A transformer operation status monitoring terminal based on artificial intelligence

Abstract

The present invention relates to the technical field of online monitoring of electric power metering, and specifically to an operating state monitoring terminal for a transformer based on artificial intelligence, comprising: a data acquisition module, used for acquiring error data existing in the operating state of the transformer, the error data including temperature, current, voltage and change value of current boundary during the operating state of the transformer; a time acquisition module, used for acquiring an error time series corresponding to the error data, determining a change trend of the error time series during the operating state of the transformer, and a change amount of a predicted value; a state detection module, used for detecting the state of the transformer, determining a change coefficient of the corresponding inductance of the transformer when the error data appears, and a probability of a change in the electrical stress of the transformer; an error verification module, used for verifying the influence coefficient of the change in the electrical stress of the transformer on the actual output power of the current transformer under different error conditions; the operating reliability of the transformer is improved, and the risk of failure is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of online monitoring of electric power metering, and in particular to an artificial intelligence-based transformer operating status monitoring terminal. Background Art

[0002] As an important part of the electric energy metering device, the accuracy and reliability of the transformer's metering performance is directly related to the fairness and justice of electric energy trade settlement. Capacitor voltage transformer (CVT) is a voltage transformer that is divided by a series capacitor, and then reduced and isolated by an electromagnetic transformer. As an instrument for converting voltage, the capacitor voltage transformer can also couple the carrier frequency to the transmission line for long-distance communication, selective line high-frequency protection, remote control and other functions. Compared with conventional electromagnetic voltage transformers, capacitor voltage transformers have many advantages in terms of economy and safety in addition to their high impact insulation strength, simple manufacturing, small size, and light weight.

[0003] During the actual operation of CVT, the transformer error will exceed the measurement deviation limit during its working life due to the influence of the acquisition principle and harsh environment. Therefore, it is necessary not only to be able to accurately and quickly diagnose when the measurement error exceeds the tolerance, but also to make timely predictions on the deterioration trend of the CVT measurement error so that relevant operation and maintenance personnel can arrange inspection and maintenance work. If the degradation of the transformer status cannot be discovered in time, it will affect the operation of the power grid.

[0004] For example, Chinese patent publication number CN113821938A discloses a short-term prediction method and device for the measurement error state of a transformer. First, the error stripping method is used to preprocess the CVT error data, and the ARIMA algorithm is used to construct a trend prediction model. The transformer's own error prediction value and the additional error prediction value are added together to predict the error value in a relatively recent time period and obtain the transformer error state prediction curve.

[0005] The prior art mainly monitors the transformer by means of error frequency, predicted value, etc. when performing error analysis on the transformer. However, in this case, it is necessary to clarify the trend of the error generated by the transformer and the specific problem corresponding to the trend, and verify the relevant data corresponding to the current and voltage to determine the problem corresponding to the transformer, so as to improve the effect of transformer status monitoring. Summary of the invention

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a transformer operation status monitoring terminal based on artificial intelligence, including: a data acquisition module, used to obtain error data existing in the transformer during operation, the error data including the temperature, current, voltage and current boundary change value of the transformer during operation.

[0007] The time acquisition module is used to obtain the error time series corresponding to the error data, determine the change trend of the error time series when the transformer is running, and the change amount of the predicted value.

[0008] The state detection module is used to detect the state of the transformer, determine the change coefficient of the corresponding inductance of the transformer when the error data appears, and the probability of the electrical stress change of the transformer.

[0009] The error verification module is used to verify the influence coefficient of the change of electrical stress on the actual output power of the current transformer under different error conditions.

[0010] The state evaluation module is used to evaluate the working state of the transformer according to the acquired influence coefficient to obtain the evaluation coefficient of the current transformer.

[0011] The beneficial effects of the present invention are as follows: the present invention uses a time series analysis method to more accurately identify and predict the changing trend of error data; by comprehensively considering the inductance variation coefficient and the probability of electric stress variation, the state of the transformer can be detected more comprehensively; by verifying the influence of electric stress on the actual output power under different error conditions, a reliable basis is provided for fault diagnosis; by weighted summation of multiple indicators, a comprehensive evaluation coefficient can be obtained to comprehensively reflect the working state of the transformer; through a hierarchical evaluation method, a detailed evaluation can be performed on different variables to improve the granularity and accuracy of the evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0013] Figure 1 It is a system framework diagram of a transformer operation status monitoring terminal based on artificial intelligence.

[0014] Figure 2 It is a system flow chart of a transformer operation status monitoring terminal based on artificial intelligence.

[0015] Figure 3 The present invention is a schematic diagram of a device for monitoring the operation status of a transformer based on artificial intelligence.

[0016] In the figure: 500, electronic device; 510, memory; 511, computer program; 520, processor. DETAILED DESCRIPTION

[0017] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. If no specific techniques or conditions are specified in the embodiments, the techniques or conditions described in the literature in the art or the product specifications are used.

[0018] See also Figure 1 , Figure 2 , a transformer operation status monitoring terminal based on artificial intelligence, including: a data acquisition module, a time acquisition module, a status detection module, an error verification module and a status evaluation module; the data acquisition module inputs data into the time acquisition module, the time acquisition module inputs the processed data of the error time series into the status detection module, the status detection module inputs the corresponding coefficient into the error verification module, and after verification, the error verification module evaluates the whole through the status evaluation module.

[0019] The data acquisition module is used to obtain error data of the transformer during operation, wherein the error data includes the temperature, current, voltage and the change value of the current boundary of the transformer during operation.

[0020] The time acquisition module is used to obtain the error time series corresponding to the error data, determine the change trend of the error time series when the transformer is running, and the change amount of the predicted value.

[0021] The state detection module is used to detect the state of the transformer, determine the change coefficient of the corresponding inductance of the transformer when the error data appears, and the probability of the electrical stress change of the transformer.

[0022] The error verification module is used to verify the influence coefficient of the change of electrical stress on the actual output power of the current transformer under different error conditions.

[0023] The state evaluation module is used to evaluate the working state of the transformer according to the acquired influence coefficient to obtain the evaluation coefficient of the current transformer.

[0024] At this time, the electrical stress indicates the maximum current and voltage that the transformer can accept; when the maximum allowable current and voltage change, it means that the current transformer is aging or has other problems, resulting in a decrease in the current processing effect on the transformer.

[0025] In one embodiment of the present invention, the error data in the data acquisition module is determined by setting current, voltage and temperature sensors on the transformer to determine the errors in temperature and current when the transformer is running. The voltage sensor is used to determine the stability of the current and to assist in determining the change value of the current boundary at this time. The change value of the current boundary in this embodiment refers to the upper and lower limits of the three-phase current that triggers the protection action during normal operation. The changes in this boundary value over the time the transformer is used can detect whether the transformer has an aging fault and whether the current state of the transformer needs to be adjusted.

[0026] At this time, an ammeter is used to measure the current, a voltmeter is used to measure the voltage, and an overcurrent protection relay is used to obtain whether the current exceeds the boundary value and the change of the current boundary value identified at this time; the temperature is detected using a temperature sensor to obtain temperature data in the presence of errors.

[0027] In one embodiment of the present invention, the time acquisition module processes the error time series as follows.

[0028] For the error time series obtained, the form of time series analysis is adopted to record the changing trend of the error data at the corresponding time according to the time when the error data appears in the mutual inductor, and record the change of the corresponding value of the error data when the error occurs, such as determining the change value of temperature, current, voltage, and current boundary, and calculating the changing trend amount of each temperature, current, voltage, and current boundary in the error time series in turn, and take the comprehensive value of the changing trend amount of temperature, current, voltage and current boundary as the changing trend identified at this time.

[0029] At the same time, when verifying the changing trend, verify the changing trend under the conditions of each error data, and use the residual of the identified changing trend as the change of the output predicted value; to verify that when the predicted value changes, each error condition will cause the residual value of the error, as well as the occurrence of errors that can be affected under different conditions; the calculated residual is the difference between the changing trend and the expected value to determine the corresponding residual.

[0030] The temperature change trend quantity is expressed as follows: the temperature change trend in the error time series is extracted, and the temperature change trend quantity is calculated based on the temperature change trend; for example, the temperature change trend within 7 days is calculated to obtain the current required temperature change trend.

[0031] When calculating the temperature change trend, the value corresponding to the temperature change trend will be identified, and the moving average of the temperature at each time point will be calculated according to the sampling time point. Based on the obtained moving average, the slope value of the temperature at each time point will be determined; the temperature slope value will be detected, and the value output after detection will be used as the temperature change trend at this time.

[0032] The moving average of the temperature at each time point is expressed as: ; represents the moving average of temperature, The value representing the temperature change trend at the i-th time point; Indicates the time window, which is used to adjust the temperature value of the current time point to the average of the adjacent temperatures. For example, when k=3, the second point is the average of the first to third points, the third point is the average of the second to fourth points, and so on; ; The value of k is less than n, Indicates the number of time points.

[0033] The slope value of the temperature at this time is expressed as: ;in, represents the slope value of temperature, represents the time value of the i-th time point, Indicates the average value of the moment value. The moment value set at this time indicates how long it takes for the current temperature to change. The moment value at this time indicates the difference between the corresponding time points. represents the moving average of the temperature at the i-th time point, Indicates the standard value corresponding to the moving average of temperature. The slope value at this time can tell the changing trend of the current temperature after the corresponding time point has passed.

[0034] Then the way to test the slope value of the temperature is to verify the ratio of the calculated slope value to the standard error, and output the slope value whose ratio is greater than the preset temperature ratio as the temperature change trend at this time.

[0035] The calculation method for the standard error is: ;in, Expressed as the standard error corresponding to the temperature.

[0036] At this point, after multiple processing methods, the temperature change trend can be obtained to verify the data distribution corresponding to the temperature in the error time series, as well as the final verification result.

[0037] For the calculation of the current change trend, you can choose data within 5 days to measure the corresponding current change. For the voltage change trend, you can choose data within 3 days to calculate the corresponding voltage change.

[0038] At this time, when verifying the changing trend of current and voltage, the same time point as temperature sampling is used to verify the changing trend of current and voltage at this time.

[0039] That is, the calculation method of the change trend of current and voltage is to extract the change trend of current and voltage from the error time series, calculate the change rate of current and voltage, and use the correlation coefficient between the change rate of current and voltage and the preset change rate as the output change trend.

[0040] At this time, the rate of change of current and voltage is expressed as: ;in, represents the rate of change of current and voltage at the i-th time point, Represents the value of current and voltage at the i+1th time point; It represents the value of current and voltage at the i-th time point, where the value of i ranges from 1 to n-1, to determine the change rate of the corresponding current and voltage within n time points.

[0041] The correlation coefficient between the change rate of current and voltage and the preset change rate is shown below.

[0042] ;in, Indicates the correlation coefficient between the rate of change of current and voltage and the preset rate of change, represents the rate of change of the current at the i-th time point, represents the rate of change of voltage at the i-th time point, Indicates the standard value of the rate of change of current, A standard value that indicates the rate of change of voltage.

[0043] The correlation coefficient obtained at this time can reflect the correlation between the current and voltage under the corresponding change trend. At this time, the data corresponding to this part of the correlation is output to obtain the change trend of the current and voltage.

[0044] For the change trend of the current boundary, data within 7 days can be selected to determine the change of the corresponding boundary. At this time, the change trend of the current boundary is implemented by successively calculating the first probability value corresponding to the upper limit value of the current boundary and the second probability value corresponding to the lower limit value, subtracting the current boundary value when the first probability value and the second probability value take the maximum value from the preset boundary value, and using the subtracted value as the change trend of the current boundary identified at this time; at this time, the upper limit value corresponding to the maximum first probability value is subtracted from the preset boundary value upper limit value, and then the lower limit value of the second probability value taking the maximum value is calculated and subtracted from the preset boundary value lower limit value, and the average of the sum of the two subtracted values ​​is used as the change trend at this time.

[0045] The first probability value at this time is calculated by dividing the number of times one of the current boundary upper limit values ​​appears by the total number of values. The second probability value is calculated in the same way. What is obtained at this time is the boundary value that needs to be paid special attention to, and it is determined how much the boundary value that needs to be paid attention to differs from the normal boundary value.

[0046] Finally, the change trend quantity corresponding to the temperature, current, voltage, and current boundary is output, and the change trend quantity is subtracted from the expected value to obtain the change quantity of the predicted value.

[0047] In one embodiment of the present invention, the processing method of the state detection module is as follows.

[0048] At this time, after completing the correlation analysis of the error time series, determine whether the current inductance has changed. For example, you can calculate whether the currently identified inductance value has changed, and the change in the corresponding insulation performance of the corresponding inductance value to quantify the change coefficient of the inductance. For example, you can determine the change in the insulation performance of the current transformer when it is in use to verify the change coefficient of the current inductance. By calculating the dielectric loss rate of the transformer, the change trend of the current and voltage under different insulation conditions, you can determine the change coefficient of the current transformer's corresponding inductance.

[0049] That is, at this time, the dielectric loss rate, current and voltage change trends of the transformer are weighted and summed to obtain the change coefficient of inductance.

[0050] The dielectric loss rate at this time can be expressed as: ;in, Indicates dielectric loss rate, Indicates the power loss, Represents the total power. At this time, the ratio of the loss power to the total power is calculated to determine the dielectric loss rate corresponding to the transformer.

[0051] Then the coefficient of change of the current transformer corresponding to the inductance is, ;in, represents the coefficient of variation of inductance, represents the weight of dielectric loss rate, The weight that represents the changing trend of current and voltage.

[0052] The identified electrical stress refers to the maximum current value and the maximum voltage value; verifying the probability of electrical stress change is to extract the corresponding value from the error time series to record whether the maximum current value and the maximum voltage value have changed, and how much the change value is.

[0053] The maximum current value at this time has a different meaning from the current boundary set above. The maximum current value at this time refers to the current limit value of the transformer during design, which refers to the maximum current value that the equipment or system can carry; the current boundary indicates the value that can trigger the transformer to perform self-protection. The current boundary is more significant in distinguishing the current range corresponding to the normal working and fault states of the transformer. After testing, the two values ​​have different positions and corresponding structures. When current or voltage abnormalities occur, the scope of verification can be narrowed, thereby improving the effect of transformer status monitoring.

[0054] The implementation method of the probability of electric stress change includes: according to the obtained change trend amount of the current boundary, taking the time interval corresponding to the maximum value of the first probability value and the second probability value in the current boundary as the initial interval, and calculating the probability of electric stress change in the initial interval; when the probability of electric stress change in the initial interval reaches a preset threshold, taking the probability of electric stress change in the initial interval at this time as the output probability of electric stress change; when the probability of electric stress change in the initial interval is less than the preset threshold, obtaining the adjacent time interval of the initial interval, the size of the adjacent time interval is consistent with the size of the initial interval, and taking the probability of electric stress change calculated in the adjacent time interval as the output probability of electric stress change.

[0055] The initial interval at this time is to select the time point when the first probability value reaches the maximum value to the time point when the second probability value reaches the maximum value, to obtain the corresponding initial interval, and verify the maximum current value and maximum voltage value in this interval, so as to determine whether the current transformer has a corresponding fault.

[0056] At this time, the method for calculating the probability of change of the electric stress is to calculate the corresponding value change rate of the electric stress at the extreme point existing in the corresponding time interval to obtain the probability of change of the electric stress.

[0057] At this time, the ratio of the value of the electric stress at the extreme point to the preset standard value is obtained to determine the probability of change of the electric stress; at the same time, in order to be able to evaluate the situation in the corresponding time zone, for the probability of change of the electric stress in the initial interval, the average value of the ratio of the value of the electric stress at the extreme point to the preset standard value is calculated as the output value of the probability of change of the electric stress in the initial interval.

[0058] For the probability of electric stress change in adjacent time intervals, a set of data with the largest difference between the ratio of the electric stress value at the extreme point and the preset standard value at adjacent time points is obtained as the output electric stress change probability.

[0059] In one embodiment of the present invention, the error verification module is implemented as follows.

[0060] This is used when the current transformer can normally output power and process values ​​due to certain faults or when the transformer is worn out due to long-term use, so as to verify whether the performance of the transformer has declined.

[0061] The influence coefficient obtained is the ratio of the actual output power of the transformer under the error data to the theoretical output power, which is used to quantify the operating state of the transformer under fault conditions.

[0062] In one embodiment of the present invention, the evaluation coefficient of the state evaluation module is obtained as shown below.

[0063] In the final evaluation, the current transformer is comprehensively evaluated based on the multiple values ​​obtained to identify the corresponding value.

[0064] The evaluation coefficient is expressed as: the inductance variation coefficient, the electric stress variation probability, the predicted value variation and the actual output power influence coefficient are normalized, and the normalized inductance variation coefficient, the electric stress variation probability, the predicted value variation and the actual output power influence coefficient are calculated to obtain the evaluation coefficient.

[0065] Therefore, the evaluation coefficient is expressed as: ;in, Expressed as the evaluation coefficient, Expressed as the coefficient of variation of inductance, Expressed as the probability of electrical stress change, Expressed as the change in the predicted value, Expressed as the influence coefficient of actual output power; Expressed as the weight of the coefficient of variation of the inductance, Expressed as the weight of the probability of electrical stress change, Expressed as the weight of the change in the predicted value, Expressed as the weight of the influence coefficient on the actual output power.

[0066] When calculating the evaluation coefficient, since the change in the predicted value represents the difference between the changing trend of temperature, the changing trend of current and voltage, and the changing trend of current boundary and the expected value, the evaluation coefficient calculated at this time needs to calculate three evaluation coefficients separately, which are divided into an evaluation coefficient for the changing trend of temperature, an evaluation coefficient for the changing trend of current and voltage, and an evaluation coefficient for the changing trend of current boundary, and these three evaluation coefficients are combined to obtain the final evaluation coefficient; the combination method of these three evaluation coefficients is shown as follows.

[0067] ;in, represents the evaluation coefficient of the final output, Represents the evaluation coefficient for the temperature change trend, Indicates the evaluation coefficient for the trend of current and voltage changes, Represents the evaluation coefficient of the change trend quantity for the current boundary, Represents the weight of the evaluation coefficient for the temperature change trend, Represents the weight of the evaluation coefficient for the trend of current and voltage changes, represents the weight of the evaluation coefficient for the change trend of the current boundary, It represents the average value of the evaluation coefficient, which is the average value of the three evaluation coefficients. Represents an exponential constant.

[0068] At this time, the evaluation coefficient finally outputted can express the current situation of the transformer under the corresponding temperature, current, voltage and current boundary conditions, to quantify the final evaluation coefficient, and use this final evaluation coefficient as the output evaluation coefficient.

[0069] like Figure 3 An embodiment of the present invention provides an electronic device 500, including a memory 510, a processor 520, and a computer program 511 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the computer program 511, the following steps are implemented.

[0070] S1. Collect error data of the transformer during operation.

[0071] S2. Determine the time series change trend of the error data and predict the future change amount, calculate the change trend amount of each variable; calculate the change amount of the predicted value based on the difference between the change trend amount and the expected value.

[0072] S3, extracting temperature time series data; calculating the moving average at each time point; calculating the slope value of the temperature at each time point based on the moving average; detecting the slope value and using it as the temperature change trend quantity.

[0073] S4. Extract the current and voltage time series data, calculate the change rate of the current and voltage, calculate the correlation coefficient between the change rate and the preset change rate, and use the correlation coefficient as the output change trend quantity.

[0074] S5. Extract the time series data of the upper and lower limits of the current boundary; calculate the first probability value corresponding to the upper limit of the current boundary and the second probability value corresponding to the lower limit; take the maximum value of the first probability value and the second probability value; subtract the current boundary value corresponding to the maximum value from the preset boundary value to obtain the change trend amount of the current boundary.

[0075] S6. Obtain the change trend of dielectric loss ratio, current and voltage, perform weighted summation on the change trend of dielectric loss ratio, current and voltage, and obtain the inductance change coefficient.

[0076] S7. Determine the initial interval according to the change trend of the current boundary, and calculate the probability of electric stress change within the initial interval; if the probability of electric stress change reaches a preset threshold, use it as the output probability; otherwise, obtain the adjacent time interval and repeat the calculation until the condition is met.

[0077] S8. Calculate the actual output power, calculate the theoretical output power, and calculate the influence coefficient, that is, the ratio of the actual output power to the theoretical output power.

[0078] S9. Normalize the inductance variation coefficient, the electric stress variation probability, the variation of the predicted value, and the influence coefficient of the actual output power; and calculate the evaluation coefficient by weighted summation.

[0079] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention and they are still covered by the protection scope of the present invention.

Claims

1. A transformer operation status monitoring terminal based on artificial intelligence, characterized in that: include: A data acquisition module is used to acquire error data of the transformer during operation, wherein the error data includes the change values ​​of temperature, current, voltage and current boundary during operation of the transformer; A time acquisition module is used to obtain the error time series corresponding to the error data, determine the change trend of the error time series when the transformer is running, and the change amount of the predicted value; A state detection module is used to detect the state of the transformer, determine the change coefficient of the corresponding inductance of the transformer when the error data appears, and the probability of the electrical stress change of the transformer; The error verification module is used to verify the influence coefficient of the change of electrical stress on the actual output power of the current transformer under different error conditions; A state evaluation module is used to evaluate the working state of the transformer according to the acquired influence coefficient to obtain the evaluation coefficient of the current transformer; The dielectric loss rate, current and voltage variation trends of the transformer are weighted and summed to obtain the variation coefficient of the inductance; The implementation method of the probability of electric stress change includes: according to the change trend amount of the current boundary obtained, taking the time interval corresponding to the maximum value of the first probability value and the second probability value in the current boundary as the initial interval, and calculating the probability of electric stress change in the initial interval; when the probability of electric stress change in the initial interval reaches a preset threshold, taking the probability of electric stress change in the initial interval at this time as the output probability of electric stress change; when the probability of electric stress change in the initial interval is less than the preset threshold, obtaining the adjacent time interval of the initial interval, the size of the adjacent time interval is consistent with the size of the initial interval, and taking the probability of electric stress change calculated in the adjacent time interval as the output probability of electric stress change; The inductance variation coefficient, the electric stress variation probability, the predicted value variation, and the actual output power influence coefficient are normalized, and the normalized inductance variation coefficient, the electric stress variation probability, the predicted value variation, and the actual output power influence coefficient are calculated to obtain the evaluation coefficient: ； in, Expressed as the evaluation coefficient, Expressed as the coefficient of variation of inductance, Expressed as the probability of electrical stress change, Expressed as the change in the predicted value, Expressed as the influence coefficient of actual output power; Expressed as the weight of the coefficient of variation of the inductance, Expressed as the weight of the probability of electrical stress change, Expressed as the weight of the change in the predicted value, It is expressed as the weight of the influence coefficient of the actual output power; When calculating the evaluation coefficient, three evaluation coefficients need to be calculated separately, which are divided into the evaluation coefficient for the temperature change trend quantity, the evaluation coefficient for the current and voltage change trend quantity, and the evaluation coefficient for the current boundary change trend quantity, and these three evaluation coefficients are combined to obtain the final evaluation coefficient; the combination method of these three evaluation coefficients is shown as follows: ； in, represents the evaluation coefficient of the final output, Represents the evaluation coefficient for the temperature change trend, Indicates the evaluation coefficient for the trend of current and voltage changes, Represents the evaluation coefficient of the change trend quantity for the current boundary, Represents the weight of the evaluation coefficient for the temperature change trend, Represents the weight of the evaluation coefficient for the change trend of current and voltage, represents the weight of the evaluation coefficient for the change trend of the current boundary, It represents the average value of the evaluation coefficient, which is the average value of the three evaluation coefficients. Represents an exponential constant.

2. The terminal for monitoring the operation status of a transformer based on artificial intelligence according to claim 1, characterized in that: The time acquisition module processes the error time series as follows: For the error time series obtained, the form of time series analysis is adopted. The error time series is recorded according to the time when the error data appears in the mutual inductor, and the changing trend of the error data at the corresponding time is recorded to determine the respective change values ​​of temperature, current, voltage and current boundary. The changing trend quantities of temperature, current, voltage and current boundary in the error time series are calculated in turn, and the comprehensive value of the changing trend quantities of temperature, current, voltage and current boundary is used as the changing trend identified at this time.

3. The terminal for monitoring the operation status of a transformer based on artificial intelligence according to claim 2 is characterized in that: The change amount of the predicted value is expressed as the change amount of the predicted value obtained by subtracting the change trend amount corresponding to the temperature, current, voltage and current boundary from the expected value.

4. The terminal for monitoring the operation status of a transformer based on artificial intelligence according to claim 2 is characterized in that: When calculating the temperature change trend, the value corresponding to the temperature change trend is identified, and the moving average value of the temperature at each time point is calculated according to the sampling time point. Based on the obtained moving average value, the slope value of the temperature at each time point is determined; The slope value of the temperature is detected, and the value output after the detection is used as the change trend of the temperature at this time.

5. The terminal for monitoring the operation status of a transformer based on artificial intelligence according to claim 2 is characterized in that: The calculation method of the change trend of current and voltage is as follows: extract the change trend of current and voltage from the error time series, calculate the change rate of current and voltage, and use the correlation coefficient between the change rate of current and voltage and the preset change rate as the output change trend.

6. The terminal for monitoring the operation status of a transformer based on artificial intelligence according to claim 2, characterized in that: For the change trend quantity of the current boundary, the implementation method is to calculate the first probability value corresponding to the upper limit value of the current boundary and the second probability value corresponding to the lower limit value in sequence, subtract the current boundary value when the first probability value and the second probability value take the maximum value from the preset boundary value, and use the subtracted value as the change trend quantity of the current boundary identified at this time.

7. The terminal for monitoring the operation status of a transformer based on artificial intelligence according to claim 1, characterized in that: The influence coefficient indicates the ratio of the actual output power of the transformer to the theoretical output power when error data occurs.

Images