Simulation identification method of transformer inter-turn insulation degradation fault based on box cover temperature

Abstract

The present invention belongs to the technical field of simulation analysis, and specifically relates to a transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature, the steps include constructing a three-dimensional geometric model of the original size of the transformer, and importing it into the finite element simulation software to obtain the simulation geometric model, and establish the "electromagnetic-fluid temperature field" coupling model of the transformer; based on the transformer entity test, the simulation results are compared and verified; the temperature change of the oil tank shell under different active loss levels and different fault positions is simulated, and a sample database of active loss, fault position, and oil tank shell temperature change rate is established; the threshold value of the active loss change rate and the oil tank shell temperature change rate when the inter-turn insulation degradation fault of the winding occurs is determined; during the operation of the transformer, the inter-turn insulation degradation fault position of the winding is estimated according to the three-phase winding voltage and current obtained by real-time measurement. The present invention can realize non-stop detection and avoid further expansion of the fault.

Description

Technical Field

[0001] The invention belongs to the technical field of simulation analysis, and in particular relates to a transformer interturn insulation degradation fault simulation identification method based on box cover temperature. Background Art

[0002] Transformers are key equipment in power systems, and their operational reliability is directly related to the safety and stability of the power grid. Transformer winding interturn insulation degradation fault is one of the common transformer fault types. However, since this type of fault is mainly manifested as a decrease in the insulation performance inside the winding in the early stage, and has a weak impact on the main circuit current, conventional electrical protection devices cannot effectively detect these subtle changes. Therefore, traditional detection methods often fail to detect early insulation degradation faults in the winding in a timely manner, resulting in further development of the fault, which may eventually cause serious power system accidents.

[0003] Existing detection methods mainly focus on main circuit current, transformer vibration, gas analysis, etc. However, these methods have obvious deficiencies in the early detection of inter-turn insulation degradation faults of windings. Not only can they not provide sufficient sensitivity and accuracy, but they can also not give the location of insulation degradation faults. In addition, the internal structure of transformer windings is complex and the fault characteristics are diverse. Traditional detection methods are difficult to fully reflect the insulation condition inside the windings, making it difficult to identify and prevent early faults. Therefore, an effective method for identifying inter-turn insulation degradation faults of transformer windings is urgently needed. Summary of the invention

[0004] In view of the above deficiencies in the prior art, the purpose of the present invention is to provide a transformer interturn insulation degradation fault simulation identification method based on the box cover temperature, which can not only realize non-power-off detection, but also make up for the shortcomings of traditional detection means, and provide an effective means for evaluating the insulation status of the winding and effectively identifying the early stage faults of transformer winding interturn insulation degradation.

[0005] To achieve the above objectives, the present invention provides a transformer inter-turn insulation degradation fault simulation identification method based on box cover temperature, comprising the following steps:

[0006] S1. Use 3D drawing software to construct a 3D geometric model of the original size of the transformer, import the 3D geometric model into finite element simulation software to obtain a simulation geometric model, and establish an "electromagnetic-fluid temperature field" coupling model of the transformer based on the simulation geometric model;

[0007] S2. Based on the transformer entity test, the simulation results of the "electromagnetic-fluid temperature field" coupling model are compared and verified, and it is determined whether the simulation geometry model needs to be readjusted according to the comparison results;

[0008] S3. Based on the simulated geometric model after comparison and verification, simulate the temperature change of the fuel tank shell under different active power loss levels and different fault locations, and establish a sample database of active power loss, fault location, and fuel tank shell temperature change rate;

[0009] S4, determining the thresholds of the rate of change of active power loss and the rate of change of oil tank casing temperature when a fault of inter-turn insulation degradation of the winding occurs;

[0010] S5. Based on the determined threshold value and sample database, during the operation of the transformer, the fault location of the insulation degradation between turns of the winding is estimated according to the winding voltage and current obtained by real-time measurement and the temperature change rate of the oil tank shell measured at the time of the fault.

[0011] As a preferred solution of the present invention, in S1, the specific process of establishing the "electromagnetic-fluid temperature field" coupling model of the transformer is:

[0012] S1.1. Use 3D drawing software to construct a 3D geometric model of the transformer in its original size, including the high-voltage winding, low-voltage winding, iron core, end insulation, clamps, transformer oil and oil tank shell, and model and number the low-voltage winding with each turn as the smallest unit;

[0013] S1.2, import the three-dimensional geometric model into ANSYS finite element simulation software, obtain the simulation geometric model, divide the electromagnetic field calculation grid in the electromagnetic field analysis module, set the electromagnetic field parameters of each component material, and import the BH DC magnetization curve and BP iron loss curve;

[0014] S1.3. Construct a rectangular solution domain that surrounds the simulation geometry model, apply zero tangential magnetic field boundary conditions on the six faces of the rectangular solution domain, and set up an external circuit equivalently according to the connection group, number of winding turns, winding voltage and resistance of the transformer winding to achieve "electromagnetic field-circuit" coupling and perform transient field calculation;

[0015] S1.4. In the fluid temperature field analysis module, the simulation geometry model is divided into a fluid field calculation grid, the thermal parameters of each component material are set, the equivalent boundary conditions of ambient temperature, atmospheric pressure and wind speed are set, the heat dissipation coefficient of the oil tank shell is calculated, and the construction of the "electromagnetic-fluid temperature field" coupling model is completed; the core loss and winding loss obtained by the electromagnetic field calculation are used as the initial heat source, and the Couple solver is selected for steady-state field calculation.

[0016] As a preferred embodiment of the present invention, in S1.2, the electromagnetic field parameters of each component material include density, relative magnetic permeability and volume conductivity; in S1.4, the thermal parameters of each component material include density, thermal conductivity, specific heat capacity and viscosity, and the heat dissipation coefficient of the oil tank shell is calculated using the basic principles of heat transfer.

[0017] As a preferred solution of the present invention, in S2, judging whether it is necessary to readjust the simulation geometric model according to the comparison situation includes:

[0018] S2.1. Taking the main circuit current, load loss and no-load loss of the transformer under rated working conditions as standard reference quantities, the electromagnetic field simulation results are compared with the main circuit current, no-load loss and load loss of the high-voltage winding and the low-voltage winding actually measured. When the error between the two is within 5%, the simulation geometric model does not need to be adjusted;

[0019] S2.2. Use the transformer short-circuit method to perform a temperature rise test to measure the hotspot temperature, top oil temperature and shell temperature of the high-voltage winding and low-voltage winding. Compare the measured test values ​​with the simulation values ​​of the "electromagnetic-fluid temperature field" coupling model. When the error between the two is within 5%, the simulation geometry model does not need to be adjusted.

[0020] S2.3. Use a copper wire with the same coil specifications as the low-voltage winding to make a short-circuit turn, and put it on the upper end of the low-voltage winding of the A-phase iron core. Use a current transformer to measure the short-circuit turn current. During the test, open the high-voltage winding, and detect and record the short-circuit turn current when different currents flow through the low-voltage winding; set an inter-turn short-circuit fault equivalently in the simulation geometry model, and compare the short-circuit loop current simulation value with the test value. When the error between the two is within 5%, the simulation geometry model does not need to be adjusted;

[0021] S2.4. When adjustments are needed, the simulation geometry model can be adjusted by correcting the dimensional parameters of each component, adjusting the electromagnetic parameters of each component material, adjusting the thermal parameters of each component material, optimizing the meshing, and adjusting the boundary conditions and external circuit settings.

[0022] As a preferred solution of the present invention, in the S3, the simulation of the temperature change of the oil tank shell under different active loss levels and different fault positions is specifically to set 100mΩ as the critical resistance value of inter-turn insulation degradation, and 10mΩ as the critical resistance value of inter-turn insulation failure, and the fault identification is only for the case when the inter-turn insulation resistance is 10mΩ~100mΩ;

[0023] In the electromagnetic field analysis module, insulation degradation faults are set at different winding positions, and the inter-turn insulation resistance is set to gradually decrease from normal to 10mΩ. The high-voltage winding main circuit current, low-voltage winding main circuit current, insulation degradation turn current, short-circuit ring current and active power loss are obtained by simulation calculation under different fault positions, different insulation levels and different load rates.

[0024] The losses during the fault are coupled to the "fluid temperature field", the steady-state temperature of the transformer under rated conditions is used as the initial temperature, the SIMPLE transient solver is selected, the solution step size and solution time are set, and the transient field calculation is performed; during the calculation process, the temperature variation characteristics of the high-voltage winding, low-voltage winding, core, top oil temperature, and tank shell with the fault time are recorded.

[0025] As a preferred solution of the present invention, in the above S3, the simulation data of the "electromagnetic-fluid temperature field" coupling model is processed to establish a sample database including active power loss, active power loss change rate, load rate, fault location, and tank shell temperature change rate. The establishment process is:

[0026] According to the principle of energy conservation, the active power loss of the winding is solved by the difference between the active power of the high-voltage input side and the low-voltage output side. The power loss of each phase of the three-phase winding is:

[0027] ;

[0028] Where P A , P B , P C are the winding losses of phase A, phase B, and phase C, i.e., active power losses; U A , U B , U C are the phase voltages of the high-voltage side windings of phase A, phase B, and phase C respectively; I A ,I B ,I C They are the phase currents of the high voltage side winding of phase A, phase B and phase C respectively; , , are the power factor angles of the high-voltage side windings of phase A, phase B, and phase C respectively; U a , U b , U c are the phase voltages of the low-voltage windings of phase A, phase B, and phase C respectively; I a ,I b ,I c They are the phase currents of the low-voltage side winding of phase A, phase B, and phase C respectively; , , are the power factor angles of the low voltage side windings of phase A, phase B and phase C respectively;

[0029] The total active power loss of the transformer is P S for:

[0030] ;

[0031] The change rate of active loss of each phase winding is:

[0032] ;

[0033] In the formula, , , are the active loss change rates of phase A, phase B, and phase C windings respectively; P AN , P BN , P CN They are the rated active power losses of phase A, phase B and phase C windings respectively;

[0034] Total active power loss change rate for:

[0035] ;

[0036] Where P N is the total rated active power loss;

[0037] Assume that the load rate of the transformer during normal operation is L, and the corresponding rated active losses of the A-phase, B-phase, and C-phase windings are P AN , P BN , P CN , take the maximum temperature of the upper end of the oil tank shell corresponding to the A phase, B phase, and C phase winding at time t0, and record them as , , ;

[0038] Assume that the insulation degradation fault occurs in the i-th turn of the winding at time t0, and the corresponding active losses of the A-phase, B-phase, and C-phase windings are P A , P B , P C At time t1, t minutes after the fault occurs, the maximum temperatures of the upper ends of the oil tank shells of phase A, phase B, and phase C windings are recorded as , , The change rates of the upper end temperature of the oil tank shell corresponding to the A-phase, B-phase, and C-phase windings are , , ,but:

[0039] .

[0040] As a preferred embodiment of the present invention, the value range of L is 0.5~1.5.

[0041] As a preferred solution of the present invention, in said S4, the thresholds of the active loss change rate and the tank shell temperature change rate when the inter-turn insulation degradation fault of the winding occurs are specifically determined by setting two thresholds of the three-phase winding active loss change rate, which are respectively , ,in Take the active loss change rate of each phase winding when the insulation resistance between turns of the winding is 100mΩ, Take the active loss change rate of each phase winding when the insulation resistance between turns of the winding is 10mΩ, ;use express , , or , that is, X = A, B, C, S, when When the winding insulation is normal; when When , the inter-turn insulation degradation fault occurs; when When the insulation between winding turns fails.

[0042] As a preferred embodiment of the present invention, in S5, the method for identifying the insulation state between turns of the transformer winding is:

[0043] Using voltage transformer to obtain U A , U B , U C and U a , U b , U c , using current transformer to obtain I A ,I B ,I C and I a ,I b ,I c , calculate their active power loss P respectively A , P B , P C , P S , and then calculate the active power loss change rate , , , ,when When the winding insulation is normal; when When the insulation between the winding turns fails, the transformer relay protection system sends a trip signal to immediately cut off the fault; When the inter-turn insulation degradation fault occurs, an early warning signal is issued, the temperature sensor is started to measure the temperature of the oil tank shell, and further judgment is made:

[0044] if , determine that the insulation degradation fault occurs in the A-phase winding, use the temperature sensor to continuously measure the temperature of the oil tank shell, and calculate the temperature change rate of the A-phase winding oil tank shell Then, the neural network algorithm is used to find the closest match in the established sample database and find the A and The matching fault position, assuming the fault position is j, means that the insulation degradation fault occurs in the jth turn of the A-phase winding;

[0045] if , judge that the insulation degradation fault occurs in the B-phase winding, use the temperature sensor to continuously measure the temperature of the oil tank shell, and calculate the temperature change rate of the B-phase winding oil tank shell Then, the neural network algorithm is used to find the closest match in the established sample database and find the B and The matching fault position, assuming the fault position is j, indicates that the insulation degradation fault occurs in the jth turn of the B-phase winding;

[0046] if , judge that the C-phase winding has insulation degradation fault, use the temperature sensor to continuously measure the temperature of the oil tank shell, and calculate the temperature change rate of the C-phase winding oil tank shell Then, the neural network algorithm is used to find the closest match in the established sample database and find the C and If the fault position is j, it means that the insulation degradation fault occurs in the jth turn of the C-phase winding.

[0047] As a preferred solution of the present invention, the neural network algorithm adopts a closest matching algorithm based on feature weighting and local sensitive hashing, and the steps are as follows:

[0048] Step 1: Collect feature data in a sample database as sample data, and collect input feature data measured in real time to form a real-time database;

[0049] Step 2: Use the Min-Max normalization method to normalize all feature data to the interval [0, 1] to eliminate dimensional differences;

[0050] Step 3: Assign weights to different types of features based on their importance;

[0051] Step 4: Generate a local sensitive hash function using a random projection method;

[0052] Step 5: Map each sample data in the sample database into a hash table through a local sensitive hash function to form multiple hash buckets;

[0053] Step 6: Map the input feature data in the real-time database into the hash bucket through the same local sensitive hash function;

[0054] Step 7: Find the sample data mapped to the same hash bucket as the input feature data from the hash table to form a candidate set;

[0055] Step 8: For each sample data in the candidate set, calculate its weighted Euclidean distance with the input feature data;

[0056] Step 9: Select the sample data with the smallest distance as the closest match.

[0057] The simulation and algorithm involved in the present invention can be executed by an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The simulation and algorithm calculation mentioned above are implemented by executing the software by the processor.

[0058] The beneficial effects of the present invention are:

[0059] The present invention can detect the faulty phase winding by the active loss change rate in the early stage of the inter-turn insulation degradation fault of the winding, that is, when the main circuit current does not change significantly, and judge the position of the faulty turn by the change rate of the oil tank shell temperature, which makes up for the shortcomings of the traditional detection method and avoids further expansion of the fault. The voltage, current and active loss can be obtained by using the voltage transformer and current transformer of the distribution transformer itself, without adding other equipment, which effectively reduces the equipment cost.

[0060] The present invention provides a detection method based on active power loss and temperature, which can be detected online, enriches the detection method of transformer winding interturn insulation degradation fault, and provides new technical support for the safe operation of the power system. Compared with traditional electrical protection devices and detection methods, the present invention uses the changing characteristics of active power loss and oil tank shell temperature to provide higher detection sensitivity and accuracy, which helps to timely discover and locate insulation degradation faults. By identifying faults in advance, preventive and maintenance measures can be taken in time to prevent the continued development of insulation degradation faults, reduce the scope of faults and their hazards, and improve the operational reliability and safety of transformers. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a schematic diagram of the process of the present invention;

[0062] Figure 2 Flow chart for estimating the fault location of insulation degradation between turns of winding. DETAILED DESCRIPTION

[0063] The embodiments of the present invention are further described below in conjunction with the accompanying drawings:

[0064] like Figure 1 As shown, the transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature includes the following steps:

[0065] S1. Use 3D drawing software (ANSYS, SolidWorks, etc.) to construct a 3D geometric model of the original size of the transformer, import the 3D geometric model into finite element simulation software to obtain a simulation geometric model, and establish an "electromagnetic-fluid temperature field" coupling model of the transformer based on the simulation geometric model;

[0066] S2. Based on the transformer entity test, the simulation results of the "electromagnetic-fluid temperature field" coupling model are compared and verified, and it is determined whether the simulation geometry model needs to be readjusted according to the comparison results;

[0067] S3. Based on the simulated geometric model after comparison and verification, simulate the temperature change of the fuel tank shell under different active power loss levels and different fault locations, and establish a sample database of active power loss, fault location, and fuel tank shell temperature change rate;

[0068] S4, determining the thresholds of the rate of change of active power loss and the rate of change of oil tank casing temperature when a fault of inter-turn insulation degradation of the winding occurs;

[0069] S5. Based on the determined threshold value and sample database, during the operation of the transformer, the fault location of the insulation degradation between turns of the winding is estimated according to the winding voltage and current obtained by real-time measurement and the temperature change rate of the oil tank shell measured at the time of the fault.

[0070] This embodiment is mainly aimed at three-phase transformers, especially three-phase oil-immersed distribution transformers. For single-phase transformers, since the parameters of each phase of the three-phase transformer are usually the same, the fault position of the insulation degradation between the winding turns can be estimated by using any phase.

[0071] In S1, the specific process of establishing the "electromagnetic-fluid temperature field" coupling model of the transformer is:

[0072] S1.1. Use 3D drawing software to construct a 3D geometric model (high precision) of the original size of the transformer, including the high-voltage winding, low-voltage winding, iron core, end insulation, clamps, transformer oil and oil tank shell. Considering the actual structure of the low-voltage winding, the low-voltage winding is modeled and numbered with each turn as the smallest unit;

[0073] S1.2. Import the three-dimensional geometric model into ANSYS finite element simulation software to obtain the simulation geometric model, divide the electromagnetic field calculation grid in the electromagnetic field analysis module, set the electromagnetic field parameters of each component material, consider the nonlinear characteristics of the core material, import the BH DC magnetization curve and BP iron loss curve to accurately characterize its electromagnetic characteristics;

[0074] S1.3. Construct a rectangular solution domain that surrounds the simulation geometry model, apply zero tangential magnetic field boundary conditions on the six faces of the rectangular solution domain, and set up an external circuit equivalently according to the connection group, number of winding turns, winding voltage and resistance of the transformer winding to achieve "electromagnetic field-circuit" coupling and perform transient field calculation;

[0075] S1.4. In the fluid temperature field analysis module, the simulation geometry model is divided into a fluid field calculation grid, the thermal parameters of each component material are set, the equivalent boundary conditions of ambient temperature, atmospheric pressure and wind speed are set, the heat dissipation coefficient of the oil tank shell is calculated, and the construction of the "electromagnetic-fluid temperature field" coupling model is completed; the core loss and winding loss obtained by the electromagnetic field calculation are used as the initial heat source, and the Couple solver is selected for steady-state field calculation.

[0076] The BH DC magnetization curve (BH curve) is a curve that describes the magnetization characteristics of ferromagnetic materials in a DC magnetic field. It reflects the degree of magnetization of the material under different magnetic field strengths and is one of the basic characteristics of ferromagnetic materials. The BP iron loss curve is a curve that describes the relationship between the loss characteristics of ferromagnetic materials in an AC magnetic field and the magnetic induction intensity. It reflects the loss of the material under different magnetic induction intensities and is an important tool for evaluating the performance of ferromagnetic materials in AC applications.

[0077] In S1.2, the electromagnetic field parameters of each component material include density, relative magnetic permeability and volume conductivity; in S1.4, the thermal parameters of each component material include density, thermal conductivity, specific heat capacity and viscosity. The heat dissipation coefficient of the fuel tank shell is calculated using the basic principles of heat transfer (known heat dissipation coefficient calculation formula).

[0078] In S2, it is determined whether the simulation geometric model needs to be readjusted according to the comparison results, including:

[0079] S2.1. According to the technical parameters of the transformer, the main circuit current, load loss and no-load loss of the transformer under rated conditions are used as standard reference quantities. The electromagnetic field simulation results are compared with the main circuit current, no-load loss and load loss of the high-voltage winding and the low-voltage winding actually measured. When the error between the two is within 5% (which can be adjusted according to actual conditions, the same below), the simulation geometry model does not need to be adjusted;

[0080] S2.2. Use the transformer short-circuit method to perform a temperature rise test to measure the hotspot temperature, top oil temperature and shell temperature of the high-voltage winding and low-voltage winding. Compare the measured test values ​​with the simulation values ​​of the "electromagnetic-fluid temperature field" coupling model. When the error between the two is within 5%, the simulation geometry model does not need to be adjusted.

[0081] S2.3. To simulate the inter-turn insulation degradation fault of the low-voltage winding, according to the actual structure of the transformer low-voltage winding, a copper wire with the same coil specification as the low-voltage winding is selected to manufacture a short-circuit turn, and it is mounted on the upper end of the low-voltage winding of the A-phase iron core. A current transformer is used to measure the short-circuit turn current. During the test, the high-voltage winding is open-circuited, and the short-circuit turn current when different currents flow through the low-voltage winding is detected and recorded; an inter-turn short-circuit fault is set equivalently in the simulation geometry model, and the simulation value of the short-circuit ring current is compared with the test value. When the error between the two is within 5%, the simulation geometry model does not need to be adjusted;

[0082] S2.4. When adjustments are needed, the simulation geometry model can be adjusted by correcting the dimensional parameters of each component, adjusting the electromagnetic parameters of each component material, adjusting the thermal parameters of each component material, optimizing the meshing, and adjusting the boundary conditions and external circuit settings.

[0083] In S3, the temperature change of the oil tank shell under different active loss levels and different fault locations is simulated as follows: according to the electromagnetic field simulation results of the transformer interturn insulation degradation fault and actual experience, 100mΩ can be set as the critical resistance value of interturn insulation degradation, and 10mΩ can be set as the critical resistance value of interturn insulation failure. Fault identification is only for the case when the interturn insulation resistance is 10mΩ~100mΩ;

[0084] When the inter-turn insulation resistance is greater than 100mΩ, only the temperature of the faulty turn increases, while the top oil temperature and the tank shell temperature remain almost unchanged, and the transformer can maintain normal operation. When the inter-turn insulation resistance is less than 10mΩ, the temperature of the faulty turn increases sharply, reaching thousands of degrees in a very short time, causing the copper wire to fuse, making it difficult to take effective protection measures in time;

[0085] In the electromagnetic field analysis module, insulation degradation faults are set at different winding positions, and the inter-turn insulation resistance is set to gradually decrease from normal to 10mΩ. The high-voltage winding main circuit current, low-voltage winding main circuit current, insulation degradation turn current, short-circuit ring current and active power loss are obtained by simulation calculation under different fault positions, different insulation levels and different load rates.

[0086] The losses during the fault are coupled to the "fluid temperature field", the steady-state temperature of the transformer under rated conditions is used as the initial temperature, the SIMPLE transient solver is selected, the solution step size and solution time are set, and the transient field calculation is performed; during the calculation process, the temperature variation characteristics of the high-voltage winding, low-voltage winding, core, top oil temperature, and tank shell with the fault time are recorded.

[0087] In S3, the simulation data of the "electromagnetic-fluid temperature field" coupling model is processed to establish a sample database including active power loss, active power loss change rate, load rate, fault location, and tank shell temperature change rate. The establishment process is as follows:

[0088] According to the principle of energy conservation, the active power loss of the winding is solved by the difference between the active power of the high-voltage input side and the low-voltage output side. The power loss of each phase of the three-phase winding is:

[0089] ;

[0090] Where P A , P B , P C are the winding losses of phase A, phase B, and phase C, i.e., active power losses; U A , U B , U C are the phase voltages of the high-voltage side windings of phase A, phase B, and phase C respectively; I A ,I B ,I C They are the phase currents of the high voltage side winding of phase A, phase B and phase C respectively; , , are the power factor angles of the high-voltage side windings of phase A, phase B, and phase C respectively; U a , U b , U c are the phase voltages of the low-voltage windings of phase A, phase B, and phase C respectively; I a ,I b ,I c They are the phase currents of the low-voltage side winding of phase A, phase B, and phase C respectively; , , are the power factor angles of the low voltage side windings of phase A, phase B and phase C respectively;

[0091] The total active power loss of the transformer is P S for:

[0092] ;

[0093] The change rate of active loss of each phase winding is:

[0094] ;

[0095] In the formula, , , are the active loss change rates of phase A, phase B, and phase C windings respectively; P AN , P BN , P CN They are the rated active power losses of phase A, phase B and phase C windings respectively;

[0096] Total active power loss change rate for:

[0097] ;

[0098] Where P N is the total rated active power loss;

[0099] Assume that the load rate of the transformer during normal operation is L (the value range is 0.5~1.5), and the corresponding rated active losses of the A-phase, B-phase, and C-phase windings are P respectively. AN , P BN , P CN , take the maximum temperature of the upper end of the oil tank shell corresponding to the A phase, B phase, and C phase winding at time t0, and record them as , , ;

[0100] Assume that the insulation degradation fault occurs in the i-th turn of the winding at time t0, and the corresponding active losses of the A-phase, B-phase, and C-phase windings are P A , P B , P C At time t1, t minutes after the fault occurs, the maximum temperatures of the upper ends of the oil tank shells of phase A, phase B, and phase C windings are recorded as , , The change rates of the upper end temperature of the oil tank shell corresponding to the A-phase, B-phase, and C-phase windings are , , ,but:

[0101] .

[0102] For different load rates and different inter-turn insulation resistances, each possible fault location, that is, each turn of the winding, is calculated separately to obtain the corresponding active power loss, active power loss change rate, and temperature change rate of the upper end of the fuel tank shell under various inter-turn insulation resistances at different load rates. By summarizing various types of data, a complete sample database including winding current, winding voltage, load rate, inter-turn insulation resistance, fault location, active power loss, active power loss change rate, and temperature change rate of the upper end of the fuel tank shell can be obtained.

[0103] In S4, the thresholds of the active power loss change rate and the tank shell temperature change rate when the inter-turn insulation degradation fault of the winding occurs are determined. Specifically, two thresholds of the three-phase winding active power loss change rate are set, which are , ,in Take the active loss change rate of each phase winding when the insulation resistance between turns of the winding is 100mΩ ( , , are usually the same), Take the active loss change rate of each phase winding when the insulation resistance between turns of the winding is 10mΩ, ;use express , , or , that is, X = A, B, C, S, when When the winding insulation is normal; when When , the inter-turn insulation degradation fault occurs; when When the insulation between winding turns fails.

[0104] When the inter-turn insulation degradation position is different, the temperature change rate of the oil tank shell in a certain period of time is also different. The closer to the upper end of the winding, the greater the temperature change rate of the oil tank shell. Therefore, when the load rate of the distribution transformer is L, when an insulation degradation fault occurs at position i, its active loss is P, and the temperature change rate of the top of the oil tank shell is taken as When the actual measured temperature change When , it indicates that insulation degradation fault occurs at the i-th turn of the winding.

[0105] For S13-250kVA oil-immersed distribution transformer, under rated conditions, , The values ​​can be 1% and 14% respectively.

[0106] In S5, the method for identifying the insulation state between turns of the transformer winding is:

[0107] like Figure 2 As shown in the figure, during the operation of the transformer, the voltage transformer is used to obtain U A , U B , U C and U a , U b , U c , using current transformer to obtain I A ,I B ,I C and I a ,I b ,I c , calculate their active power loss P respectively A , P B , P C , P S , and then calculate the active power loss change rate , , , ,when When the winding insulation is normal; when When the insulation between the winding turns fails, the transformer relay protection system (existing system) sends a trip signal to immediately cut off the fault; when When the inter-turn insulation degradation fault occurs, an early warning signal is issued, the temperature sensor is started to measure the temperature of the oil tank shell, and further judgment is made:

[0108] if , determine that the insulation degradation fault occurs in the A-phase winding, use the temperature sensor to continuously measure the temperature of the oil tank shell, and calculate the temperature change rate of the A-phase winding oil tank shell (Reference The calculation formula can be obtained, and the rest is the same), and then the neural network algorithm is used to find the closest match in the established sample database to find the closest match with P A and The matching fault position, assuming the fault position is j, means that the insulation degradation fault occurs in the jth turn of the A-phase winding;

[0109] if , judge that the insulation degradation fault occurs in the B-phase winding, use the temperature sensor to continuously measure the temperature of the oil tank shell, and calculate the temperature change rate of the B-phase winding oil tank shell Then, the neural network algorithm is used to find the closest match in the established sample database and find the B and The matching fault position, assuming the fault position is j, indicates that the insulation degradation fault occurs in the jth turn of the B-phase winding;

[0110] if , judge that the C-phase winding has insulation degradation fault, use the temperature sensor to continuously measure the temperature of the oil tank shell, and calculate the temperature change rate of the C-phase winding oil tank shell Then, the neural network algorithm is used to find the closest match in the established sample database and find the C and If the fault position is j, it means that the insulation degradation fault occurs in the jth turn of the C-phase winding.

[0111] The neural network algorithm uses the closest matching algorithm based on feature weighting and local sensitive hashing. The steps are as follows:

[0112] Step 1: Collect feature data in a sample database as sample data, and collect input feature data measured in real time to form a real-time database;

[0113] Step 2: Use the Min-Max normalization method to normalize all feature data to the interval [0, 1] to eliminate dimensional differences;

[0114] Step 3: Assign weights to different types of features based on their importance;

[0115] Step 4: Generate a local sensitive hash function using a random projection method;

[0116] Step 5: Map each sample data in the sample database into a hash table through a local sensitive hash function to form multiple hash buckets;

[0117] Step 6: Map the input feature data in the real-time database into the hash bucket through the same local sensitive hash function;

[0118] Step 7: Find the sample data mapped to the same hash bucket as the input feature data from the hash table to form a candidate set;

[0119] Step 8: For each sample data in the candidate set, calculate its weighted Euclidean distance with the input feature data;

[0120] Step 9: Select the sample data with the smallest distance as the closest match.

[0121] Other well-known neural network algorithms may also be used for matching.

Claims

1. A transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature is characterized by The following steps are involved: S1. Use 3D drawing software to construct a 3D geometric model of the original size of the transformer, import the 3D geometric model into finite element simulation software to obtain a simulation geometric model, and establish an "electromagnetic-fluid temperature field" coupling model of the transformer based on the simulation geometric model; S2. Based on the transformer entity test, the simulation results of the "electromagnetic-fluid temperature field" coupling model are compared and verified, and it is determined whether the simulation geometry model needs to be readjusted according to the comparison results; S3. Based on the simulated geometric model after comparison and verification, simulate the temperature change of the fuel tank shell under different active power loss levels and different fault locations, and establish a sample database of active power loss, fault location, and fuel tank shell temperature change rate; S4, determining the thresholds of the rate of change of active power loss and the rate of change of oil tank casing temperature when a fault of inter-turn insulation degradation of the winding occurs; S5. Based on the determined threshold value and sample database, during the operation of the transformer, the fault location of the insulation degradation between turns of the winding is estimated according to the winding voltage and current obtained by real-time measurement and the temperature change rate of the oil tank shell measured at the time of the fault.

2. The transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature according to claim 1 is characterized in that: In the above S1, the specific process of establishing the "electromagnetic-fluid temperature field" coupling model of the transformer is as follows: S1.

1. Use 3D drawing software to construct a 3D geometric model of the transformer in its original size, including the high-voltage winding, low-voltage winding, iron core, end insulation, clamps, transformer oil and oil tank shell, and model and number the low-voltage winding with each turn as the smallest unit; S1.2, import the three-dimensional geometric model into ANSYS finite element simulation software, obtain the simulation geometric model, divide the electromagnetic field calculation grid in the electromagnetic field analysis module, set the electromagnetic field parameters of each component material, and import the BH DC magnetization curve and BP iron loss curve; S1.

3. Construct a rectangular solution domain that surrounds the simulation geometry model, apply zero tangential magnetic field boundary conditions on the six faces of the rectangular solution domain, and set up an external circuit equivalently according to the connection group, number of winding turns, winding voltage and resistance of the transformer winding to achieve "electromagnetic field-circuit" coupling and perform transient field calculation; S1.

4. In the fluid temperature field analysis module, the simulation geometry model is divided into a fluid field calculation grid, the thermal parameters of each component material are set, the equivalent boundary conditions of ambient temperature, atmospheric pressure and wind speed are set, the heat dissipation coefficient of the oil tank shell is calculated, and the construction of the "electromagnetic-fluid temperature field" coupling model is completed; the core loss and winding loss obtained by the electromagnetic field calculation are used as the initial heat source, and the Couple solver is selected for steady-state field calculation.

3. The transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature according to claim 2 is characterized in that: In the above S1.2, the electromagnetic field parameters of each component material include density, relative magnetic permeability and volume conductivity; in S1.4, the thermal parameters of each component material include density, thermal conductivity, specific heat capacity and viscosity, and the heat dissipation coefficient of the oil tank shell is calculated using the basic principles of heat transfer.

4. The transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature according to claim 1 is characterized in that: In the above S2, judging whether it is necessary to readjust the simulation geometric model according to the comparison situation includes: S2.

1. Taking the main circuit current, load loss and no-load loss of the transformer under rated working conditions as standard reference quantities, the electromagnetic field simulation results are compared with the main circuit current, no-load loss and load loss of the high-voltage winding and the low-voltage winding actually measured. When the error between the two is within 5%, the simulation geometric model does not need to be adjusted; S2.

2. Use the transformer short-circuit method to perform a temperature rise test to measure the hotspot temperature, top oil temperature and shell temperature of the high-voltage winding and low-voltage winding. Compare the measured test values ​​with the simulation values ​​of the "electromagnetic-fluid temperature field" coupling model. When the error between the two is within 5%, the simulation geometry model does not need to be adjusted. S2.

3. Use a copper wire with the same coil specifications as the low-voltage winding to make a short-circuit turn, and put it on the upper end of the low-voltage winding of the A-phase iron core. Use a current transformer to measure the short-circuit turn current. During the test, open the high-voltage winding, and detect and record the short-circuit turn current when different currents flow through the low-voltage winding; set an inter-turn short-circuit fault equivalently in the simulation geometry model, and compare the short-circuit loop current simulation value with the test value. When the error between the two is within 5%, the simulation geometry model does not need to be adjusted; S2.

4. When adjustments are needed, the simulation geometry model can be adjusted by correcting the dimensional parameters of each component, adjusting the electromagnetic parameters of each component material, adjusting the thermal parameters of each component material, optimizing the meshing, and adjusting the boundary conditions and external circuit settings.

5. The transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature according to claim 1 is characterized in that: In the above S3, the temperature change of the oil tank shell under different active power loss levels and different fault positions is simulated by setting 100mΩ as the critical resistance value of inter-turn insulation degradation and 10mΩ as the critical resistance value of inter-turn insulation failure. Fault identification is only for the case when the inter-turn insulation resistance is 10mΩ~100mΩ; In the electromagnetic field analysis module, insulation degradation faults are set at different winding positions, and the inter-turn insulation resistance is set to gradually decrease from normal to 10mΩ. The high-voltage winding main circuit current, low-voltage winding main circuit current, insulation degradation turn current, short-circuit ring current and active power loss are obtained by simulation calculation under different fault positions, different insulation levels and different load rates. The loss during the fault is coupled to the "fluid temperature field", the steady-state temperature of the transformer under rated conditions is used as the initial temperature, the SIMPLE transient solver is selected, the solution step size and solution time are set, and the transient field calculation is performed; during the calculation process, the variation characteristics of the high-voltage winding, low-voltage winding, core, top oil temperature, and tank shell temperature with the fault time are recorded.

6. The transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature according to claim 5 is characterized in that: In the above S3, the simulation data of the "electromagnetic-fluid temperature field" coupling model is processed to establish a sample database including active power loss, active power loss change rate, load rate, fault location, and tank shell temperature change rate. The establishment process is: According to the principle of energy conservation, the active power loss of the winding is solved by the difference between the active power of the high-voltage input side and the low-voltage output side. The power loss of each phase of the three-phase winding is: ； Where P A , P B , P C are the winding losses of phase A, phase B, and phase C, i.e., active power losses; U A , U B , U C are the phase voltages of the high-voltage side windings of phase A, phase B, and phase C respectively; I A ,I B ,I C They are the phase currents of the high voltage side winding of phase A, phase B and phase C respectively; , , are the power factor angles of the high-voltage side windings of phase A, phase B, and phase C respectively; U a , U b , U c are the phase voltages of the low-voltage windings of phase A, phase B, and phase C respectively; I a ,I b ,I c They are the phase currents of the low voltage side winding of phase A, phase B and phase C respectively; , , are the power factor angles of the low voltage side windings of phase A, phase B and phase C respectively; The total active power loss of the transformer is P S for: ； The change rate of active loss of each phase winding is: ； In the formula, , , are the active loss change rates of phase A, phase B, and phase C windings respectively; P AN , P BN , P CN They are the rated active power losses of phase A, phase B and phase C windings respectively; Total active power loss change rate for: ； Where P N is the total rated active power loss; Assume that the load rate of the transformer during normal operation is L, and the corresponding rated active losses of the A-phase, B-phase, and C-phase windings are P AN , P BN , P CN , take the maximum temperature of the upper end of the oil tank shell corresponding to the A phase, B phase, and C phase winding at time t0, and record them as , , ; Assume that the insulation degradation fault occurs in the i-th turn of the winding at time t0, and the corresponding active losses of the A-phase, B-phase, and C-phase windings are P A , P B , P C At time t1, t minutes after the fault occurs, the maximum temperatures of the upper ends of the oil tank shells of phase A, phase B, and phase C windings are recorded as , , The change rates of the upper end temperature of the oil tank shell corresponding to the A-phase, B-phase, and C-phase windings are , , ,but: 。 7. The transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature according to claim 6 is characterized in that: The value range of L is 0.5~1.

5.

8. The transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature according to claim 6 is characterized in that: In the above S4, the thresholds of the active power loss change rate and the tank shell temperature change rate when the inter-turn insulation degradation fault of the winding occurs are determined specifically by setting two thresholds of the three-phase winding active power loss change rate, which are respectively , ,in Take the active loss change rate of each phase winding when the insulation resistance between turns of the winding is 100mΩ, Take the active loss change rate of each phase winding when the insulation resistance between turns of the winding is 10mΩ, ;use express , , or , that is, X = A, B, C, S, when When the winding insulation is normal; when When , the inter-turn insulation degradation fault occurs; when When the insulation between winding turns fails.

9. The transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature according to claim 8 is characterized in that: In the above S5, the method for identifying the insulation state between turns of the transformer winding is: Using voltage transformer to obtain U A , U B , U C and U a , U b , U c , using current transformer to obtain I A ,I B ,I C and I a ,I b ,I c , calculate their active power loss P respectively A , P B , P C , P S , and then calculate the active power loss change rate , , , ,when When the winding insulation is normal; when When the insulation between the winding turns fails, the transformer relay protection system sends a trip signal to immediately cut off the fault; When the inter-turn insulation degradation fault occurs, an early warning signal is issued, the temperature sensor is started to measure the temperature of the oil tank shell, and further judgment is made: if , determine that the insulation degradation fault occurs in the A-phase winding, use the temperature sensor to continuously measure the temperature of the oil tank shell, and calculate the temperature change rate of the A-phase winding oil tank shell Then, the neural network algorithm is used to find the closest match in the established sample database and find the A and The matching fault position, assuming the fault position is j, means that the insulation degradation fault occurs in the jth turn of the A-phase winding; if , judge that the insulation degradation fault occurs in the B-phase winding, use the temperature sensor to continuously measure the temperature of the oil tank shell, and calculate the temperature change rate of the B-phase winding oil tank shell Then, the neural network algorithm is used to find the closest match in the established sample database and find the B and The matching fault position, assuming the fault position is j, indicates that the insulation degradation fault occurs in the jth turn of the B-phase winding; if , judge that the C-phase winding has insulation degradation fault, use the temperature sensor to continuously measure the temperature of the oil tank shell, and calculate the temperature change rate of the C-phase winding oil tank shell Then, the neural network algorithm is used to find the closest match in the established sample database and find the C and If the fault position is j, it means that the insulation degradation fault occurs in the jth turn of the C-phase winding.

10. The transformer inter-turn insulation degradation fault simulation identification method based on the box cover temperature according to claim 9 is characterized in that: The neural network algorithm uses the closest matching algorithm based on feature weighting and local sensitive hashing. The steps are as follows: Step 1: Collect feature data in a sample database as sample data, and collect input feature data measured in real time to form a real-time database; Step 2: Use the Min-Max normalization method to normalize all feature data to the interval [0, 1] to eliminate dimensional differences; Step 3: Assign weights to different types of features based on their importance; Step 4: Generate a local sensitive hash function using a random projection method; Step 5: Map each sample data in the sample database into a hash table through a local sensitive hash function to form multiple hash buckets; Step 6: Map the input feature data in the real-time database into the hash bucket through the same local sensitive hash function; Step 7: Find the sample data mapped to the same hash bucket as the input feature data from the hash table to form a candidate set; Step 8: For each sample data in the candidate set, calculate its weighted Euclidean distance with the input feature data; Step 9: Select the sample data with the smallest distance as the closest match.

Images