodiment sets the hydrogen trend response sub-window after the temperature response sub-window ends. That is, the local monitoring terminal starts the hydrogen trend response sub-window after the temperature response sub-window ends. In this embodiment, the hydrogen trend response sub-window is set to 30 minutes.

[0067] Within the hydrogen trend response sub-window, the local monitoring terminal calculates the hydrogen growth slope K2 and hydrogen growth deviation D2 within the hydrogen trend response sub-window based on the H2 concentration data in the oil obtained by the online hydrogen monitoring device.

[0068] Where K2 represents the slope of the linear fit of the hydrogen (H2) concentration in oil over time within the hydrogen trend response sub-window. D2 represents the deviation of K2 from the baseline hydrogen growth slope K0 within the historical stable window. D2 is calculated according to the following formula: D2 = K2 - K0 Where K2 represents the hydrogen growth slope within the hydrogen trend response sub-window, K0 represents the hydrogen baseline growth slope within the historical stable window, and D2 represents the hydrogen growth deviation within the hydrogen trend response sub-window.

[0069] When D2 is less than or equal to 0, the local monitoring terminal will record D2 as 0 in the hydrogen risk assessment.

[0070] The change in hydrogen growth slope, DK, represents the change in the hydrogen growth slope after cooling response verification relative to the pre-verification window. In this embodiment, the hydrogen growth slope within the pre-verification window is denoted as Kpre, and the change in hydrogen growth slope, DK, is calculated according to the following formula: DK = K2 - Kpre Where K2 represents the hydrogen growth slope within the hydrogen trend response sub-window, Kpre represents the hydrogen growth slope within the pre-validation window, and DK represents the change in the hydrogen growth slope.

[0071] The change in hydrogen growth deviation DD represents the change in hydrogen growth deviation after cooling response verification relative to the pre-verification window. In this embodiment, the hydrogen growth deviation within the pre-verification window is denoted as Dpre, and the change in hydrogen growth deviation DD is calculated according to the following formula: DD = D2 - Dpre Where D2 represents the hydrogen growth deviation within the hydrogen trend response sub-window, Dpre represents the hydrogen growth deviation within the pre-verification window, and DD represents the change in hydrogen growth deviation.

[0072] When DK is not greater than the preset slope stabilization threshold DKs, and DD is not greater than the preset deviation stabilization threshold DDs, the hydrogen growth trend is determined to have not continued to deteriorate. This judgment condition is expressed as follows: DK<= DKs and: DD<= DDs When DK exceeds the preset slope deterioration threshold DKb, or DD exceeds the preset deviation deterioration threshold DDb, the hydrogen growth trend is determined to continue to deteriorate. This judgment condition is expressed as follows: DK>DKb or: DD>DDb In this embodiment, the temperature response sub-window and the hydrogen trend response sub-window are set in series, rather than completely overlapping. This is because the top-layer oil temperature response is relatively fast after the cooling device is adjusted, while the change in hydrogen (H2) concentration in the oil is affected by oil-gas diffusion, oil circulation, and the detection cycle, and the trend change usually lags behind the temperature change. By observing the temperature response first and then the hydrogen trend response, it is possible to more accurately determine whether the hydrogen anomaly slows down with improved cooling, thereby improving the accuracy of distinguishing between heat load anomalies, non-thermal hydrogen anomalies, insufficient cooling response anomalies, and combined deterioration anomalies.

[0073] In other embodiments, the duration of the temperature response sub-window and the hydrogen trend response sub-window is adjusted according to the transformer capacity, cooling device type, oil circulation speed and sampling cycle of the online hydrogen monitoring device, but it should be ensured that the temperature response sub-window precedes the hydrogen trend response sub-window, and at least two valid hydrogen concentration data points can be obtained within the hydrogen trend response sub-window.

[0074] VI. Generation of Exception Decoupling Types After the cooling response verification window ends, the local monitoring terminal generates anomaly decoupling types based on the cooling response data. The cooling response data includes the top oil temperature drop magnitude Td, the top oil temperature drop rate Vd, the actual operating status of the cooling device, the change in hydrogen growth slope DK, and the change in hydrogen growth deviation DD after the execution of the first cooling adjustment command.

[0075] Among them, the decrease in top oil temperature Td is used to reflect whether the top oil temperature has been effectively improved after the first cooling adjustment command is executed; the change in hydrogen growth slope DK is used to reflect whether the growth slope of hydrogen H2 in the oil continues to increase relative to the window before verification after the cooling response verification; and the change in hydrogen growth deviation DD is used to reflect whether the growth deviation of hydrogen H2 in the oil continues to expand relative to the window before verification after the cooling response verification.

[0076] In this embodiment, the preset effective temperature drop threshold is denoted as Tdth. When Td is greater than or equal to Tdth, the temperature response after cooling adjustment is determined to be effective; when Td is less than Tdth, the temperature response after cooling adjustment is determined to be insufficient.

[0077] A preset slope stabilization threshold is denoted as DKs, and a preset slope deterioration threshold is denoted as DKb. To avoid a gap in the judgment between the stability and deterioration judgments of the hydrogen growth slope change DK, this embodiment uses the same numerical boundary for DKs and DKb. That is, when DK is not greater than DKs, it is determined that the hydrogen growth slope has not continued to deteriorate; when DK is greater than DKb, it is determined that the hydrogen growth slope continues to deteriorate.

[0078] The preset deviation stabilization threshold is denoted as DDs, and the preset deviation deterioration threshold is denoted as DDb. To avoid a gap in the judgment between the stability and deterioration judgments for the hydrogen growth deviation change DD, DDs and DDb use the same numerical boundary in this embodiment. That is, when DD is not greater than DDs, it is determined that the hydrogen growth deviation has not continued to deteriorate; when DD is greater than DDb, it is determined that the hydrogen growth deviation continues to deteriorate.

[0079] The local monitoring terminal generates four types of anomaly decoupling based on the temperature response and hydrogen trend response results. The specific judgment process is as follows.

[0080] First, the local monitoring terminal generates a heat load type anomaly when the following conditions are met: Td>= Tdth; DK <= DKs; DD <= DDs.

[0081] The above conditions indicate that after the first cooling adjustment command is executed, the temperature drop of the top layer oil reaches the preset effective temperature drop threshold, indicating that the cooling adjustment effectively improves the top layer oil temperature. Simultaneously, the changes in the hydrogen growth slope and the hydrogen growth deviation do not exceed the stability boundary, indicating that the hydrogen (H2) growth trend in the oil has not continued to worsen. Therefore, this type of anomaly mainly manifests as a thermal response caused by increased load or environmental thermal influences, and the local monitoring terminal classifies it as a heat load-type anomaly.

[0082] Second, the local monitoring terminal generates a non-thermal hydrogen gas anomaly when the following conditions are met: Td>= Tdth; And it satisfies at least one of the conditions DK>DKb or DD>DDb.

[0083] The above conditions indicate that after the first cooling adjustment command was executed, the temperature drop of the top layer oil had reached the preset effective temperature drop threshold, indicating that the cooling adjustment was effective in temperature control; however, the change in the hydrogen growth slope or the change in the hydrogen growth deviation still exceeded the corresponding deterioration boundary, indicating that the hydrogen (H2) growth trend in the oil did not slow down with the improvement of temperature. At this time, the hydrogen anomaly should not be simply attributed to the increase in heat load, and the local monitoring terminal classifies it as a non-thermal hydrogen anomaly.

[0084] Third, the local monitoring terminal generates an insufficient cooling response anomaly when the following conditions are met: The cooling device has executed the first cooling adjustment command; Td <Tdth; DK <= DKs; DD <= DDs.

[0085] The above conditions indicate that the first cooling adjustment command has been actually executed by the cooling device, but the temperature drop of the top oil layer has not reached the preset effective temperature drop threshold, indicating that the temperature drop effect after cooling adjustment is insufficient; at the same time, the growth trend of hydrogen (H2) in the oil has not continued to worsen. Therefore, this type of anomaly mainly reflects factors such as insufficient heat dissipation capacity of the cooling device, decreased cooling efficiency, excessively high ambient temperature, or strong load heat release. The local monitoring terminal classifies it as an anomaly of insufficient cooling response.

[0086] Fourth, the local monitoring terminal generates a composite degradation anomaly when the following conditions are met: The cooling device has executed the first cooling adjustment command; Td <Tdth; And it satisfies at least one of the conditions DK>DKb or DD>DDb.

[0087] The above conditions indicate that the first cooling adjustment command has been actually executed by the cooling device, but the temperature drop of the top oil layer has not reached the preset effective temperature drop threshold, indicating insufficient cooling response. At the same time, the growth slope or deviation of hydrogen (H2) in the oil continues to worsen, indicating that the abnormal hydrogen trend is still intensifying. At this time, insufficient cooling response and deteriorating hydrogen trend coexist, and the local monitoring terminal classifies it as a compound deterioration anomaly. In this embodiment, if the cooling device command feedback status indicates that the first cooling adjustment command was not actually executed by the cooling device, the local monitoring terminal prioritizes recording the cooling execution anomaly and corrects the anomaly decoupling direction to a cooling response insufficiency anomaly. Simultaneously, the system outputs a cooling device execution anomaly prompt to remind maintenance personnel to review the cooling fan, oil pump, control circuit, power supply line, or communication feedback status. This priority processing rule is used to avoid misjudging insufficient temperature drop caused by the cooling device not actually operating as internal transformer degradation.

[0088] In the above four types of anomaly decoupling judgments, the drop in top oil temperature, Td, is used to distinguish between "effective cooling" and "insufficient cooling," while the change in hydrogen growth slope, DK, and the change in hydrogen growth deviation, DD, are used to distinguish between "hydrogen trend not deteriorating" and "hydrogen trend continuing to deteriorate." This leads to the following correspondence: When Td reaches Tdth and neither DK nor DD deteriorates, a heat load-type anomaly is generated. When Td reaches Tdth and at least one of DK and DD deteriorates, the generation of nonthermal hydrogen gas is abnormal. When Td does not reach Tdth and neither DK nor DD deteriorates, an anomaly of insufficient cooling response is generated. When Td does not reach Tdth and at least one of DK and DD deteriorates, a compound deterioration anomaly is generated.

[0089] VII. Cooling efficiency coefficient and hydrogen deterioration coefficient To improve the stability of the anomaly decoupling judgment, this embodiment further calculates the cooling effectiveness coefficient and the hydrogen deterioration coefficient in step S6. The cooling effectiveness coefficient is used to characterize the actual improvement of the top oil temperature by the cooling device after the execution of the first cooling adjustment command; the hydrogen deterioration coefficient is used to characterize whether the hydrogen (H2) growth trend in the oil continues to deteriorate after the cooling response verification.

[0090] In this embodiment, the effective cooling coefficient is denoted as C, and the hydrogen deterioration coefficient is denoted as G. The larger the effective cooling coefficient C, the more significant the improvement effect of the first cooling adjustment command on the top oil temperature; the larger the hydrogen deterioration coefficient G, the more the hydrogen (H2) growth trend in the oil still tends to worsen after the cooling response verification.

[0091] The effective cooling coefficient C is calculated based on the top oil temperature drop amplitude Td, the top oil temperature drop rate Vd, and the actual operating conditions of the cooling system. The calculation process is as follows: First, calculate the temperature drop evaluation value C1 based on the temperature drop amplitude Td of the top oil layer and the preset effective temperature drop threshold Tdth: C1 = Td / Tdth When C1 is greater than 1, C1 is set to 1; when C1 is less than 0, C1 is set to 0. C1 is used to indicate the degree to which the temperature drop of the top oil reaches the preset effective temperature drop threshold.

[0092] Secondly, the cooling rate evaluation value C2 is calculated based on the top oil temperature drop rate Vd and the preset cooling rate threshold Vdth: C2 = Vd / Vdth When C2 is greater than 1, C2 is set to 1; when C2 is less than 0, C2 is set to 0. C2 is used to indicate the degree to which the rate of temperature drop of the top layer oil reaches the preset cooling rate threshold.

[0093] Next, the cooling execution status value C3 is determined based on whether the actual operating status of the cooling device is consistent with the first cooling adjustment command. If the actual operating status of the cooling device is consistent with the first cooling adjustment command, then C3 is set to 1; if the actual operating status of the cooling device is inconsistent with the first cooling adjustment command, then C3 is set to 0.

[0094] In this embodiment, the cooling effectiveness coefficient C is calculated according to the following formula: C = 0.45*C1 + 0.35*C2 + 0.20*C3 Wherein, C1 represents the temperature drop evaluation value, C2 represents the cooling rate evaluation value, and C3 represents the cooling execution status value. The above weighting settings ensure that the top oil temperature drop magnitude, top oil temperature drop rate, and actual cooling device performance all participate in the cooling effectiveness judgment, avoiding the reliance on a single data point—the top oil temperature—to determine the effectiveness of cooling.

[0095] In this embodiment, the preset effective cooling threshold is denoted as Cth. When C is greater than or equal to Cth, it is determined that the first cooling adjustment command generates an effective cooling response; when C is less than Cth, it is determined that the cooling response is insufficient. The value of Cth is 0.70.

[0096] The hydrogen deterioration coefficient G is calculated based on the change in the hydrogen growth slope DK and the change in the hydrogen growth deviation DD. The calculation process is as follows: First, the slope deterioration evaluation value G1 is calculated based on the change in hydrogen growth slope DK and the preset slope deterioration threshold DKb: G1 = DK / DKb When G1 is greater than 1, G1 is set to 1; when G1 is less than 0, G1 is set to 0. G1 is used to indicate the extent to which the hydrogen growth slope continues to increase relative to the pre-validation window.

[0097] Secondly, the deviation deterioration evaluation value G2 is calculated based on the change in hydrogen growth deviation DD and the preset deviation deterioration threshold DDb: G2 = DD / DDb When G2 is greater than 1, G2 is set to 1; when G2 is less than 0, G2 is set to 0. G2 is used to represent the extent to which the hydrogen growth deviation continues to expand relative to the pre-verification window.

[0098] In this embodiment, the hydrogen deterioration coefficient G is calculated according to the following formula: G = 0.60*G1 + 0.40*G2 Wherein, G1 represents the slope deterioration evaluation value, and G2 represents the deviation deterioration evaluation value. Since the hydrogen growth slope can directly reflect the increasing trend of hydrogen (H2) concentration in oil over time, G1 has a greater weight than G2 in this embodiment.

[0099] In this embodiment, a preset hydrogen deterioration threshold is denoted as Gth. When G is greater than or equal to Gth, it is determined that the hydrogen (H2) growth trend in the oil continues to worsen; when G is less than Gth, it is determined that the hydrogen (H2) growth trend in the oil does not continue to worsen. The value of Gth is 0.50.

[0100] The local monitoring terminal generates the abnormal decoupling type based on the cooling effectiveness coefficient C and the hydrogen deterioration coefficient G. The specific judgment rules are as follows: When C is greater than or equal to Cth and G is less than Gth, it indicates that the top oil temperature has improved significantly after the first cooling adjustment command is executed, and the hydrogen growth trend has not continued to deteriorate, resulting in a heat load anomaly.

[0101] When C is greater than or equal to Cth and G is greater than or equal to Gth, it indicates that the top oil temperature has improved significantly after the first cooling adjustment command is executed, but the hydrogen growth trend still deteriorates and the generation of non-thermal hydrogen is abnormal.

[0102] When C is less than Cth and G is less than Gth, it indicates that the top oil temperature is not improved enough after the first cooling adjustment command is executed, but the hydrogen growth trend does not continue to worsen, resulting in an anomaly of insufficient cooling response.

[0103] When C is less than Cth and G is greater than or equal to Gth, it indicates that the top oil temperature improvement is insufficient after the first cooling adjustment command is executed, and the hydrogen growth trend continues to deteriorate, resulting in a compound deterioration anomaly.

[0104] In this embodiment, the cooling response effect and hydrogen trend change are converted into a cooling effectiveness coefficient C and a hydrogen deterioration coefficient G, respectively. Then, an anomaly decoupling type is generated based on the combination of C and G, providing the anomaly judgment process with clear data sources, calculation procedures, and discrimination rules. This method avoids directly classifying faults based solely on oil temperature or hydrogen thresholds, improving the accuracy of anomaly source identification.

[0105] VIII. Hierarchical Control Strategy The local monitoring terminal generates corresponding hierarchical control strategies based on the anomaly decoupling type. These hierarchical control strategies include maintaining enhanced cooling, reducing allowable load limits, issuing early warnings for abnormal hydrogen output, providing verification prompts for the output cooling device, and issuing maintenance prompts for the output.

[0106] In this embodiment, the permissible load limit is denoted as Pmax, the rated permissible load of the transformer is denoted as Pr, and the load limiting ratio is denoted as Q. After implementing load limiting control, the new permissible load limit Pmax is determined according to the following formula: Pmax = Pr*(1 - Q) Where Pr represents the rated allowable load of the transformer, Q represents the load limiting ratio, and Pmax represents the allowable load limit after the load limiting control is implemented.

[0107] When the abnormal decoupling type is a heat load type abnormality, it indicates that after the execution of the first cooling adjustment command, the cooling effectiveness coefficient C reaches the preset cooling effectiveness threshold Cth, and the hydrogen deterioration coefficient G does not reach the preset hydrogen deterioration threshold Gth. At this time, the local monitoring terminal maintains the current enhanced cooling state and enters the observation cycle. The observation cycle is set to 30 minutes. During the observation cycle, the local monitoring terminal continues to collect top oil temperature data and hydrogen (H2) concentration data in the oil. When the top oil temperature continues to decrease during the observation cycle and the hydrogen deterioration coefficient G is still less than Gth, the load limit ratio Q is set to 0; when the top oil temperature does not continue to decrease during the observation cycle, the load limit ratio Q is set to the first load limit ratio Q1. Q1 is set to 5%.

[0108] When the abnormal decoupling type is non-thermal hydrogen anomaly, it indicates that after the execution of the first cooling adjustment command, the cooling effectiveness coefficient C reaches the preset cooling effectiveness threshold Cth, but the hydrogen deterioration coefficient G reaches the preset hydrogen deterioration threshold Gth. At this time, the oil temperature has improved, but the hydrogen growth trend is still deteriorating, indicating that the hydrogen anomaly should not be simply attributed to changes in heat load. The local monitoring terminal reduces the allowable load limit according to the second load limit ratio Q2 and outputs a hydrogen anomaly warning. Q2 is determined according to the hydrogen deterioration coefficient G, and in this embodiment, it is taken according to the following rules: when G is greater than or equal to Gth and less than 0.75, Q2 is 10%; when G is greater than or equal to 0.75, Q2 is 15%.

[0109] When the abnormal decoupling type is insufficient cooling response, it indicates that after the first cooling adjustment command is executed, the cooling effectiveness coefficient C has not reached the preset cooling effectiveness threshold Cth, and the hydrogen deterioration coefficient G has not reached the preset hydrogen deterioration threshold Gth. At this time, the local monitoring terminal outputs a cooling device verification prompt and reduces the allowable load limit according to the third load limit ratio Q3. Q3 is determined based on the cooling effectiveness coefficient C, and in this embodiment, it is taken according to the following rules: when C is greater than or equal to 0.50 and less than Cth, Q3 is 10%; when C is less than 0.50, Q3 is 20%. The cooling device verification prompt includes whether the cooling fan is started, whether the cooling fan operation level has reached the command level, whether the oil pump operation status is normal, whether the cooling power supply is abnormal, and whether the heat dissipation channel is blocked.

[0110] When the abnormal decoupling type is a composite deterioration anomaly, it indicates that after the execution of the first cooling adjustment command, the cooling effectiveness coefficient C has not reached the preset cooling effectiveness threshold Cth, and the hydrogen deterioration coefficient G has reached the preset hydrogen deterioration threshold Gth. At this time, there is a risk of insufficient cooling response and hydrogen trend deterioration occurring simultaneously. The local monitoring terminal reduces the allowable load limit according to the fourth load limit ratio Q4, and simultaneously outputs a hydrogen anomaly warning, a cooling device verification prompt, and a maintenance prompt. Q4 is determined based on the cooling effectiveness coefficient C and the hydrogen deterioration coefficient G. In this embodiment, it is valued according to the following rules: when C is greater than or equal to 0.50 and G is less than 0.75, Q4 is 20%; when C is less than 0.50 or G is greater than or equal to 0.75, Q4 is 30%; when C is less than 0.50 and G is greater than or equal to 0.75, Q4 is 40%.

[0111] The first load limit ratio Q1, the second load limit ratio Q2, the third load limit ratio Q3, and the fourth load limit ratio Q4 mentioned above satisfy the following relationship: Q1 <Q2 Q2 <= Q3 Q3 <Q4 Through the aforementioned graded control strategy, this embodiment matches the load limit range with the source of the anomaly, the effectiveness of cooling, and the degree of hydrogen deterioration. For heat load anomalies, a large load limit is not directly implemented to avoid unnecessary load losses caused by normal heat load response; for non-thermal hydrogen anomalies, even if the temperature has improved, the hydrogen anomaly handling level is still increased; for anomalies with insufficient cooling response, the cooling device is specifically prompted to be rechecked and the load is appropriately limited; for anomalies with complex deterioration, a higher level of load limit is implemented and multiple warning prompts are issued simultaneously.

[0112] IX. Threshold and Weight Adjustment for the Next Detection Period In order to adapt the control method to the long-term operating conditions of the transformer, this embodiment modifies the judgment threshold and abnormal decoupling judgment weight for the next detection cycle based on the top oil temperature change, hydrogen growth trend change and actual operating status of the cooling device after the implementation of the hierarchical control strategy.

[0113] The anomaly decoupling judgment weights include temperature rise judgment weight, hydrogen gas judgment weight, and cooling response judgment weight. The temperature rise judgment weight is denoted as W1, the hydrogen gas judgment weight as W2, and the cooling response judgment weight as W3. In this embodiment, the initial values ​​of W1, W2, and W3 are 0.35, 0.35, and 0.30, respectively, and they satisfy the following relationship: W1 + W2 + W3 = 1 Among them, W1 is used to characterize the influence of the rate of change of oil temperature on the judgment of abnormal decoupling, W2 is used to characterize the influence of the hydrogen growth slope and hydrogen growth deviation on the judgment of abnormal decoupling, and W3 is used to characterize the influence of the cooling effectiveness coefficient and the actual operating state of the cooling device on the judgment of abnormal decoupling.

[0114] When the anomaly decoupling type is a heat load anomaly, and the cooling effectiveness coefficient C is greater than or equal to Cth and the hydrogen deterioration coefficient G is less than Gth for three consecutive detection cycles, it indicates that this type of anomaly is mainly caused by heat load changes, and the cooling device responds effectively. In this case, the local monitoring terminal increases the temperature rise judgment threshold RTth for the next detection cycle by 5% and decreases the hydrogen judgment weight W2 by 0.05. The reduced 0.05 is allocated to the temperature rise judgment weight W1 and the cooling response judgment weight W3, with W1 increasing by 0.03 and W3 increasing by 0.02. After the adjustment, W1, W2, and W3 still satisfy W1 + W2 + W3 = 1.

[0115] When the abnormal decoupling type is non-thermal hydrogen anomaly, and C is greater than or equal to Cth and G is greater than or equal to Gth, it indicates that the cooling action is effective in improving the temperature, but the hydrogen trend is still deteriorating. In this case, the local monitoring terminal reduces the hydrogen slope threshold Kth by 5% and the hydrogen deviation threshold Dth by 5% for the next detection cycle, and increases the hydrogen judgment weight W2 by 0.05; the increased 0.05 is deducted from the temperature rise judgment weight W1. After the adjustment, W1, W2, and W3 still satisfy W1 + W2 + W3 = 1.

[0116] When the anomaly decoupling type is insufficient cooling response, and the actual operating status of the cooling device is inconsistent with the first cooling adjustment command, it indicates that the cooling device has an execution anomaly or a feedback anomaly. In this case, the local monitoring terminal increases the cooling response judgment weight W3 by 0.05 and outputs a cooling device execution anomaly prompt; the increased 0.05 is deducted from the temperature rise judgment weight W1. This prompt is used to remind maintenance personnel to check the cooling fan, oil pump, control circuit, power supply line, and heat dissipation channel.

[0117] When the anomaly decoupling type is a composite degradation anomaly, it indicates that insufficient cooling response and deteriorating hydrogen trend coexist. In this case, the local monitoring terminal reduces the temperature rise judgment threshold RTth, the hydrogen slope threshold Kth, and the hydrogen deviation threshold Dth by 5% for the next detection cycle, while simultaneously increasing the hydrogen judgment weight W2 and the cooling response judgment weight W3. Specifically, W2 is increased by 0.03, and W3 is increased by 0.02, with the increased portion deducted from the temperature rise judgment weight W1. After the adjustment, W1, W2, and W3 still satisfy W1 + W2 + W3 = 1.

[0118] To prevent excessive shifts in thresholds and weights during continuous operation, this embodiment sets boundary conditions for threshold and weight corrections. The single correction magnitude of the temperature rise judgment threshold RTth, hydrogen slope threshold Kth, and hydrogen deviation threshold Dth does not exceed 5% of their original values, and the cumulative shift after continuous corrections does not exceed 20% of the initial set values. The temperature rise judgment weight W1, hydrogen judgment weight W2, and cooling response judgment weight W3 are all maintained between 0.20 and 0.50.

[0119] Through the aforementioned threshold and weight correction methods, this embodiment can adjust the judgment conditions for the next cycle based on the control results of the previous cycle. For recurring normal heat load responses, the system reduces the probability of hydrogen false triggering; for situations where hydrogen deteriorates even after temperature improvement, the system increases the sensitivity to hydrogen anomalies; for malfunctions in the cooling device, the system increases the impact of cooling response judgment; for compound deterioration anomalies, the system simultaneously increases the priority of identifying both hydrogen trends and insufficient cooling response. This makes the cooling control, load limiting control, and anomaly early warning of oil-immersed transformers more consistent with actual operating conditions.

[0120] When an oil-immersed transformer exhibits an abnormal temperature rise trend or an abnormal hydrogen growth trend, this invention does not directly rely on oil temperature thresholds or hydrogen thresholds for alarm, load limiting, or fault determination. Instead, it first sends a first cooling adjustment command to the cooling device, causing it to perform limited cooling adjustments. The actual response after the cooling adjustment is then used to determine the source of the anomaly. Thus, the cooling device not only functions as a cooling actuator but also as an active response mechanism in the anomaly source verification process, transforming transformer anomaly judgment from static threshold judgment to a closed-loop judgment process of "trigger—cooling verification—response analysis—tiered control."

[0121] This invention, by setting a cooling response verification window and further dividing it into a temperature response sub-window and a hydrogen trend response sub-window, can respectively acquire the top oil temperature drop magnitude, top oil temperature drop rate, hydrogen growth slope change, and hydrogen growth deviation change after the execution of the first cooling adjustment command. Since the top oil temperature typically responds to cooling actions faster than the hydrogen trend change in the oil, this invention employs a series window approach—observing the temperature response first and then the hydrogen trend response—which better reflects the actual thermal response and gas change patterns of oil-immersed transformers.

[0122] This invention calculates the cooling effectiveness coefficient C based on the drop in top-layer oil temperature, the rate of drop in top-layer oil temperature, and the actual operating state of the cooling device. It also calculates the hydrogen deterioration coefficient G based on the change in the hydrogen growth slope and the change in the hydrogen growth deviation. This allows both the cooling response effect and the degree of hydrogen deterioration to be expressed using comparable quantitative parameters. This approach avoids relying solely on a single temperature value, a single hydrogen concentration value, or a simple growth rate to make fault judgments, thus improving the stability and interpretability of the anomaly judgment process.

[0123] This invention classifies abnormal operation of oil-immersed transformers into four categories based on the combined relationship between the effective cooling coefficient C and the hydrogen deterioration coefficient G: thermal load-related abnormalities, non-thermal hydrogen-related abnormalities, insufficient cooling response-related abnormalities, and combined deterioration-related abnormalities. Compared with existing technologies that treat mixed issues such as oil temperature rise, hydrogen growth, or insufficient cooling, this invention can more accurately distinguish the different sources of abnormalities, thereby avoiding misjudging thermal responses caused by normal load increases as internal faults, and also avoiding misjudging insufficient cooling device response as combined internal deterioration of the transformer.

[0124] This invention maintains enhanced cooling and enters an observation period under thermal load anomalies; outputs a hydrogen anomaly warning and reduces the allowable load limit under non-thermal hydrogen anomalies; outputs a cooling device review prompt and implements load limiting under insufficient cooling response anomalies; and simultaneously outputs a hydrogen anomaly warning, a cooling device review prompt, and a maintenance prompt under combined degradation anomalies. Therefore, this invention enables the control strategy to match the source of the anomaly, avoiding the crude "alarm upon trigger" approach of existing methods.

[0125] This invention sets different load limiting ratios based on different anomaly decoupling types, and progressively increases the first, second, third, and fourth load limiting ratios according to the degree of risk. This setting can reduce unnecessary load reduction while ensuring the safe operation of the transformer. For thermal load anomalies, this invention prioritizes enhanced cooling and observation strategies; for non-thermal hydrogen anomalies, anomalies with insufficient cooling response, and complex deterioration anomalies, the load limiting range is increased step by step according to the degree of risk, thereby improving the rationality of load limiting control.

[0126] This invention determines whether the first cooling adjustment command has been actually executed by the cooling device by collecting the feedback status of the cooling device's commands. When the actual operating status of the cooling device is inconsistent with the first cooling adjustment command, this invention can prioritize identifying insufficient cooling response anomalies and output a cooling device execution anomaly prompt. This method can avoid the system misjudging a serious internal transformer fault due to abnormal operation of the cooling fan, oil pump, control circuit, or power supply line.

[0127] This invention performs time-synchronized processing on load data, top oil temperature data, ambient temperature data, cooling device operating status data, and hydrogen (H2) concentration data in the oil. Furthermore, if hydrogen data has not been updated within a preset effective retention period, it prohibits triggering load limiting control solely based on the hydrogen growth slope or hydrogen growth deviation. This method resolves the issue of inconsistent sampling periods for temperature, load, cooling status, and hydrogen concentration, reducing the risk of false alarms or incorrect load limiting due to the use of expired hydrogen data.

[0128] This invention determines the baseline hydrogen growth slope by setting a historical stability window, and discards transitional data before and after a sudden change in the cooling device's operating level, excluding this transitional data from the calculation of the baseline hydrogen growth slope. This avoids interference from unstable data during cooling level switching phases in the calculation of hydrogen growth deviation, improving the accuracy of judging abnormal hydrogen trends.

[0129] After completing the graded control, this invention corrects the temperature rise judgment threshold, hydrogen slope threshold, hydrogen deviation threshold, and abnormal decoupling judgment weight for the next detection cycle based on changes in top-layer oil temperature, hydrogen growth trend, and the actual operating status of the cooling device. Through this feedback correction method, this invention can adjust the judgment conditions for the next cycle based on the control results of the previous cycle, allowing the system to gradually adapt to the actual operating conditions of the transformer and reducing misjudgments caused by fixed thresholds.

[0130] By implementing a control chain of "abnormal triggering - first cooling regulation - temperature response verification - hydrogen trend verification - dual coefficient decoupling - graded load limiting - threshold and weight feedback correction", the system achieves detailed identification and targeted handling of abnormal operation of oil-immersed transformers. This improves the accuracy of cooling control, load limiting control and abnormal early warning, and reduces the possibility of false alarms, false load limiting and misjudgment of fault sources.

[0131] 10. Specific Application Examples Taking an oil-immersed transformer as an example, its on-site configuration includes a top-layer oil temperature sensor, an ambient temperature sensor, a load current acquisition unit, a cooling fan control unit, and an online hydrogen monitoring device. The local monitoring terminal collects load data, top-layer oil temperature data, ambient temperature data, and cooling device operating status data every minute, and acquires hydrogen (H2) concentration data in the oil every 10 minutes. The current detection window is set to 30 minutes, the pre-verification window is set to 30 minutes before the first cooling adjustment command is issued, the temperature response sub-window is set to 20 minutes, and the hydrogen trend response sub-window is set to 30 minutes.

[0132] In this embodiment, the preset temperature rise change threshold RTth is set to 0.03 degrees Celsius / minute, the preset hydrogen slope threshold Kth is set to 0.020 ppm / minute, the preset hydrogen deviation threshold Dth is set to 0.015 ppm / minute, the preset effective temperature drop threshold Tdth is set to 1.0 degrees Celsius, the preset cooling rate threshold Vdth is set to 0.04 degrees Celsius / minute, the preset effective cooling threshold Cth is set to 0.70, and the preset hydrogen deterioration threshold Gth is set to 0.50. The preset slope deterioration threshold DKb is set to 0.015 ppm / minute, and the preset deviation deterioration threshold DDb is set to 0.010 ppm / minute.

[0133] The effective cooling coefficient C is calculated according to the following formula: C = 0.45*C1 + 0.35*C2 + 0.20*C3 Wherein, C1 is the temperature drop evaluation value, C1 = Td / Tdth; C2 is the cooling rate evaluation value, C2 = Vd / Vdth; and C3 is the cooling execution status value. When C1 or C2 is greater than 1, it is counted as 1; when C1 or C2 is less than 0, it is counted as 0. When the actual operating status of the cooling device is consistent with the first cooling adjustment command, C3 is set to 1; when the actual operating status of the cooling device is inconsistent with the first cooling adjustment command, C3 is set to 0.

[0134] The hydrogen degradation factor G is calculated using the following formula: G = 0.60*G1 + 0.40*G2 Wherein, G1 is the slope deterioration evaluation value, G1 = DK / DKb; G2 is the deviation deterioration evaluation value, G2 = DD / DDb. When G1 or G2 is greater than 1, it is counted as 1; when G1 or G2 is less than 0, it is counted as 0.

[0135] 1. Example of heat load-type anomaly Within a given monitoring window, the transformer load rate increases from 65% to 82%, the top oil temperature rises from 61°C to 65°C, and the ambient temperature increases from 27°C to 28°C. Therefore, the oil temperature rises from 34°C to 37°C. The current monitoring window duration is 30 minutes. Therefore, the rate of change of the oil temperature RT is: RT = (37 - 34) / 30 = 0.10 degrees Celsius / minute Since RT is greater than RTth, the local monitoring terminal determines that the temperature rise trigger condition is met. At this time, the cooling fan is in the off state, and the local monitoring terminal sends the first cooling adjustment command to make the cooling fan switch from the off state to the low-speed operation state, and starts the cooling response verification window.

[0136] Within the temperature response sub-window, the top layer oil temperature decreased from 65 degrees Celsius to 63.6 degrees Celsius, and the decrease in top layer oil temperature, Td, was: Td = 65 - 63.6 = 1.4 degrees Celsius The temperature response sub-window duration Tw is 20 minutes, and the top oil temperature drop rate Vd is: Vd = 1.4 / 20 = 0.07 degrees Celsius / minute Since Td is greater than Tdth, it indicates that the low-level cooling effectively cools the top oil temperature.

[0137] Within the pre-validation window, the hydrogen growth slope Kpre is 0.018 ppm / min, and the historical stable window's baseline hydrogen growth slope K0 is 0.010 ppm / min. Therefore, the hydrogen growth deviation Dpre within the pre-validation window is: Dpre = 0.018 - 0.010 = 0.008 ppm / min Within the hydrogen trend response sub-window, the hydrogen growth slope K2 is 0.015 ppm / min, and the hydrogen growth deviation D2 is: D2 = 0.015 - 0.010 = 0.005 ppm / min Therefore, the change in the hydrogen growth slope, DK, is: DK = 0.015 - 0.018 = -0.003 ppm / min The change in hydrogen growth deviation DD is: DD = 0.005 - 0.008 = -0.003 ppm / min Since DK is less than 0 and DD is less than 0, the hydrogen growth trend has not continued to worsen.

[0138] Calculate the cooling efficiency coefficient C based on the above data: C1 = 1.4 / 1.0 = 1, count as 1; C2 = 0.07 / 0.04 = 1.75, rounded down to 1; C3 = 1; C = 0.45*1 + 0.35*1 + 0.20*1 = 1 Based on the above data, calculate the hydrogen deterioration coefficient G: G1 = -0.003 / 0.015, which is less than 0, so it is counted as 0; G2 = -0.003 / 0.010, which is less than 0, so it is counted as 0; G = 0.60*0 + 0.40*0 = 0 Since C is greater than Cth and G is less than Gth, the local monitoring terminal generates a heat load-type anomaly. This indicates that the temperature rise after the load increase can be mitigated by cooling adjustments, and the hydrogen growth trend has not worsened. Therefore, the local monitoring terminal maintains low-level enhanced cooling and enters an observation period, without immediately reducing the allowable load limit. This handling avoids misjudging the thermal response caused by a normal load increase as an internal fault.

[0139] 2. Examples of non-thermal hydrogen gas anomalies In another operating scenario, if the rate of change of oil temperature RT within the current detection window exceeds RTth, the system triggers a cooling response verification. After the first cooling adjustment command is executed, the cooling fan switches from low to medium speed operation.

[0140] Within the temperature response sub-window, the top oil temperature decreased from 66 degrees Celsius to 64.7 degrees Celsius, and the decrease in top oil temperature, Td, was: Td = 66 - 64.7 = 1.3 degrees Celsius The rate of temperature drop (Vd) of the top oil layer is: Vd = 1.3 / 20 = 0.065 degrees Celsius / minute Since Td is greater than Tdth, it indicates that the top oil temperature has been effectively improved after cooling adjustment.

[0141] Within the pre-validation window, the hydrogen growth slope Kpre was 0.020 ppm / min, and within the historical stable window, the baseline hydrogen growth slope K0 was 0.010 ppm / min. Therefore, Dpre is: Dpre = 0.020 - 0.010 = 0.010 ppm / min Within the hydrogen trend response sub-window, the hydrogen growth slope K2 is 0.040 ppm / min, and the hydrogen growth deviation D2 is: D2 = 0.040 - 0.010 = 0.030 ppm / min Therefore, DK is: DK = 0.040 - 0.020 = 0.020 ppm / min DD is: DD = 0.030 - 0.010 = 0.020 ppm / min Calculate the cooling efficiency coefficient C based on the above data: C1 = 1.3 / 1.0 = 1, count as 1; C2 = 0.065 / 0.04 = 1.625, rounded down to 1; C3 = 1; C = 0.45*1 + 0.35*1 + 0.20*1 = 1.00 Calculate the hydrogen degradation factor G: G1 = 0.020 / 0.015 = 1.33, counted as 1; G2 = 0.020 / 0.010 = 2.00, rounded down to 1; G = 0.60*1 + 0.40*1 = 1 Since C is greater than Cth and G is greater than Gth, the local monitoring terminal generates an abnormal non-thermal hydrogen gas. This indicates that cooling regulation has effectively improved the top oil temperature, but the hydrogen (H2) growth trend in the oil is still significantly worsening. This anomaly should not be simply attributed to increased heat load. The local monitoring terminal reduces the allowable load limit according to the second load limit ratio and outputs a hydrogen anomaly warning.

[0142] In this embodiment, when G is greater than or equal to 0.75, the second load limiting ratio Q2 is 15%. If the rated allowable load of the transformer is denoted as Pr, then the allowable load limit Pmax after load limiting is: Pmax = Pr*(1 - 0.15) = 0.85*Pr 3. Example of an anomaly due to insufficient cooling response In another operating scenario, the rate of change of oil temperature RT within the current detection window exceeds RTth, triggering a cooling response verification by the system. After the first cooling adjustment command is executed, the cooling fan feedback status shows that the fan has switched from the off state to low-speed operation.

[0143] Within the temperature response sub-window, the top layer oil temperature decreased from 64 degrees Celsius to 63.6 degrees Celsius, and the decrease in top layer oil temperature, Td, was: Td = 64 - 63.6 = 0.4 degrees Celsius The rate of temperature drop (Vd) of the top oil layer is: Vd = 0.4 / 20 = 0.02 degrees Celsius / minute Since Td is less than Tdth, it indicates that the cooling response is insufficient.

[0144] Within the pre-validation window, the hydrogen growth slope Kpre was 0.018 ppm / min, and within the historical stable window, the baseline hydrogen growth slope K0 was 0.010 ppm / min. Therefore, Dpre is: Dpre = 0.018 - 0.010 = 0.008 ppm / min Within the hydrogen trend response sub-window, the hydrogen growth slope K2 is 0.016 ppm / min, and the hydrogen growth deviation D2 is: D2 = 0.016 - 0.010 = 0.006 ppm / min Therefore, DK is: DK = 0.016 - 0.018 = -0.002 ppm / min DD is: DD = 0.006 - 0.008 = -0.002 ppm / min Calculate the cooling efficiency coefficient C based on the above data: C1 = 0.4 / 1.0 = 0.40; C2 = 0.02 / 0.04 = 0.50; C3 = 1; C = 0.45*0.40 + 0.35*0.50 + 0.20*1 = 0.555 Calculate the hydrogen degradation factor G: G1 = -0.002 / 0.015, which is less than 0, so it is counted as 0; G2 = -0.002 / 0.010, which is less than 0, so it is counted as 0; G = 0.60*0 + 0.40*0 = 0 Since C is less than Cth and G is less than Gth, the local monitoring terminal generates a cooling response insufficiency anomaly. This indicates that the hydrogen growth trend has not continued to worsen, but the top oil temperature drop after the cooling fan starts is insufficient. The system outputs a cooling device verification prompt and reduces the allowable load limit according to the third load limit ratio. The cooling device verification prompt includes checking whether the cooling fan is actually operating at low speed, whether the fan power supply is abnormal, whether the heat dissipation channel is blocked, and whether the temperature control execution circuit is normal.

[0145] In this embodiment, when C is greater than or equal to 0.50 and less than Cth, the third load limiting ratio Q3 is 10%. Therefore, the allowable load limit Pmax after load limiting is: Pmax = Pr*(1 - 0.10) = 0.90*Pr 4. Examples of composite degradation anomalies Under further operating conditions, if the rate of change of oil temperature RT within the current detection window exceeds RTth, and the hydrogen growth deviation D1 exceeds Dth, the system triggers a cooling response verification. After the first cooling adjustment command is executed, the cooling fan switches from low to medium speed operation.

[0146] Within the temperature response sub-window, the top layer oil temperature decreased from 67 degrees Celsius to 66.7 degrees Celsius, and the decrease in top layer oil temperature, Td, was: Td = 67 - 66.7 = 0.3 degrees Celsius The rate of temperature drop (Vd) of the top oil layer is: Vd = 0.3 / 20 = 0.015 degrees Celsius / minute Since Td is less than Tdth, it indicates that the cooling response is insufficient.

[0147] Within the pre-validation window, the hydrogen growth slope Kpre was 0.020 ppm / min, and within the historical stable window, the baseline hydrogen growth slope K0 was 0.010 ppm / min. Therefore, Dpre is: Dpre = 0.020 - 0.010 = 0.010 ppm / min Within the hydrogen trend response sub-window, the hydrogen growth slope K2 is 0.045 ppm / min, and the hydrogen growth deviation D2 is: D2 = 0.045 - 0.010 = 0.035 ppm / min Therefore, DK is: DK = 0.045 - 0.020 = 0.025 ppm / min DD is: DD = 0.035 - 0.010 = 0.025 ppm / min Calculate the cooling efficiency coefficient C based on the above data: C1 = 0.3 / 1.0 = 0.30; C2 = 0.015 / 0.04 = 0.375; C3 = 1; C = 0.45*0.30 + 0.35*0.375 + 0.20*1 = 0.466 Calculate the hydrogen degradation factor G: G1 = 0.025 / 0.015 = 1.67, rounded down to 1; G2 = 0.025 / 0.010 = 2.50, rounded down to 1; G = 0.60*1 + 0.40*1 = 1 Since C is less than Cth and G is greater than Gth, the local monitoring terminal generates a compound deterioration anomaly. This indicates that after the first cooling adjustment command was executed, the top oil temperature was not effectively improved, and the hydrogen (H2) growth trend in the oil continued to worsen, posing a risk of insufficient cooling response combined with internal anomalies. The local monitoring terminal reduces the allowable load limit according to the fourth load limit ratio, and simultaneously outputs a hydrogen anomaly warning, a cooling device verification prompt, and a maintenance prompt.

[0148] In this embodiment, since C is less than 0.50 and G is greater than 0.75, the fourth load limiting ratio Q4 is set to 40%. Therefore, the allowable load limit Pmax after load limiting is: Pmax = Pr*(1 - 0.40) = 0.60*Pr As can be seen from the four operating scenarios described above, this embodiment does not directly trigger an alarm or limit the load after the oil temperature rises or hydrogen levels increase. Instead, it first executes a first cooling adjustment command with a limited range, then uses a temperature response sub-window and a hydrogen trend response sub-window to obtain the temperature response and hydrogen trend response after cooling, respectively. Furthermore, it decouples and judges the source of the anomaly using the cooling effectiveness coefficient C and the hydrogen deterioration coefficient G. This method can distinguish between heat load anomalies, non-thermal hydrogen anomalies, insufficient cooling response anomalies, and combined deterioration anomalies, ensuring that cooling control, load limiting ratios, hydrogen warnings, cooling device verification prompts, and maintenance prompts match the actual source of the anomaly, thereby improving the accuracy and reliability of the oil-immersed transformer's operation control.

[0149] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention are included within the protection scope of the present invention.

Claims

1. A cooling response verification and control method for abnormal operation of an oil-immersed transformer, characterized in that: The method is executed by a local monitoring terminal, temperature controller or edge controller of the oil-immersed transformer, the oil-immersed transformer is equipped with a cooling device and an online hydrogen monitoring device, and the method includes the following steps: S1 Collect the operating status data of the oil-immersed transformer, the operating status data includes load data, top oil temperature data, ambient temperature data, cooling device operating status data and hydrogen (H2) concentration data in the oil obtained by the online hydrogen monitoring device; S2 performs time synchronization processing on the operating status data to form a current detection window, a pre-verification window, and a historical stable window. It then calculates initial anomaly judgment parameters based on the current detection window, the pre-verification window, and the historical stable window. The initial anomaly judgment parameters include load change rate, oil temperature rise change rate, hydrogen growth slope, and hydrogen growth deviation. S3 determines whether the cooling response verification trigger condition is met based on the initial anomaly judgment parameters. The cooling response verification trigger condition includes a temperature rise trigger condition and a hydrogen trigger condition. When the oil temperature rise rate exceeds the preset temperature rise threshold, the temperature rise trigger condition is met. When the hydrogen growth slope exceeds the preset hydrogen slope threshold, or the hydrogen growth deviation exceeds the preset hydrogen deviation threshold, the hydrogen trigger condition is met. When either the temperature rise trigger condition or the hydrogen trigger condition is met, the cooling response verification trigger condition is determined to be met. When the cooling response verification trigger condition is met, S4 sends a first cooling adjustment command to the cooling device. When the cooling device is in the off state, low-level cooling is started; when the cooling device is in the running state but not at the highest level, the cooling intensity is increased by one level based on the current running intensity; when the cooling device is already at the highest level, the current cooling intensity is maintained, and the cooling device is marked as being in full-load cooling state. S5 starts the cooling response verification window and collects cooling response data within the cooling response verification window. The cooling response data includes the drop in top oil temperature after cooling adjustment, the rate of drop in top oil temperature, the actual operating status of the cooling device, the change in hydrogen growth slope, and the change in hydrogen growth deviation. S6 generates anomaly decoupling types based on the cooling response data, specifically including: When the temperature drop of the top oil reaches the preset effective temperature drop threshold, and the change in hydrogen growth slope is not greater than the preset stable slope threshold and the change in hydrogen growth deviation is not greater than the preset stable deviation threshold, a heat load anomaly is generated. When the temperature drop of the top oil reaches the preset effective temperature drop threshold, and the change in the hydrogen growth slope is greater than the preset slope deterioration threshold, or the change in the hydrogen growth deviation is greater than the preset deviation deterioration threshold, non-thermal hydrogen abnormality is generated. When the cooling device has executed the first cooling adjustment command, if the drop in top oil temperature does not reach the preset effective temperature drop threshold, and the change in hydrogen growth slope is not greater than the preset slope stability threshold and the change in hydrogen growth deviation is not greater than the preset deviation stability threshold, an insufficient cooling response anomaly is generated. When the cooling device has executed the first cooling adjustment command, if the drop in top oil temperature does not reach the preset effective temperature drop threshold, and the change in hydrogen growth slope is greater than the preset slope deterioration threshold, or the change in hydrogen growth deviation is greater than the preset deviation deterioration threshold, a compound deterioration anomaly is generated. S7 generates a corresponding hierarchical control strategy based on the aforementioned anomaly decoupling type, specifically including: When the abnormal decoupling type is a thermal load type abnormality, maintain enhanced cooling and enter the observation period. When the abnormal decoupling type is a non-thermal hydrogen abnormality, reduce the allowable load limit and output a hydrogen abnormality warning. When the abnormal decoupling type is a cooling response inadequate type abnormality, output a cooling device verification prompt and reduce the allowable load limit. When the abnormal decoupling type is a compound deterioration type abnormality, reduce the allowable load limit and simultaneously output a hydrogen abnormality warning, a cooling device verification prompt, and a maintenance prompt. Based on the changes in top-level oil temperature, hydrogen growth trend, and actual operating status of the cooling device after the implementation of the graded control strategy, S8 makes corresponding corrections to the temperature rise judgment threshold, hydrogen slope threshold, hydrogen deviation threshold, cooling response verification trigger condition, or abnormal decoupling judgment weight for the next detection cycle.

2. The cooling response verification and control method for abnormal operation of an oil-immersed transformer according to claim 1, characterized in that: The cooling device includes a cooling fan, an oil pump, or a combination thereof, for regulating the top oil temperature of the oil-immersed transformer; The online hydrogen monitoring device is selected from the following devices: a single online hydrogen monitoring device, a multi-gas online monitoring device, or an online gas analysis device that includes a hydrogen detection channel in oil, used to obtain the hydrogen concentration in oil and its changing trend.

3. The cooling response verification and control method for abnormal operation of an oil-immersed transformer according to claim 2, characterized in that: The load data includes one or more of load current, load rate, and active power; The operating status data of the cooling device includes one or more of the following: cooling fan start / stop status, cooling fan operating speed, oil pump start / stop status, oil pump operating frequency, and cooling device command feedback status. The online hydrogen monitoring device is either a single online hydrogen monitoring device or an online oil gas monitoring device that includes a hydrogen detection channel; The cooling device command feedback status is used to determine whether the first cooling adjustment command is actually executed by the cooling device. When the first cooling adjustment command is not actually executed by the cooling device, the abnormal decoupling type is preferentially corrected to the insufficient cooling response type abnormality.

4. The cooling response verification and control method for abnormal operation of an oil-immersed transformer according to claim 2, characterized in that: In step S2, the time synchronization process includes: The load data, top oil temperature data, ambient temperature data, and cooling device operating status data are synchronized according to the first sampling period; The H2 concentration data in the oil was synchronized according to the second sampling cycle; When the second sampling period is longer than the first sampling period, the load data, top oil temperature data, ambient temperature data and cooling device operating status data in the first sampling period are aggregated into the corresponding second sampling period and then correlated with the hydrogen (H2) concentration data in the oil. If no new H2 concentration data in the oil is obtained after the preset effective holding time has expired, it is prohibited to trigger the load limit control solely based on the hydrogen growth slope or hydrogen growth deviation, and the cooling response verification trigger condition is limited to the temperature rise trigger condition.

5. The cooling response verification and control method for abnormal operation of an oil-immersed transformer according to claim 2, characterized in that: The rate of change of oil temperature rise is calculated based on the change in temperature difference between the top oil temperature and the ambient temperature; The hydrogen growth slope is calculated based on the changing trend of the hydrogen (H2) concentration in the oil over time within the current detection window. The hydrogen growth deviation is the amount by which the hydrogen growth slope in the current detection window deviates from the hydrogen baseline growth slope in the historical stable window. The historical stability window is a historical time interval that simultaneously meets the following conditions: the load change rate is less than the preset load stability threshold, the oil temperature rise change rate is less than the preset temperature rise stability threshold, the cooling device operating status does not experience a sudden change in gear, and the hydrogen growth slope does not exceed the preset hydrogen stability threshold. When the operating status of the cooling device changes abruptly within the historical stable window, the transitional data before and after the change is discarded, and the transitional data is not used as the basis for calculating the hydrogen reference growth slope.

6. The cooling response verification and control method for abnormal operation of an oil-immersed transformer according to claim 2, characterized in that: The cooling response verification window includes a temperature response sub-window and a hydrogen trend response sub-window set in series. The temperature response sub-window starts timing from the moment the first cooling adjustment command is actually executed by the cooling device, and is used to calculate the drop in top oil temperature and the rate of drop in top oil temperature. The hydrogen trend response sub-window starts timing after the end of the temperature response sub-window and is used to calculate the change in hydrogen growth slope and the change in hydrogen growth deviation. The change in hydrogen growth slope is the change in the hydrogen growth slope within the hydrogen trend response sub-window relative to the hydrogen growth slope within the pre-verification window. The change in hydrogen growth deviation is the change in hydrogen growth deviation within the hydrogen trend response sub-window relative to the hydrogen growth deviation within the pre-verification window.

7. The cooling response verification and control method for abnormal operation of an oil-immersed transformer according to claim 6, characterized in that: The duration of the temperature response sub-window is determined based on the rated capacity of the oil-immersed transformer, the type of cooling device, and the ambient temperature. The duration of the hydrogen trend response sub-window is determined based on the sampling cycle of the online hydrogen monitoring device and the hydrogen growth slope within the pre-verification window. When the sampling period of the online hydrogen monitoring device increases, the hydrogen trend response sub-window is extended; When the hydrogen growth slope in the pre-verification window exceeds the preset rapid growth threshold, the hydrogen trend response sub-window is shortened, and the judgment priority of non-thermal hydrogen anomalies and compound deterioration anomalies is increased.

8. The cooling response verification and control method for abnormal operation of an oil-immersed transformer according to claim 6, characterized in that: In step S6, the cooling effectiveness coefficient and hydrogen deterioration coefficient are further calculated; The effective cooling coefficient is calculated based on the drop in top oil temperature, the rate of drop in top oil temperature, and the actual operating status of the cooling device. The hydrogen deterioration coefficient is calculated based on the change in the hydrogen growth slope and the change in the hydrogen growth deviation. When the cooling effectiveness coefficient reaches the preset cooling effectiveness threshold and the hydrogen deterioration coefficient does not reach the preset hydrogen deterioration threshold, a heat load anomaly is generated. When the cooling effectiveness coefficient reaches the preset cooling effectiveness threshold and the hydrogen deterioration coefficient reaches the preset hydrogen deterioration threshold, non-thermal hydrogen is generated abnormally. When the cooling effectiveness coefficient does not reach the preset cooling effectiveness threshold and the hydrogen deterioration coefficient does not reach the preset hydrogen deterioration threshold, an insufficient cooling response anomaly is generated. When the cooling effectiveness coefficient fails to reach the preset cooling effectiveness threshold and the hydrogen deterioration coefficient reaches the preset hydrogen deterioration threshold, a compound deterioration anomaly is generated.

9. The cooling response verification and control method for abnormal operation of an oil-immersed transformer according to claim 8, characterized in that: In step S7, different abnormal decoupling types correspond to different load limiting ratios; When the abnormal decoupling type is thermal load type abnormality, the allowable load limit shall not be reduced during the observation period, or the allowable load limit shall be reduced according to the first load limit ratio. When the abnormal decoupling type is non-thermal hydrogen abnormality, the allowable load limit is reduced according to the second load limit ratio. When the abnormal decoupling type is insufficient cooling response, the allowable load limit is reduced according to the third load limit ratio. When the abnormal decoupling type is a composite deterioration type abnormality, the allowable load limit is reduced according to the fourth load limit ratio. The first load limit ratio, the second load limit ratio, the third load limit ratio, and the fourth load limit ratio are determined based on the cooling effectiveness coefficient and the hydrogen deterioration coefficient, and the first load limit ratio is less than the second load limit ratio, the second load limit ratio is less than or equal to the third load limit ratio, and the third load limit ratio is less than the fourth load limit ratio.

10. The cooling response verification and control method for abnormal operation of an oil-immersed transformer according to claim 8, characterized in that: In step S8, the abnormal decoupling judgment weights include temperature rise judgment weights, hydrogen judgment weights, and cooling response judgment weights. When the abnormal decoupling type is heat load type abnormality, and the cooling effective coefficient reaches the preset cooling effective threshold and the hydrogen deterioration coefficient does not reach the preset hydrogen deterioration threshold in multiple consecutive detection cycles, the temperature rise judgment threshold of the next detection cycle is increased and the hydrogen judgment weight is reduced. When the abnormal decoupling type is non-thermal hydrogen abnormality, and the cooling effectiveness coefficient reaches the preset cooling effectiveness threshold and the hydrogen deterioration coefficient reaches the preset hydrogen deterioration threshold, the hydrogen slope threshold and hydrogen deviation threshold of the next detection cycle are reduced, and the hydrogen judgment weight is increased. When the abnormal decoupling type is insufficient cooling response, and the actual operating status of the cooling device is inconsistent with the first cooling adjustment command, the weight of the cooling response judgment is increased, and an abnormal cooling device execution prompt is output. When the abnormal decoupling type is a composite deterioration type, the temperature rise judgment threshold, hydrogen slope threshold and hydrogen deviation threshold of the next detection cycle are reduced, and the hydrogen judgment weight and cooling response judgment weight are increased.

Images