Transformer Health Assessment Method Based on Fuzzy Logic

Abstract

The present invention relates to a method for evaluating the health of a transformer based on fuzzy logic, comprising the following steps: S10: Collect the operation data of the transformer, screen and summarize the data, and extract useful data at the same time; S20: Preset a health evaluation model based on fuzzy logic, input the extracted data into each health evaluation model, and evaluate the health of the transformer; S30: Through deep learning methods, optimize the health evaluation model with the data collected later, and the number of required characteristic samples for evolution is greater than or equal to 1000; S40: Regularly verify the health evaluation model through preset standard data values; S50: Output the health index of the transformer and the maintenance time plan. The present invention comprehensively considers various parameters during the operation of the transformer, judges the health status of the transformer through fuzzy logic rules, and specifies a maintenance plan according to the health status.

Description

Technical Field

[0001] The present invention belongs to the technical field of transformer health condition detection methods, and specifically relates to a transformer health degree evaluation method based on fuzzy logic. Background Technique

[0002] Electricity is an indispensable energy source in national production and life. With the acceleration of the urbanization and modernization processes, the scale of the power system is also constantly expanding.

[0003] As the core equipment of the power system, the operating conditions and health conditions of transformers are closely related to the safe and stable operation of the power system. However, during the long-term operation of transformers, affected by many factors and their own aging conditions, various faults will occur during the operation of transformers. Therefore, it is necessary to detect the operating conditions of transformers and perform regular maintenance.

[0004] Currently, with the modernization and intelligent transformation of transformers, parameters during the operation of transformers, such as winding temperature, oil temperature, input and output voltage and current of the transformer, are detected in real time through sensors and uploaded to the control center. Substation operation and maintenance personnel judge the operating conditions of transformers based on these parameters.

[0005] However, currently, the impacts of these parameters on the health condition of transformers and the maintenance cycle are not comprehensively considered. The determination of the maintenance cycle is still carried out in a rough time period for maintenance, such as once a year or once every six months. Summary of the Invention

[0006] The technical problem to be solved by the present invention is: to overcome the deficiencies of the prior art and provide a transformer health degree evaluation method based on fuzzy logic. The present invention comprehensively considers various parameters during the operation of transformers, judges the health condition of transformers through fuzzy logic rules, and formulates a maintenance plan according to the health condition.

[0007] The technical solution adopted by the present invention to solve the problems existing in the prior art is:

[0008] A transformer health degree evaluation method based on fuzzy logic, comprising the following steps:

[0009] S10: Collect the operation data of the transformer, screen and summarize the data, and extract useful data at the same time;

[0010] S20: Preset a health evaluation model based on fuzzy logic, input the extracted data into each health evaluation model, and evaluate the health degree of the transformer;

[0011] S30: Optimize the health assessment model through deep learning methods using the data collected later, with the number of required characteristic samples for evolution being greater than or equal to 1000;

[0012] S40: Regularly verify the health assessment model against preset standard data values;

[0013] S50: Output the transformer health index and the maintenance time plan.

[0014] Preferably, the operating data of the transformer collected in S10 includes the winding temperature, oil temperature, input and output current, and voltage of the transformer collected by sensors;

[0015] Each parameter is classified and stored in the corresponding unit of the memory, and each parameter is classified according to a preset parameter judgment method, mainly into qualified parameters and unqualified parameters.

[0016] If the obtained parameter is an unqualified parameter, the controller controls the sensor to repeatedly collect the parameter subsequently, and analyzes whether the parameter is abnormal or the sensor is faulty based on the data collected subsequently.

[0017] If the sensor is faulty, an alarm is issued, a work order is formed, and manual maintenance is requested; if the parameter is abnormal, it is brought into the health assessment model to evaluate the health of the transformer.

[0018] Preferably, both the winding temperature and the oil temperature are preset with an interval range. When extracting the data of both, first judge whether it is within the preset interval range. If it is within this interval range, it is determined that the data meets the standard and the transformer health index is good; if it is not within this interval range, it is determined that the data is abnormal.

[0019] Preferably, the two data of the winding temperature and the oil temperature collected at intervals are analyzed through the following formula:

[0020] Winding temperature rise = (I 2 ×R)×(C + k×t1) / (1 + h×(C + k×t1))

[0021] where I is the transformer input current, R is the winding resistance, C is the constant loss, t1 is the transformer operating time, k is the temperature constant, and h is the cooling coefficient;

[0022] Oil temperature rise = (I 2 ×R)×(k1×t1 + k2×t2) / (1 + h ×(k1×t1 + k2×t2))

[0023] Wherein, I is the input current of the transformer, R is the winding resistance, k1 and k2 are the oil temperature rise rate constants, t1 is the operating time of the transformer, t2 is the square of the operating time of the transformer, and h is the cooling coefficient;

[0024] If two data collected at an interval conform to the above formula, it indicates that the data is normal;

[0025] If two data collected at an interval do not conform to the above formula, it indicates that the data is abnormal and feedback is made.

[0026] Preferably, the winding temperature rise value and the oil temperature rise value calculated by the formula are respectively multiplied by 0.7 and 1.25 to form an interval value.

[0027] The actual temperature difference between two data collected at an interval is corresponded to it to determine whether the collected data is abnormal.

[0028] Preferably, in step S20, through fuzzy logic, based on the input data (winding temperature, oil temperature, transformer input and output current and voltage), the specific algorithm for establishing a transformer health assessment model includes the following steps:

[0029] S21. Determine variables and membership functions:

[0030] Determine the input variables: winding temperature, oil temperature, transformer input and output current and voltage;

[0031] Determine the output variable: transformer health status;

[0032] S22. Construct a set of fuzzy rules:

[0033] Based on the knowledge and experience in this field, determine a set of fuzzy rules to associate the input variables with the output variables;

[0034] S23. Conduct fuzzy reasoning:

[0035] According to the membership degrees of the input data, apply these data to the fuzzy rules to calculate the fuzzy measure of the output variable;

[0036] S24. Conduct defuzzification processing:

[0037] Use the defuzzification method to convert the fuzzy measure into a specific health status value;

[0038] S25. Establish model verification and optimization:

[0039] Use the known data set to verify the model, evaluate the accuracy and performance of the model, and for the indicators that need to be optimized, conduct parameter adjustment and model improvement.

[0040] Preferably, each fuzzy rule consists of a condition part and a conclusion part.

[0041] Preferably, the condition part consists of the membership functions of the input variables, describing the conditional situation of the input variables; the conclusion part consists of the membership functions of the output variables, describing the result of the output variables.

[0042] Preferably, for each fuzzy rule, a weight can be set to represent its importance or confidence level.

[0043] It is also necessary to determine the operators of fuzzy logic, such as using "AND" and "OR" to combine multiple conditions.

[0044] Preferably, by using data analysis methods, the rule weights are learned from the sample data to determine the weights of each fuzzy rule.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] (1) The health status of the transformer is affected by various factors, including environmental conditions, working load, electrical performance, etc. There is a certain degree of uncertainty and ambiguity between these factors and the health status, and it is difficult to judge using precise logical methods. Using fuzzy logic can fully consider this uncertainty and ambiguity, improving the accuracy and reliability of the evaluation results.

[0047] (2) Compared with other complex machine learning models, the fuzzy logic rule model has a relatively simple modeling process, is easy to understand and maintain. At the same time, the rules in the rule model can be obtained through the knowledge and experience in this field, and have strong interpretability and credibility.

[0048] (3) The health status of the transformer involves multiple aspects of factors, such as electrical performance, mechanical structure, etc. Using fuzzy logic can comprehensively evaluate these factors to obtain a comprehensive health index, facilitating quantitative and qualitative analysis.

[0049] (4) The fuzzy logic model can be used to monitor the health status of the transformer in real time, detect abnormal situations and problems in a timely manner, improving the reliability and safety of the equipment. At the same time, based on historical data and the fuzzy logic model, the prediction of the health status can be carried out, providing a scientific basis for maintenance and repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The present invention will be further described below in conjunction with the drawings and embodiments.

[0051] Figure 1 It is a flowchart of the steps of the transformer health assessment method based on fuzzy logic of the present invention.

[0052] Figure 2 Schematic diagram of the terminal device in the transformer health assessment method based on fuzzy logic of the present invention. Detailed implementation manners

[0053] As used in the specification and claims, certain terms are used to refer to specific components. Those skilled in the art should understand that hardware manufacturers may use different terms to refer to the same component. The specification and claims do not use the difference in names as a way to distinguish components, but use the difference in functions of components as the criterion for distinction. As mentioned throughout the specification and claims, "comprising" is an open-ended term and should be interpreted as "including but not limited to". "Substantially" means within an acceptable error range. Those skilled in the art can solve the technical problems within a certain error range and basically achieve the technical effects.

[0054] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "upper", "lower", "front", "rear", "left", "right", "horizontal", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.

[0055] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed", etc. should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0056] The following further details the transformer health assessment method based on fuzzy logic of the present invention with reference to the drawings, but does not limit the present invention.

[0057] The transformer health assessment method based on fuzzy logic includes the following steps:

[0058] S10: Collect the operation data of the transformer, screen and summarize the data, and extract useful data at the same time;

[0059] Collect data through the existing detection modules (i.e., various sensors) on the transformer. The operation data of the collected transformer includes the winding temperature, oil temperature, input and output current, and voltage of the transformer collected by the sensors.

[0060] The various parameters detected by the detection module are classified and stored in the corresponding units of the storage module, and the various parameters are classified according to a preset parameter judgment method. The classification is mainly qualified parameters and unqualified parameters.

[0061] The preset parameter judgment method adopts the interval judgment method, that is, there are preset interval ranges for the current and voltage of the transformer input and output, winding temperature and oil temperature. This interval range is the parameter interval range during the normal operation of the transformer. At the same time, the current and voltage of the transformer input and output are corresponding through the formula:

[0062] Transformer turns ratio formula:

[0063] According to the turns ratio relationship of the transformer, the relationship between the input voltage and the output voltage can be expressed by the turns ratio of the transformer. The turns ratio represents the ratio of the number of turns of the input side to the output side, that is:

[0064] V out = (N out / N in )×V in

[0065] Among them, V out and V in represent the output voltage and the input voltage respectively, N out and N in represent the number of turns of the output side and the input side.

[0066] Transformer equivalent circuit model:

[0067] By establishing the equivalent circuit model of the transformer, the relationship between the input and output current and voltage can be described by Ohm's law and Kirchhoff's voltage law. This model includes parameters such as resistance, inductive reactance and mutual inductance. For a single-phase ideal transformer, the following formula can be used:

[0068] V out = V in - (I out ×R) - (j ω ×L×I out )

[0069] Among them, V out and V in represent the output voltage and the input voltage respectively, I out represents the output current, R represents the winding resistance, L represents the equivalent inductive reactance, j ω represents the angular frequency.

[0070] In the process of parameter judgment, for the winding temperature and the oil temperature, if the collected data is within the preset interval range, then the parameter is a qualified parameter and is stored in the qualified parameter area inside the storage device.

[0071] For the input and output current and voltage of the transformer, the collected data should not only be compared with the preset range individually, but also be substituted into the above two formulas to determine whether the input and output voltage and current match. If they match, they are determined as qualified parameters and stored in the qualified parameter area inside the storage device.

[0072] If the collected parameters are within the preset range, however, they cannot match the above formulas, especially the transformer turns ratio formula, then this set of parameters will be stored in the to-be-determined parameter area of the storage device.

[0073] An algorithm is embedded in the to-be-determined parameter area to analyze the reasons for the mismatch between the input and output voltage and current of the transformer through an algorithm model.

[0074] The algorithm embedded in the to-be-determined parameter area is as follows:

[0075] V in = (R c + jX c )×I in + (M / R m + jM / X m )×I out

[0076] V out = (R c + jX c )×I out + (M / R m + jM / X m )×I in

[0077] Among them, the coil resistance (R c ) represents the resistance of the transformer coil, the coil reactance (X c ) represents the inductance of the transformer coil, the magnetic circuit resistance (R m ) represents the magnetic circuit resistance of the iron core material, taking into account the loss of the iron core, the magnetic circuit reactance (X m ) represents the magnetic circuit inductance of the iron core material, taking into account the saturation characteristic of the iron core, and the mutual inductance (M) represents the mutual inductance between the input side and the output side.

[0078] Among them, j represents the imaginary unit. When the imaginary unit j appears in the calculation of elements such as inductance and reactance, then jX c represents the coil reactance, and jM represents the magnetic circuit reactance of the iron core material.

[0079] R c , X c , jX c and jM in the formula are all set as constant fixed values, Vin 、V out and I in 、I out are the detection parameters of the detection module. Therefore, when R m remains unchanged, the value of X m can be calculated. This value is compared with the set value range. If it is within the range, it indicates that there is no problem with the saturation characteristic of the iron core. If it is not within the range, a signal indicating a problem with the saturation characteristic of the iron core is output.

[0080] Similarly, when X m remains unchanged, the value of R m is calculated. This value is compared with the set value range. If it is within the range, it indicates that there is no problem with the iron core loss. If it is not within the range, a signal indicating an iron core loss is output.

[0081] And the parameters are placed in the unqualified parameter area of the storage module in the form of a time axis.

[0082] For the parameters of winding temperature and oil temperature, if they are not within the preset range, they are also placed in the unqualified parameter area of the storage module in the form of a time axis.

[0083] For the winding temperature and oil temperature parameters placed in the qualified area of the storage module, within a certain time span, they can be rechecked through the following algorithm:

[0084] Two data of winding temperature and oil temperature collected at intervals are analyzed through the following formula:

[0085] Winding temperature rise = (I 2 ×R)×(C + k×t1) / (1 + h×(C + k×t1))

[0086] where I is the transformer input current, R is the winding resistance, C is the constant loss, t1 is the transformer operation time, k is the temperature constant, and h is the cooling coefficient;

[0087] Oil temperature rise = (I 2 ×R)×(k1×t1 + k2×t2) / (1 + h ×(k1×t1 + k2×t2))

[0088] where I is the transformer input current, R is the winding resistance, k1 and k2 are the oil temperature rise rate constants, t1 is the transformer operation time, t2 is the square of the transformer operation time, and h is the cooling coefficient;

[0089] If the two data collected at intervals conform to the above formula, it indicates that the data is normal;

[0090] If two pieces of data collected at an interval do not conform to the above formula, it indicates that the data is abnormal and feedback is given.

[0091] The values of the winding temperature rise and the oil temperature rise calculated by the formula are multiplied by 0.7 and 1.25 respectively to form an interval value, allowing the value to fluctuate within this range.

[0092] The actual temperature difference between two pieces of data collected at an interval is corresponded to it to determine whether the collected data is abnormal.

[0093] Through the re-inspection of the above algorithm, it can be judged whether there is a fault in the cooling system of the transformer. If the re-inspection is qualified, it proves that the cooling system is operating well. If there is a deviation in the re-inspection value, it proves that there is a problem with the operating state of the cooling system, and a report is output and self-adjustment is performed.

[0094] The output report and adjustment method are as follows:

[0095] 1. If the actual temperature difference is less than the temperature difference calculated by the algorithm, it proves that the power of the transformer cooling system is excessive. An instruction is given through the controller to reduce the power of the cooling system;

[0096] 2. If the actual temperature difference is greater than the temperature difference calculated by the algorithm, it proves that the power of the transformer cooling system is insufficient. An instruction is given through the controller to increase the power of the cooling system.

[0097] Through step S10, the data collected by each detection module can be sorted and classified into qualified parameters and unqualified parameters on the same time axis.

[0098] If the obtained parameter is an unqualified parameter, the controller controls the sensor to repeatedly collect the parameter subsequently, and analyzes whether the parameter is abnormal or the sensor is faulty through the subsequently collected data.

[0099] If the sensor is faulty, an alarm is issued, a work order is formed, and manual maintenance is requested; if the parameter is abnormal, it is brought into the health assessment model to evaluate the health of the transformer.

[0100] S20: Preset a health assessment model based on fuzzy logic, input the extracted data into each health assessment model, and evaluate the health of the transformer;

[0101] If machine learning algorithms such as support vector machines (SVMs), random forests, neural networks, etc., and deep learning algorithms such as convolutional neural networks (CNNs) or recurrent neural networks (RNNs) are used for transformer health assessment, these algorithms belong to precise logical methods. However, the health status of a transformer is affected by multiple factors, including environmental conditions, workload, electrical performance, etc. There is a certain degree of uncertainty and ambiguity between these factors and the health status, making it difficult to judge using precise logical methods. Using fuzzy logic can fully consider this uncertainty and ambiguity, improving the accuracy and reliability of the evaluation results.

[0102] In this step, through fuzzy logic, the specific algorithm for establishing a transformer health assessment model based on input data (winding temperature, oil temperature, transformer input and output current and voltage) includes the following steps:

[0103] S21. Determine variables and membership functions:

[0104] Determine the input variables: winding temperature, oil temperature, transformer input and output current and voltage; determine the output variable: transformer health status.

[0105] Define membership functions for each variable, such as high, medium, and low levels of temperature.

[0106] The membership function of a variable is defined in fuzzy logic and is used to describe the degree of membership of a variable at different values. The membership function is usually represented in the form of a graph and can be a triangular function, trapezoidal function, Gaussian function, etc.

[0107] Taking the winding temperature as an example, three membership functions can be defined: High, Medium, and Low. The specific definitions can be as follows:

[0108] High membership function:

[0109] Within an appropriate temperature range, the degree of membership gradually increases, indicating the degree of higher winding temperature. It can be represented using a triangular or trapezoidal function.

[0110] Medium membership function:

[0111] Within an appropriate temperature range, the degree of membership reaches the maximum value, indicating that the winding temperature is at a medium level. It can be represented using a triangular or trapezoidal function.

[0112] Low membership function:

[0113] Within an appropriate temperature range, the degree of membership gradually increases, indicating the degree of lower winding temperature. It can be represented using a triangular or trapezoidal function.

[0114] In this embodiment, the above membership functions are all represented by triangular functions, specifically as follows:

[0115] When using a triangular function to identify a variable, the shape of the triangle can be determined by defining three key points, namely the left boundary, the peak, and the right boundary. The specific steps are as follows:

[0116] Determine the value range of the variable:

[0117] First, determine the minimum and maximum values of the variable. For example, for the variable of winding temperature, assume the value range is from 0 to 100 degrees.

[0118] Determine the key points of the triangular function:

[0119] Define three key points: the left boundary (a), the peak (b), and the right boundary (c). The values of these points should be within the value range of the variable and be distributed according to certain rules.

[0120] Draw the graph of the triangular function:

[0121] In the coordinate system, connect the three key points to form a triangle. The connection method is: connect from (a, 0) to (b, 1), and then from (b, 1) to (c, 0).

[0122] Note that the value of the triangular function outside the key points is 0, indicating that the membership degree outside this point is 0.

[0123] Set the membership degree of the variable:

[0124] Substitute the specific value of the variable into the triangular function to calculate the corresponding membership degree. According to the shape of the triangular function, the membership degree value reaches the maximum (usually 1) at the peak point and gradually decreases near the boundary points.

[0125] S22. Construct a set of fuzzy rules:

[0126] Based on the knowledge and experience in this field, determine a set of fuzzy rules. Each rule consists of a condition part and a conclusion part. The condition part is composed of the membership functions of the input variables, describing the condition of the input variables, and the conclusion part is composed of the membership functions of the output variables, describing the result of the output variables.

[0127] At the same time, set the rule weights and operators:

[0128] For each rule, a weight can be set to represent its importance or confidence.

[0129] It is also necessary to determine the operators of fuzzy logic, such as using "AND" and "OR" to combine multiple conditions.

[0130] S23. Perform fuzzy inference:

[0131] According to the membership degrees of the input data, apply these data to the fuzzy rules to calculate the fuzzy measure of the output variable;

[0132] S24. Perform defuzzification:

[0133] Use the defuzzification method to convert the fuzzy measure into a specific health status value;

[0134] S25. Establish model verification and optimization:

[0135] Use the known data set to verify the model, evaluate the accuracy and performance of the model, and for the indicators that need to be optimized, perform parameter adjustment and model improvement.

[0136] The basic fuzzy rules are as follows in the table:

[0137]

[0138] Based on the above table, add weights. The parameters of the weights adopt the current and voltage of the transformer input and output. Based on the judgment of the current and voltage parameters of the transformer input and output in step S10, if the current and voltage parameters are determined to be unqualified parameters, do not refer to the winding temperature and the oil temperature, and directly output the conclusion - the transformer is unhealthy.

[0139] In step S10, when the current and voltage parameters of the transformer input and output on the same time axis are determined to be qualified parameters, perform step S20, and judge the health status of the transformer through fuzzy rules.

[0140] S30: Through the deep learning method, optimize the health assessment model with the data collected later, and the number of required characteristic samples for evolution is greater than or equal to 1000;

[0141] S40: Regularly verify the health assessment model with the preset standard data values;

[0142] S50: Output the transformer health index and the maintenance time plan.

[0143] The health index of the transformer is judged and output through fuzzy rules. For the maintenance time plan, mainly consider the later development trends of the two parameters of the winding temperature and the oil temperature.

[0144] The prediction methods for the later development trends of the two are the same. Taking the winding temperature as an example, the following description is given:

[0145] First, winding temperature data needs to be collected over multiple time periods. These data should cover different operating conditions and environmental changes to obtain a more comprehensive and accurate model.

[0146] Then, the winding temperature is analyzed through the prediction model and algorithm embedded in the controller, and the model is used to predict the winding temperature in a certain time period in the future. The prediction result matches the maintenance time plan. For example, by predicting the winding temperature, it is predicted that the winding temperature value will evolve to a high value in a certain time period, and then the output conclusion of the transformer health will change from healthy to unhealthy. Therefore, the maintenance time can be scheduled before this time period.

[0147] The prediction model of winding temperature adopts the Autoregressive Integrated Moving Average (ARIMA) model, which is suitable for time series data with obvious trends and seasonality. The model is established by combining autoregression (AR) and moving average (MA), as follows:

[0148] Y(t) = c + φ(1)Y(t-1) + ... + φ(p)Y(tp) + θ(1)e(t-1) + ... + θ(q)e(tq) + e(t)

[0149] Where Y(t) is the winding temperature at time t, c is a constant term, φ(i) and θ(i) are model parameters, p and q are the autoregressive order and moving average order, respectively, and e(t) represents the white noise error term.

[0150] In the process of using the model, it is necessary to use data to continuously revise it, that is, to continuously predict the later time t n The winding temperature Y(t n ), and then through the arrival time t n The actual winding temperature after Y(t n ) to determine the deviation between the two and see if the deviation is within the specified range. If so, it meets the requirements. If the deviation deviates from the specified range, bring the actual winding temperature into the above model, calculate the p and q values, and the calculation method can be to fix one of the values ​​first and perform calculations separately.

[0151] The p and q values ​​are corrected through p and q in multiple time periods. The correction can use the intermediate values ​​of certain parameters to optimize the model algorithm.

[0152] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments, and various changes can be made without departing from the spirit of the present invention within the scope of knowledge possessed by those of ordinary skill in the relevant art.

Claims

1. A method for evaluating the health of a transformer based on fuzzy logic, characterized in that, It includes the following steps: S10: Collect the operation data of the transformer, screen and summarize the data, and extract useful data at the same time. The collected operation data of the transformer includes the winding temperature, oil temperature, input and output current and voltage of the transformer collected by sensors. Each parameter is classified and stored in the corresponding unit of the memory, and each parameter is classified according to the preset parameter judgment method. The classification is mainly qualified parameters and unqualified parameters. Both the winding temperature and the oil temperature are preset with an interval range. When extracting the data of the two, first judge whether it is within the preset interval range. If it is within this interval range, it is determined that the data meets the standard and the health index of the transformer is good. If it is not within this interval range, it is determined that the data is abnormal. The input and output current and voltage of the transformer are corresponding through the formula: Transformer turns ratio formula: According to the turns ratio relationship of the transformer, the relationship between the input voltage and the output voltage is represented by the turns ratio of the transformer. The turns ratio represents the ratio of the number of turns of the input side and the output side, that is: V out = (N out / N in )×V in Among them, V out and V in represent the output voltage and the input voltage respectively, and N out and N in represent the number of turns of the output side and the input side, Transformer equivalent circuit model: By establishing the equivalent circuit model of the transformer, the relationship between the input and output current and voltage is described by Ohm's law and Kirchhoff's voltage law. This model includes resistance, inductive reactance and mutual inductance parameters. For a single-phase ideal transformer, the following formula is used: V out = V in - (I out ×R) - (j ω ×L×I out ) Among them, I out represents the output current, R represents the winding resistance, L represents the equivalent inductive reactance, and j ω represents the angular frequency, In the process of parameter judgment, for the input and output current and voltage of the transformer, the collected data not only needs to be compared with the preset interval range separately, but also needs to be brought into the above two formulas to judge whether the input and output voltage and current match. If they match, they are judged as qualified parameters and stored in the qualified parameter area inside the storage device. If the collected parameter is within the preset interval range, but cannot match the transformer turns ratio formula, then the parameter is stored in the to-be-judged parameter area of the storage device. The two data of the winding temperature and the oil temperature collected at intervals are analyzed through the following formula: Winding temperature rise = (I 2 ×R)×(C + k×t1) / (1 + h×(C + k×t1)) Where, I is the input current of the transformer, R is the winding resistance, C is the constant loss, t1 is the operation time of the transformer, k is the temperature constant, and h is the cooling coefficient; Oil temperature rise = (I 2 ×R)×(k1×t1 + k2×t2) / (1 + h×(k1×t1 + k2×t2)) Where, I is the input current of the transformer, R is the winding resistance, k1 and k2 are the oil temperature rise rate constants, t1 is the operation time of the transformer, t2 is the square of the operation time of the transformer, and h is the cooling coefficient; If the two data collected at intervals meet the above formula, it indicates that the data is normal; If the two data collected at intervals do not meet the above formula, it indicates that the data is abnormal and feedback is made; S20: Preset a health assessment model based on fuzzy logic, input the extracted data into each health assessment model, and evaluate the health of the transformer; S30: Through deep learning method, optimize the health assessment model with the data collected later. The number of required characteristic samples for optimization is greater than or equal to 1000; S40: Regularly verify the health assessment model with the preset standard data value; S50: Output the health index of the transformer and the maintenance time plan.

2. The method for evaluating the health of a transformer based on fuzzy logic according to claim 1, characterized in that: If the parameter obtained in S10 is unqualified, the controller controls the sensor to subsequently collect the parameter repeatedly, and analyzes whether the parameter is abnormal or the sensor is faulty based on the subsequently collected data. If the sensor is faulty, an alarm is issued, a work order is formed, and manual maintenance is requested. If the parameter is abnormal, it is brought into the health assessment model to evaluate the health of the transformer.

3. The method for evaluating the health of a transformer based on fuzzy logic according to claim 2, characterized in that: The calculated winding temperature rise value and oil temperature rise value are respectively multiplied by 0.7 and 1.25 through a formula to form an interval value. The actual temperature difference between two data collected at intervals is corresponded to it to determine whether the collected data is abnormal.

4. The method for evaluating the health of a transformer based on fuzzy logic according to claim 2 or 3, characterized in that: In step S20, the input data is the winding temperature, oil temperature, and transformer input and output current and voltage. Then, based on the input data, the specific algorithm for establishing a transformer health assessment model by fuzzy logic includes the following steps: S21. Determine variables and membership functions: Determine input variables: winding temperature, oil temperature, transformer input and output current and voltage; Determine output variable: transformer health status; S22. Construct a set of fuzzy rules: Based on the knowledge and experience in this field, determine a set of fuzzy rules to associate the input variables with the output variables; S23. Perform fuzzy inference: According to the membership degrees of the input data, apply these data to the fuzzy rules to calculate the fuzzy measure of the output variable; S24. Perform defuzzification: Use the defuzzification method to convert the fuzzy measure into a specific health status value; S25. Establish model verification and optimization: Use known data sets for model verification, evaluate the accuracy and performance of the model, and perform parameter adjustment and model improvement for the indicators that need to be optimized.

5. The method for evaluating the health of a transformer based on fuzzy logic according to claim 4, characterized in that: Each fuzzy rule consists of a condition part and a conclusion part.

6. The method for evaluating the health of a transformer based on fuzzy logic according to claim 5, characterized in that: The condition part consists of the membership functions of the input variables, describing the condition of the input variables; The conclusion part consists of the membership functions of the output variables, describing the result of the output variables.

7. The method for evaluating the health of a transformer based on fuzzy logic according to claim 5 or 6, characterized in that: For each fuzzy rule, set a weight to represent its importance or confidence. It is also necessary to determine the operators of fuzzy logic, and use "AND" and "OR" to combine multiple conditions.

8. The method for evaluating the health of a transformer based on fuzzy logic according to claim 7, characterized in that: By using data analysis methods, learn the rule weights from the sample data to determine the weights of each fuzzy rule.

Images