Winding change degree monitoring method and device, medium and equipment
By comparing the sample characteristics during normal operation and the measured characteristics during faults when the transformer fails, the problem that the existing technology cannot accurately evaluate the degree of winding changes is solved, real-time monitoring and accurate evaluation of winding changes is achieved, and the reliable operation of the transformer is ensured.
Patent Information
- Application Number
- CN202411144461.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-20
- Publication Date
- 2024-11-08
AI Technical Summary
The prior art cannot accurately evaluate the degree of change in the transformer winding after a failure, which may lead to unnecessary power outage inspections or fault miss inspections, and may even cause major accidents.
By collecting sample vibration signals and electrical signals during normal operation of the transformer, dividing them into sub-sample signals, and setting the embedded collection representative calculation feature vector, comparing the difference between the actual measured signal and the target reference segment in the event of a fault, and determining the degree of winding change.
Real-time monitoring and accurate evaluation of winding changes is achieved, problem data is identified in a timely manner, and the reliable operation of the transformer is ensured and the risk of failure is reduced.
Smart Images

Figure CN118914928A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of high voltage testing of electrical equipment, and in particular to a method, device, medium and equipment for monitoring the degree of winding changes. Background Art
[0002] During operation, the transformer may be affected by faults such as short-circuit shocks. At this time, the winding state of the transformer will change to a certain extent. When the winding state changes to a large extent, a power outage inspection is required.
[0003] However, in the prior art, there is no algorithm that can accurately evaluate the degree of change in the winding in real time after a fault, which may cause unnecessary power outages for inspection, may also cause missed faults, and may even lead to major accidents. Summary of the invention
[0004] Based on this, it is necessary to provide a method, device, medium and equipment for monitoring the degree of winding change to solve the problem that there is no algorithm that can accurately evaluate the degree of winding change in real time after a fault.
[0005] A method for monitoring winding variation degree, the method comprising:
[0006] When the transformer is in a normal operating state, a sample vibration signal and a sample electrical signal of the transformer are obtained; wherein the sample vibration signal and the sample electrical signal are sampled simultaneously, and the sampling duration and sampling frequency are the same;
[0007] Dividing the sample vibration signal into N segments of sub-sample vibration signals at equal intervals, and dividing the sample electrical signal into N segments of sub-sample electrical signals at equal intervals; wherein N≥2;
[0008] Setting X uniformly distributed embedding collection representatives, respectively calculating the distance between each sub-sample vibration signal and the X embedding collection representatives, to obtain N sample feature vectors;
[0009] When the transformer is in a faulty operation state, a measured vibration signal and a measured electrical signal of the transformer are obtained; wherein the measured vibration signal and the measured electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same as the sampling duration of the sub-sample vibration signal and the sub-sample electrical signal;
[0010] Calculating the distances between the measured electrical signal and the N segments of sub-sample electrical signals respectively to obtain N candidate reference distances, taking a segment corresponding to a minimum value among the N candidate reference distances as a target reference segment, and obtaining a sample feature vector of the target reference segment from the N sample feature vectors;
[0011] Calculating the distance between the measured vibration signal and the X embedded set representatives to obtain a corresponding measured feature vector;
[0012] When the difference between the measured feature vector and the sample feature vector of the target reference segment meets a preset difference condition, determining that the measured vibration signal is problematic data;
[0013] The degree of change of the winding is determined based on the proportion of the problem data within a preset time period.
[0014] In one embodiment, setting X uniformly distributed embedding set representatives includes:
[0015] The maximum amplitude and the minimum amplitude of the sample vibration signal are obtained, and X-2 vibration amplitudes are uniformly inserted between the maximum amplitude and the minimum amplitude to obtain X embedded set representatives.
[0016] In one embodiment, respectively calculating the distance between each sub-sample vibration signal and the X embedded set representatives to obtain N sample feature vectors includes:
[0017] In each sub-sample vibration signal, the M sampling points are sequentially taken as a unit to calculate the average value of the sampling point amplitude, so as to obtain the M sample amplitude mean values; wherein M ≥ 2;
[0018] Calculate the distance between the mean value of each sample amplitude of each sub-sample vibration signal and the X embedded set representatives to obtain N sample feature vectors;
[0019] The calculating the distance between the measured vibration signal and the X embedded set representatives to obtain a corresponding measured feature vector includes:
[0020] The measured vibration signal is calculated with M sampling points as a unit to average the amplitude of the sampling points, so as to obtain M measured amplitude averages;
[0021] The distances between the M measured amplitude means and the X embedded set representatives are calculated to obtain corresponding measured feature vectors.
[0022] In one embodiment, respectively calculating the distance between the mean value of each sample amplitude of each sub-sample vibration signal and the X embedded set representatives to obtain N sample feature vectors includes:
[0023] In each sub-sample vibration signal, respectively calculating the absolute value of the difference between the mean amplitude of each sample and the X embedded set representatives to obtain X first absolute values of the difference;
[0024] Taking the minimum values of the X first difference absolute values as the first local eigenvalues of the corresponding sample amplitude mean, respectively, to obtain M first local eigenvalues, and taking the combination of the M first local eigenvalues as the sample eigenvector of the corresponding sub-sample vibration signal, to obtain N sample eigenvectors;
[0025] The calculating the distance between the M measured amplitude means and the X embedded set representatives to obtain the corresponding measured feature vector includes:
[0026] Calculate the absolute value of the difference between each measured amplitude mean and the X embedded set representatives to obtain X second difference absolute values;
[0027] The minimum values among the X second difference absolute values are respectively used as the second local eigenvalues corresponding to the measured amplitude mean value to obtain M second local eigenvalues, and the combination of the M second local eigenvalues is used as the measured eigenvector.
[0028] In one embodiment, the sub-sample electrical signal includes a sample current signal and a sample active power signal, the measured electrical signal includes a measured current signal and a measured active power signal, and the distances between the measured electrical signal and the N segments of sub-sample electrical signals are calculated respectively to obtain N candidate reference distances, including
[0029] respectively calculating normalized vector distances between the measured current signal and N segments of sample current signals to obtain N first reference distances;
[0030] respectively calculating the normalized vector distances between the measured active power signal and the N segments of sample active power signals to obtain N second reference distances;
[0031] The N first reference distances and the N second reference distances are weightedly calculated to obtain N candidate reference distances.
[0032] In one embodiment, when the difference between the measured feature vector and the sample feature vector of the target reference segment satisfies a preset difference condition, determining that the measured vibration signal is problematic data includes:
[0033] When the difference between any two elements of the sample feature vector of the target reference segment and the measured feature vector is greater than a preset first threshold, or when the normalized distance between the sample feature vector of the target reference segment and the measured feature vector is greater than a preset second threshold, the measured vibration signal is determined to be problematic data.
[0034] In one embodiment, the degree of change of the winding includes a first degree of change, a second degree of change, and a third degree of change that increase in sequence, and the determination of the degree of change of the winding based on the proportion of the problem data within a preset time period includes:
[0035] Within p preset unit time lengths, if the proportion of the problem data is less than a preset first proportion threshold, it is determined that the degree of change of the winding is the first degree of change;
[0036] Within q preset unit time lengths, if the proportion of the problem data is greater than or equal to a preset first proportion threshold and less than a preset second proportion threshold, it is determined that the degree of change of the winding is the second degree of change;
[0037] Within q preset unit time lengths, if the proportion of the problem data is greater than a preset second proportion threshold, it is determined that the degree of change of the winding is the third degree of change.
[0038] A winding variation degree monitoring device, the winding variation degree monitoring device comprising:
[0039] A normal operating state monitoring module is used to obtain a sample vibration signal and a sample electrical signal of the transformer when the transformer is in a normal operating state; wherein the sample vibration signal and the sample electrical signal are sampled simultaneously, and the sampling duration and sampling frequency are the same; the sample vibration signal is equally spaced into N segments of sub-sample vibration signals, and the sample electrical signal is equally spaced into N segments of sub-sample electrical signals; wherein N ≥ 2; X uniformly distributed embedded collection representatives are set, and the distance between each segment of the sub-sample vibration signal and the X embedded collection representatives is calculated respectively to obtain N sample feature vectors;
[0040] A fault operation state monitoring module is used to obtain a measured vibration signal and a measured electrical signal of the transformer when the transformer is in a fault operation state; wherein the measured vibration signal and the measured electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same as the sampling duration of the sub-sample vibration signal and the sub-sample electrical signal; respectively calculate the distance between the measured electrical signal and the N segments of sub-sample electrical signals to obtain N candidate reference distances, take a segment corresponding to the minimum value of the N candidate reference distances as a target reference segment, and obtain a sample feature vector of the target reference segment from the N sample feature vectors; calculate the distance between the measured vibration signal and the X embedded collection representatives to obtain the corresponding measured feature vector;
[0041] The winding change degree calculation module is used to determine that the measured vibration signal is problem data when the difference between the measured feature vector and the sample feature vector of the target reference segment meets the preset difference condition; and to determine the degree of change of the winding based on the proportion of the problem data within a preset time length.
[0042] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor executes the steps of the above-mentioned winding variation degree monitoring method.
[0043] A winding variation degree monitoring device comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the winding variation degree monitoring method.
[0044] The present invention provides a method, device, medium and equipment for monitoring the degree of winding change. First, when the transformer is operating normally, sample vibration signals and sample electrical signals are collected simultaneously, and they are divided into N sub-sample signals. Then, X embedded collection representatives are set, and the distance between each sub-sample signal and these representatives is calculated to obtain N sample feature vectors. When the transformer fails, the measured vibration signal and electrical signal are collected, and the distance between the measured signal and the N sub-sample signals is calculated to determine the target reference segment corresponding to the minimum distance, and extract its sample feature vector. By comparing the difference between the measured feature vector and the sample feature vector of the target reference segment, if the difference meets the preset conditions, it is determined to be problem data. The degree of change of the winding is judged based on the proportion of problem data within a preset time length. This method can realize real-time monitoring and accurate evaluation of winding changes by comparing the sample features under normal conditions with the measured features under fault conditions, timely identify problem data, and analyze the health status of the winding according to the proportion of problem data, so as to ensure the reliable operation of the transformer and reduce the risk of failure. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0046] in:
[0047] Figure 1 A schematic diagram of a flow chart of a method for monitoring winding variation degree in one embodiment;
[0048] Figure 2 It is a structural schematic diagram of a winding variation degree monitoring device;
[0049] Figure 3 This is the structural block diagram of the winding variation degree monitoring device. DETAILED DESCRIPTION
[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0051] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.
[0052] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0053] like Figure 1 As shown, Figure 1 : is a flow chart of a method for monitoring winding variation degree in one embodiment. The steps provided by the method for monitoring winding variation degree in this embodiment include:
[0054] S101, when the transformer is in a normal operating state, obtaining a sample vibration signal and a sample electrical signal of the transformer.
[0055] The sample vibration signal refers to the vibration data of the transformer under normal operating conditions, and the sample electrical signal refers to the electrical data such as current, voltage, resistance and power of the transformer under normal operating conditions. In this embodiment, the sample vibration signal and the sample electrical signal are sampled simultaneously, and the sampling duration and sampling frequency are the same.
[0056] Optionally, assuming that a transformer is operating normally in the power grid, we install vibration sensors at key locations of the transformer, such as the core and winding, and connect the sensors. Then, we obtain the transformer's sample vibration signal and sample electrical signal at the same sampling frequency for 48 hours.
[0057] S102 , dividing the sample vibration signal into N segments of sub-sample vibration signals at equal intervals, and dividing the sample electrical signal into N segments of sub-sample electrical signals at equal intervals.
[0058] Where N≥2, we can segment the sample vibration signal and the sample electrical signal at a time interval of 5-10 minutes to obtain N segments of sub-sample vibration signals and N segments of sub-sample electrical signals. Of course, the time interval can also be set to other values according to requirements.
[0059] Exemplarily, if each segment of the signal includes K sampling points, the N segments of sub-sample vibration signals are expressed as:
[0060] V1=(v 11 , v 12 ,···,v 1K )
[0061] V2=(v 21 , v 22 ,···,v 2K )
[0062] ···
[0063] V N =(v N1 , v N2 ,···,v NK )
[0064] If the sub-sample electrical signal includes a sample current signal and a sample active power signal, the N-segment sample current signal is expressed as:
[0065] I1=(i 11 ,i 12 ,···,i 1K )
[0066] I2=(i 21 ,i 22 ,···,i 2K )
[0067] ···
[0068] I N =(i N1 ,i N2 ,···,i NK )
[0069] The N-segment sample active power signal is expressed as:
[0070] P1=(p 11 , p 12 ,···,p 1K ),
[0071] P2=(p21,p 22 ,···,p 2K )
[0072] ···
[0073] P N =(p N1 , p N2 ,···,p NK )
[0074] S103, setting X evenly distributed embedding collection representatives, respectively calculating the distance between each sub-sample vibration signal and the X embedding collection representatives, to obtain N sample feature vectors.
[0075] The X embedded set representatives are reference points of the vibration signal for feature extraction, and the N sample feature vectors are used to describe the similarity or difference between the sub-sample vibration signal and the embedded set representative in the feature space.
[0076] Optionally, a clustering algorithm (such as K-means) may be used to generate X evenly distributed embedding collection representatives. Then, a suitable distance measurement method is selected to calculate the distance between each sub-sample vibration signal and the X embedding collection representatives, such as Euclidean distance, Manhattan distance or cosine similarity. The most suitable method may be selected according to the application scenario.
[0077] In a specific embodiment, setting X uniformly distributed embedded collection representatives specifically includes the following steps: obtaining the maximum amplitude and the minimum amplitude of the sample vibration signal, and uniformly inserting X-2 vibration amplitudes between the maximum amplitude and the minimum amplitude to obtain X embedded collection representatives.
[0078] Here, it is generally recommended that X is 5-10. In this specific embodiment, the uniformly distributed embedding collection represents the entire amplitude range of the sample vibration signal, which helps to fully capture the characteristics of the sample vibration signal. At the same time, the inserted additional amplitude points make the feature vector calculation more detailed, and can better identify and distinguish small changes in the vibration signal. This specific embodiment can make the detection of abnormal signals more sensitive and reliable.
[0079] In a specific embodiment, respectively calculating the distance between each sub-sample vibration signal and the X embedded set representatives to obtain N sample feature vectors specifically includes the following steps:
[0080] A1. In each sub-sample vibration signal, the M sampling points are sequentially taken as a unit to calculate the average value of the sampling point amplitudes to obtain the mean value of the M sample amplitudes.
[0081] Among them, M ≥ 2. We can adjust the size of M according to the amount of data. If the amount of data is too large, we can adjust it to more sampling points as one unit. If the amount of data is too small, we can adjust it to fewer sampling points as one unit.
[0082] Optionally, let M = 5, we can perform the following calculation on the first sub-sample vibration signal:
[0083] {v ave11 , v ave12 ,···,v ave1M}={(v 11 +v 12 +···+v 15 ) / 5,(v 16 +v 17 +···+v 110 ) / 5,(v 1(K-4) +v 1(K-3) +···+v 1K ) / 5}
[0084] The same operation is performed on the remaining sub-sample vibration signals, which will not be described in detail.
[0085] A2. Calculate the distance between the mean value of each sample amplitude of each sub-sample vibration signal and the X embedded set representatives to obtain N sample feature vectors.
[0086] Here, a suitable distance measurement method can be selected to calculate the distance between the mean amplitude of each sample of each sub-sample vibration signal and the X embedded set representatives, such as Euclidean distance, Manhattan distance or cosine similarity. The most suitable method can be selected according to the application scenario.
[0087] Optionally, respectively calculating the distance between the mean value of each sample amplitude of each sub-sample vibration signal and the X embedded set representatives to obtain N sample feature vectors specifically includes the following steps:
[0088] A21. In each sub-sample vibration signal, the absolute value of the difference between the mean amplitude of each sample and the X embedded set representatives is calculated to obtain X first absolute values of the difference.
[0089] The calculation process of the first difference absolute value of the first sub-sample vibration signal is expressed as follows:
[0090] Where m is any one of 1, 2, ···M; x=1, 2···X, v 嵌入x represents the xth embedding set representative.
[0091] The rest is similar and I will not elaborate on it.
[0092] A22. Taking the minimum value among the X first difference absolute values as the first local eigenvalue of the corresponding sample amplitude mean, respectively, to obtain M first local eigenvalues, and taking the combination of the M first local eigenvalues as the sample feature vector of the corresponding sub-sample vibration signal, to obtain N sample feature vectors.
[0093] That is The sample feature vector of the first sub-sample vibration signal is expressed as:
[0094] D1={d 11 , d 12 ,···,d 1M}
[0095] Similarly, repeat the above steps to calculate a total of N sample feature vectors expressed as:
[0096] D={D1,D2,···D N}.
[0097] S104, when the transformer is in a faulty operating state, obtaining a measured vibration signal and a measured electrical signal of the transformer.
[0098] The measured vibration signal refers to the vibration data of the transformer under fault operation, and the measured electrical signal refers to the electrical data such as current, voltage, resistance and power of the transformer under fault operation. The measured vibration signal and the measured electrical signal are sampled simultaneously and at the same sampling frequency, and the sampling duration is the same as the sampling duration of the sub-sample vibration signal and the sub-sample electrical signal.
[0099] Optionally, if the sampling duration of each sub-sample vibration signal and sub-sample electrical signal is 5 minutes and the sampling frequency is 10 sampling points per minute, then when the transformer is in a fault operation state, we also perform data sampling with a sampling duration of 5 minutes and a sampling frequency of 10 sampling points per minute to obtain the measured vibration signal and the measured electrical signal.
[0100] Exemplarily, the measured vibration signal is expressed as:
[0101] V 实测 =(v 实测1 , v 实测2 ,···,v 测K )
[0102] If the measured electrical signal includes a measured current signal and a measured active power signal, the measured current signal is expressed as:
[0103] I 实测 =(i实测1 ,i 实测2 ,···,i 实测K )
[0104] The measured active power signal is expressed as:
[0105] P 实测 =(p 实测1 , p 实测2 ,···,p 实测K ).
[0106] S105, respectively calculating the distances between the measured electrical signal and the N segments of sub-sample electrical signals to obtain N candidate reference distances, taking a segment corresponding to the minimum value among the N candidate reference distances as a target reference segment, and obtaining a sample feature vector of the target reference segment from the N sample feature vectors.
[0107] Optionally, a suitable distance measurement method may be selected to calculate the distance between the measured electrical signal and the N-segment sub-sample electrical signals, such as Euclidean distance, Manhattan distance or cosine similarity, and the most suitable method may be selected according to the application scenario.
[0108] In a specific embodiment, the sub-sample electrical signal includes a sample current signal and a sample active power signal, the measured electrical signal includes a measured current signal and a measured active power signal, and the distances between the measured electrical signal and the N segments of sub-sample electrical signals are calculated to obtain N candidate reference distances, which specifically includes the following steps:
[0109] B1. Calculate the normalized vector distances between the measured current signal and the N-segment sample current signals respectively to obtain N first reference distances.
[0110] The calculation process is expressed as:
[0111] Where n = 1, 2, ..., N
[0112] B2. Calculate the normalized vector distances between the measured active power signal and the N segments of sample active power signals respectively to obtain N second reference distances.
[0113] The calculation process is expressed as:
[0114] Where n = 1, 2, ..., N
[0115] B3. Perform weighted calculation on the N first reference distances and the N second reference distances to obtain N candidate reference distances.
[0116] The calculation process is expressed as:
[0117] ad 实测In +bd 实测Pn, where a and b are weighting parameters.
[0118] S106, calculating the distance between the measured vibration signal and the X embedded set representatives to obtain the corresponding measured feature vector.
[0119] Optionally, a suitable distance measurement method may be selected to calculate the distance between the measured vibration signal and the X embedded set representatives, such as Euclidean distance, Manhattan distance or cosine similarity, and the most suitable method may be selected according to the application scenario.
[0120] In a specific embodiment, calculating the distance between the measured vibration signal and the X embedded set representatives to obtain the corresponding measured feature vector specifically includes the following steps:
[0121] C1. Taking the measured vibration signal as a unit, the average value of the sampling point amplitude is calculated to obtain the M measured amplitude mean values.
[0122] Optionally, let M = 5, we can calculate the measured vibration signal as follows:
[0123] {v 实测ave1 , v 实测ave2 ,···,v 实测aveM}={(v 实测1 +v 实测2 +···+v 实测5 ) / 5,(v 实测6 +v 实测7 +···+v 实测10 ) / 5,(v 实测(K-4) +v 实测(K-3) +···+v 实测K ) / 5}.
[0124] C2. Calculate the distance between the mean of the M measured amplitudes and the X embedded set representatives to obtain the corresponding measured feature vector.
[0125] Here, a suitable distance measurement method can be selected to calculate the distance between the mean values of the M measured amplitudes and the X embedded set representatives, such as Euclidean distance, Manhattan distance or cosine similarity. The most suitable method can be selected according to the application scenario.
[0126] Optionally, calculating the distance between the mean values of the M measured amplitudes and the X embedded set representatives to obtain the corresponding measured feature vectors specifically includes the following steps:
[0127] C21. Calculate the absolute value of the difference between each measured amplitude mean and the X embedded set representatives to obtain X second difference absolute values.
[0128] The calculation process of the absolute value of the second difference is expressed as:
[0129] d实测mx =|v 实测avem -v 嵌入x |, where m is any one of 1, 2, ···M; x=1, 2···X, v 嵌入x represents the xth embedding set representative.
[0130] C22. Taking the minimum value among the X second difference absolute values as the second local eigenvalue corresponding to the measured amplitude mean value to obtain M second local eigenvalues, and taking the combination of the M second local eigenvalues as the measured eigenvector.
[0131] That is, d 实测m =mind 实测mx , the measured eigenvector is expressed as:
[0132] D measured = {d 实测1 , d 实测2 ,···,d 实测M}.
[0133] S107, when the difference between the measured feature vector and the sample feature vector of the target reference segment meets a preset difference condition, the measured vibration signal is determined to be problematic data.
[0134] The preset standard is used to determine whether the difference between the measured feature vector and the sample feature vector of the target reference segment is large enough.
[0135] Optionally, a suitable difference measurement method (such as Euclidean distance, cosine similarity, etc.) is first selected, and then the difference value between the measured feature vector and the sample feature vector of the target reference segment is calculated based on the difference measurement method. If the difference value exceeds a preset threshold, the measured vibration signal corresponding to the measured feature vector is determined to be problematic data.
[0136] In a specific embodiment, when the difference between the measured feature vector and the sample feature vector of the target reference segment meets a preset difference condition, determining that the measured vibration signal is problematic data specifically includes the following steps: when the difference between any two elements of the sample feature vector of the target reference segment and the measured feature vector is greater than a preset first threshold, or when the normalized distance between the sample feature vector of the target reference segment and the measured feature vector is greater than a preset second threshold, determining that the measured vibration signal is problematic data.
[0137] Optionally, the first threshold is set to 0.3 and the second threshold is set to 0.15.
[0138] S108, judging the degree of change of the winding based on the proportion of problem data within a preset time period.
[0139] The pre-set time period is used to count the proportion of problem data, usually expressed in hours, days or longer time units. Optionally, multiple levels (such as slight, medium, and severe) can be set according to the proportion of problem data to determine the degree of change of the winding.
[0140] In a specific embodiment, the degree of change of the winding includes a first degree of change, a second degree of change, and a third degree of change that increase in sequence. Determining the degree of change of the winding based on the proportion of problem data within a preset time period specifically includes the following steps:
[0141] D1. Within p preset unit time lengths, if the proportion of the problem data is less than a preset first proportion threshold, it is determined that the degree of change of the winding is the first degree of change.
[0142] Optionally, the preset unit time length is 1 day, p=7, and the first proportion threshold is set to 10%. At this time, it is determined that the change degree of the winding is the first change degree, that is, it is considered that the winding state of the main transformer has not changed, and the critical monitoring can be terminated.
[0143] D2. Within q preset unit time lengths, if the proportion of the problem data is greater than or equal to the preset first proportion threshold and less than the preset second proportion threshold, it is determined that the degree of change of the winding is the second degree of change.
[0144] Optionally, let q=1, and the second proportion threshold is set to 30%. At this time, it is determined that the change degree of the winding is the second change degree, that is, it is considered that the winding state of the main transformer may have a slight change, and it is necessary to continue to carry out intensive monitoring.
[0145] D3. Within q preset unit time lengths, if the proportion of the problem data is greater than the preset second proportion threshold, it is determined that the degree of change of the winding is the third degree of change.
[0146] At this time, the degree of change of the winding is determined to be the third degree of change, that is, it is believed that the winding state of the main transformer may have undergone a significant change, and a power outage inspection is required.
[0147] The above-mentioned method for monitoring the degree of winding change first collects sample vibration signals and sample electrical signals simultaneously when the transformer is operating normally, and divides them into N sub-sample signals. Then, set X embedded collection representatives, calculate the distance between each sub-sample signal and these representatives, and obtain N sample feature vectors. When the transformer fails, collect the measured vibration signal and electrical signal, calculate the distance between the measured signal and the N sub-sample signals, determine the target reference segment corresponding to the minimum distance, and extract its sample feature vector. By comparing the difference between the measured feature vector and the sample feature vector of the target reference segment, if the difference meets the preset conditions, it is determined to be problem data. The degree of change of the winding is judged based on the proportion of problem data within a preset time. This method can realize real-time monitoring and accurate evaluation of winding changes by comparing the sample features under normal conditions with the measured features under fault conditions, timely identify problem data, and analyze the health status of the winding according to the proportion of problem data, so as to ensure the reliable operation of the transformer and reduce the risk of failure.
[0148] In one embodiment, Figure 2 As shown, a winding variation degree monitoring device is proposed, the device comprising:
[0149] The normal operating state monitoring module 201 is used to obtain a sample vibration signal and a sample electrical signal of the transformer when the transformer is in a normal operating state; wherein the sample vibration signal and the sample electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same; the sample vibration signal is equally spaced into N segments of sub-sample vibration signals, and the sample electrical signal is equally spaced into N segments of sub-sample electrical signals; wherein N ≥ 2; X uniformly distributed embedded collection representatives are set, and the distance between each segment of the sub-sample vibration signal and the X embedded collection representatives is calculated respectively to obtain N sample feature vectors;
[0150] The fault operation state monitoring module 202 is used to obtain the measured vibration signal and the measured electrical signal of the transformer when the transformer is in a fault operation state; wherein the measured vibration signal and the measured electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same as the sampling duration of the sub-sample vibration signal and the sub-sample electrical signal; respectively calculate the distance between the measured electrical signal and the N-segment sub-sample electrical signal to obtain N candidate reference distances, take a segment corresponding to the minimum value of the N candidate reference distances as the target reference segment, and obtain the sample feature vector of the target reference segment from the N sample feature vectors; calculate the distance between the measured vibration signal and the X embedded set representatives to obtain the corresponding measured feature vector;
[0151] The winding change degree calculation module 203 is used to determine that the measured vibration signal is problem data when the difference between the measured feature vector and the sample feature vector of the target reference segment meets the preset difference condition; and to determine the degree of change of the winding based on the proportion of problem data within a preset time length.
[0152] Figure 3 FIG. 2 shows an internal structure diagram of a winding variation degree monitoring device in one embodiment. Figure 3 As shown, the winding variation degree monitoring device includes a processor, a memory and a network interface connected via a system bus. Among them, the memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the winding variation degree monitoring device stores an operating system and may also store a computer program. When the computer program is executed by the processor, the processor can implement the winding variation degree monitoring method. The internal memory may also store a computer program. When the computer program is executed by the processor, the processor can execute the winding variation degree monitoring method. Those skilled in the art can understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present application, and does not constitute a limitation on the winding variation degree monitoring device to which the scheme of the present application is applied. The specific winding variation degree monitoring device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.
[0153] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented: when a transformer is in a normal operating state, a sample vibration signal and a sample electrical signal of the transformer are obtained; wherein the sample vibration signal and the sample electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same; the sample vibration signal is equally spaced into N segments of sub-sample vibration signals, and the sample electrical signal is equally spaced into N segments of sub-sample electrical signals; wherein N ≥ 2; X uniformly distributed embedded collection representatives are set, and the distance between each segment of the sub-sample vibration signal and the X embedded collection representatives is calculated respectively to obtain N sample feature vectors; when the transformer is in a faulty operating state, the measured vibration signal and the measured feature vector of the transformer are obtained. electrical signal; wherein the measured vibration signal and the measured electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same as the sampling duration of the sub-sample vibration signal and the sub-sample electrical signal; the distances between the measured electrical signal and the N segments of sub-sample electrical signals are calculated respectively to obtain N candidate reference distances, a segment corresponding to the minimum value among the N candidate reference distances is used as the target reference segment, and the sample feature vector of the target reference segment is obtained from the N sample feature vectors; the distance between the measured vibration signal and the X embedded collection representatives is calculated to obtain the corresponding measured feature vector; when the difference between the measured feature vector and the sample feature vector of the target reference segment meets the preset difference condition, the measured vibration signal is determined to be problem data; and the degree of change of the winding is determined based on the proportion of the problem data within the preset duration.
[0154] A winding variation degree monitoring device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: when the transformer is in a normal operating state, a sample vibration signal and a sample electrical signal of the transformer are obtained; wherein the sample vibration signal and the sample electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same; the sample vibration signal is equally spaced into N segments of sub-sample vibration signals, and the sample electrical signal is equally spaced into N segments of sub-sample electrical signals; wherein N≥2; X uniformly distributed embedded collection representatives are set, and the distance between each segment of the sub-sample vibration signal and the X embedded collection representatives is calculated respectively to obtain N sample feature vectors; when the transformer is in a faulty operating state, the transformer is obtained The measured vibration signal and the measured electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same as the sampling duration of the sub-sample vibration signal and the sub-sample electrical signal; the distances between the measured electrical signal and the N-segment sub-sample electrical signal are calculated respectively to obtain N candidate reference distances, and a segment corresponding to the minimum value of the N candidate reference distances is used as the target reference segment, and the sample feature vector of the target reference segment is obtained from the N sample feature vectors; the distance between the measured vibration signal and the X embedded set representatives is calculated to obtain the corresponding measured feature vector; when the difference between the measured feature vector and the sample feature vector of the target reference segment meets the preset difference condition, the measured vibration signal is determined to be problem data; and the degree of change of the winding is determined based on the proportion of the problem data within the preset duration.
[0155] It should be noted that the above-mentioned winding variation degree monitoring method, device, equipment and computer-readable storage medium belong to a general inventive concept, and the contents in the embodiments of the winding variation degree monitoring method, device, equipment and computer-readable storage medium are applicable to each other.
[0156] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0157] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0158] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A method for monitoring winding variation, characterized in that: The method comprises: When the transformer is in a normal operating state, a sample vibration signal and a sample electrical signal of the transformer are obtained; wherein the sample vibration signal and the sample electrical signal are sampled simultaneously, and the sampling duration and sampling frequency are the same; Dividing the sample vibration signal into N segments of sub-sample vibration signals at equal intervals, and dividing the sample electrical signal into N segments of sub-sample electrical signals at equal intervals; wherein N≥2; Setting X uniformly distributed embedding collection representatives, respectively calculating the distance between each sub-sample vibration signal and the X embedding collection representatives, to obtain N sample feature vectors; When the transformer is in a faulty operation state, a measured vibration signal and a measured electrical signal of the transformer are obtained; wherein the measured vibration signal and the measured electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same as the sampling duration of the sub-sample vibration signal and the sub-sample electrical signal; Calculating the distances between the measured electrical signal and the N segments of sub-sample electrical signals respectively to obtain N candidate reference distances, taking a segment corresponding to a minimum value among the N candidate reference distances as a target reference segment, and obtaining a sample feature vector of the target reference segment from the N sample feature vectors; Calculating the distance between the measured vibration signal and the X embedded set representatives to obtain a corresponding measured feature vector; When the difference between the measured feature vector and the sample feature vector of the target reference segment meets a preset difference condition, determining that the measured vibration signal is problematic data; The degree of change of the winding is determined based on the proportion of the problem data within a preset time period.
2. The method according to claim 1, characterized in that The set of X uniformly distributed embedding collection representatives includes: The maximum amplitude and the minimum amplitude of the sample vibration signal are obtained, and X-2 vibration amplitudes are uniformly inserted between the maximum amplitude and the minimum amplitude to obtain X embedded set representatives.
3. The method according to claim 1, characterized in that The distance between each sub-sample vibration signal and the X embedded set representatives is calculated to obtain N sample feature vectors, including: In each sub-sample vibration signal, the M sampling points are sequentially taken as a unit to calculate the average value of the sampling point amplitude, so as to obtain the M sample amplitude mean values; wherein M ≥ 2; Calculate the distance between the mean value of each sample amplitude of each sub-sample vibration signal and the X embedded set representatives to obtain N sample feature vectors; The calculating the distance between the measured vibration signal and the X embedded set representatives to obtain a corresponding measured feature vector includes: The measured vibration signal is calculated with M sampling points as a unit to average the amplitude of the sampling points, so as to obtain M measured amplitude averages; The distances between the M measured amplitude means and the X embedded set representatives are calculated to obtain corresponding measured feature vectors.
4. The method according to claim 3, characterized in that The respectively calculating the distance between the mean value of each sample amplitude of each sub-sample vibration signal and the X embedded set representatives to obtain N sample feature vectors includes: In each sub-sample vibration signal, respectively calculating the absolute value of the difference between the mean amplitude of each sample and the X embedded set representatives to obtain X first absolute values of the difference; Taking the minimum values of the X first difference absolute values as the first local eigenvalues of the corresponding sample amplitude mean, respectively, to obtain M first local eigenvalues, and taking the combination of the M first local eigenvalues as the sample eigenvector of the corresponding sub-sample vibration signal, to obtain N sample eigenvectors; The calculating the distance between the M measured amplitude means and the X embedded set representatives to obtain the corresponding measured feature vector includes: Calculate the absolute value of the difference between each measured amplitude mean and the X embedded set representatives to obtain X second difference absolute values; The minimum values among the X second difference absolute values are respectively used as the second local eigenvalues corresponding to the measured amplitude mean value to obtain M second local eigenvalues, and the combination of the M second local eigenvalues is used as the measured eigenvector.
5. The method according to claim 1, characterized in that The sub-sample electrical signal includes a sample current signal and a sample active power signal, the measured electrical signal includes a measured current signal and a measured active power signal, and the distances between the measured electrical signal and the N segments of sub-sample electrical signals are calculated to obtain N candidate reference distances, including respectively calculating normalized vector distances between the measured current signal and N segments of sample current signals to obtain N first reference distances; respectively calculating the normalized vector distances between the measured active power signal and the N segments of sample active power signals to obtain N second reference distances; The N first reference distances and the N second reference distances are weightedly calculated to obtain N candidate reference distances.
6. The method according to claim 1, characterized in that When the difference between the measured feature vector and the sample feature vector of the target reference segment satisfies a preset difference condition, determining that the measured vibration signal is problematic data includes: When the difference between any two elements of the sample feature vector of the target reference segment and the measured feature vector is greater than a preset first threshold, or when the normalized distance between the sample feature vector of the target reference segment and the measured feature vector is greater than a preset second threshold, the measured vibration signal is determined to be problematic data.
7. The method according to claim 1, characterized in that The degree of change of the winding includes a first degree of change, a second degree of change and a third degree of change that increase in sequence, and the degree of change of the winding is determined based on the proportion of the problem data within a preset time period, including: Within p preset unit time lengths, if the proportion of the problem data is less than a preset first proportion threshold, it is determined that the degree of change of the winding is the first degree of change; Within q preset unit time lengths, if the proportion of the problem data is greater than or equal to a preset first proportion threshold and less than a preset second proportion threshold, it is determined that the degree of change of the winding is the second degree of change; Within q preset unit time lengths, if the proportion of the problem data is greater than a preset second proportion threshold, it is determined that the degree of change of the winding is the third degree of change.
8. A winding variation degree monitoring device, characterized in that: The winding variation degree monitoring device comprises: A normal operating state monitoring module is used to obtain a sample vibration signal and a sample electrical signal of the transformer when the transformer is in a normal operating state; wherein the sample vibration signal and the sample electrical signal are sampled simultaneously, and the sampling duration and sampling frequency are the same; the sample vibration signal is equally spaced into N segments of sub-sample vibration signals, and the sample electrical signal is equally spaced into N segments of sub-sample electrical signals; wherein N ≥ 2; X uniformly distributed embedded collection representatives are set, and the distance between each segment of the sub-sample vibration signal and the X embedded collection representatives is calculated respectively to obtain N sample feature vectors; A fault operation state monitoring module is used to obtain a measured vibration signal and a measured electrical signal of the transformer when the transformer is in a fault operation state; wherein the measured vibration signal and the measured electrical signal are sampled simultaneously, and the sampling duration and the sampling frequency are the same as the sampling duration of the sub-sample vibration signal and the sub-sample electrical signal; respectively calculate the distance between the measured electrical signal and the N segments of sub-sample electrical signals to obtain N candidate reference distances, take a segment corresponding to the minimum value of the N candidate reference distances as a target reference segment, and obtain a sample feature vector of the target reference segment from the N sample feature vectors; calculate the distance between the measured vibration signal and the X embedded collection representatives to obtain the corresponding measured feature vector; The winding change degree calculation module is used to determine that the measured vibration signal is problem data when the difference between the measured feature vector and the sample feature vector of the target reference segment meets the preset difference condition; and to determine the degree of change of the winding based on the proportion of the problem data within a preset time length.
9. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 7.
10. A winding variation degree monitoring device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 7.
Citation Information
Cited By
Transformer winding fault detection method and device, electronic equipment and storage medium
CN120180237A
Transformer winding fault detection method, device, electronic equipment and storage medium
CN120180237B
