Power distribution network equipment aging life prediction method and system and storage medium

Abstract

The invention belongs to the field of power distribution network equipment management, and discloses a power distribution network equipment aging life prediction method and system and a storage medium, and the method comprises the steps: S1, monitoring power distribution network equipment, periodically obtaining various types of monitoring data of the power distribution network equipment, and enabling the collection periods of the various types of monitoring data to be consistent; s2, filtering each type of monitoring data of the power distribution network equipment in a preset time window to obtain multiple types of filtered monitoring data; s3, respectively obtaining data features of each type of filtered monitoring data; and S4, inputting the data features of all types of monitoring data into a deep learning model trained in advance for prediction, and obtaining the residual life of the power distribution network equipment. According to the invention, the accuracy of the filtering result can be effectively improved, so that the accuracy of the residual life prediction result is improved.

Description

Technical Field

[0001] The present invention relates to the field of distribution network equipment management, and in particular to a distribution network equipment aging life prediction method, system and storage medium. Background Art

[0002] Since different distribution network equipment is used in different environments, the aging speed of distribution network equipment is also different, and the actual service life of power equipment may be less than or greater than the designed service life of power equipment. Therefore, by predicting the remaining service life of distribution network equipment such as transformers, the estimated remaining service life of distribution network equipment can be obtained, so that power operation and maintenance personnel can more timely replace power equipment with too short remaining service life and other operation and maintenance operations.

[0003] The process of predicting the remaining life of power equipment generally includes steps such as data preprocessing, data feature extraction, and inputting the extracted features into the trained prediction model for prediction. In the process of data and processing, it is usually necessary to filter the monitoring data of the distribution network. The existing filtering methods generally use only one filtering algorithm to obtain the filtering results of the monitoring data of the distribution network, or use two or more filtering algorithms to calculate the filtering results of the monitoring data of the distribution network respectively, and then perform weighted fusion on the filtering results of all types of filtering algorithms to obtain the final filtering results. However, these filtering algorithms can only filter based on the monitoring data itself. Therefore, if the state of the distribution network equipment suddenly changes, resulting in a large change in the monitoring data, the monitoring data currently being filtered may be considered as noise, and thus be mistakenly over-filtered, resulting in the result after filtering deviating too much from the actual situation, which will cause the features obtained in the subsequent feature extraction process to be inaccurate, affecting the accuracy of the prediction results obtained by predicting the remaining life of the distribution network equipment. Summary of the invention

[0004] The purpose of the present invention is to disclose a method, system and storage medium for predicting the aging life of distribution network equipment to solve the technical problems raised in the background technology.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] In a first aspect, the present invention provides a method for predicting the aging life of distribution network equipment, comprising:

[0007] S1, monitor the distribution network equipment and periodically obtain various types of monitoring data of the distribution network equipment, and the collection period of various types of monitoring data is consistent;

[0008] S2, filtering each type of monitoring data of the distribution network equipment within a preset time window to obtain multiple types of filtered monitoring data;

[0009] S3, respectively obtaining data features of each type of filtered monitoring data;

[0010] S4, input the data features of all types of monitoring data into the pre-trained deep learning model for prediction, and obtain the remaining life of the distribution network equipment;

[0011] Among them, S2 includes:

[0012] S21, for monitoring data of type B, storing the monitoring data of type B obtained within a preset time window into sequence U B ;

[0013] S22, using TU to represent all types of sequences of monitoring data except B, and obtaining sequences of monitoring data of each type in TU obtained within a preset time window;

[0014] S23, calculate U B and U i The correlation value between the monitoring data; U i represents the sequence of monitoring data of type i obtained within a preset time window, i∈TU;

[0015] S24, use relv max Indicates the maximum value of the correlation value obtained, judging relv max Is it greater than the set correlation value threshold? If so, a non-correlation filtering algorithm is used to filter U B The monitoring data in U is filtered. If not, the relevant filtering algorithm is used to filter the monitoring data in U B The monitoring data in is filtered.

[0016] Preferably, the types of monitoring data include output voltage, output current, vibration frequency, temperature, operation noise, partial discharge voltage and partial discharge current.

[0017] Preferably, the preset time window is [t star -Tw,t star ], t star It indicates the time when the remaining life of the distribution network equipment starts to be predicted, and Tw indicates the preset time window length.

[0018] Preferably, the data features include time domain features and frequency domain features;

[0019] Time domain features include mean, variance, peak-to-peak value, and zero-crossing rate;

[0020] Frequency domain features include spectrum center, spectrum peak and spectrum width.

[0021] Preferably, the pre-trained deep learning model includes an LSTM model or a GRU model.

[0022] Preferably, calculate U B and U i The correlation values ​​between the monitoring data include:

[0023] The correlation value is calculated using the following formula:

[0024]

[0025] relv B,i Indicates U B and U i The correlation value between the monitoring data of type B, N represents the total number of monitoring data of type B obtained within the preset time window; value B,n Indicates the nth monitoring data of type B obtained within the preset time window, value i,n It indicates the nth monitoring data of type i obtained within the preset time window.

[0026] Preferably, a non-correlated filtering algorithm is used to filter U B The monitoring data in is filtered, including:

[0027] Adaptive filtering algorithm is used to B Filter the monitoring data in;

[0028] Adaptive filtering algorithms include:

[0029] For U B Monitoring data in B , for d B The filtering formula is:

[0030]

[0031] sd B Indicates d B The corresponding filtered monitoring data, M represents the total number of monitoring data in the sequence seq, d j represents the jth monitoring data in seq, η1 represents the first control value; η2 represents the second control value;

[0032] The process of obtaining seq is as follows:

[0033] Will U B In the example, the time of acquisition is d B Before, and the acquisition time is the same as d BThe sequence formed by sorting the first M monitoring data with the smallest absolute value of the difference in acquisition time in order from early to late in acquisition time is taken as seq, and M is the range control value.

[0034] Preferably, a correlation filtering algorithm is used to filter U B Filter the data in, including:

[0035] For U B Monitoring data in B , will get the time at d B Before and get the time and d B The monitoring data with the smallest absolute value of the acquisition time difference is expressed as zd B ; Bthr is used to represent the upper limit of the value range of monitoring data of type B;

[0036] like Adaptive filtering algorithm is used to filter d B Filtering is performed, thrs is the preset ratio;

[0037] like C represents the type of monitoring data corresponding to the maximum value of the correlation value; U represents C represents a sequence of monitoring data of type C obtained within a preset time window;

[0038] Calculate the judgment value:

[0039]

[0040] d C and C Indicates U C Get time and d B and C For the same monitoring data, jdval is the judgment value;

[0041] If jdval is less than or equal to the preset judgment value threshold, the adaptive filtering algorithm is used to filter d B Perform filtering;

[0042] If jdval is greater than the preset judgment value threshold, the following method is used to B To filter:

[0043]

[0044] β represents the filter weight, mid B Represents the average value of the monitoring data in seq.

[0045] In a second aspect, the present invention provides a distribution network equipment aging life prediction system, comprising:

[0046] Acquisition module, filtering module, feature extraction module and prediction module;

[0047] The acquisition module is used to monitor the distribution network equipment and periodically obtain various types of monitoring data of the distribution network equipment. The acquisition period of various types of monitoring data is consistent;

[0048] The filtering module is used to filter each type of monitoring data of the distribution network equipment within a preset time window to obtain multiple types of filtered monitoring data;

[0049] The feature extraction module is used to obtain the data features of each type of filtered monitoring data respectively;

[0050] The prediction module is used to input the data features of all types of monitoring data into the pre-trained deep learning model for prediction and obtain the remaining life of the distribution network equipment;

[0051] Wherein, each type of monitoring data of the distribution network equipment within the preset time window is filtered respectively to obtain multiple types of filtered monitoring data, including:

[0052] For monitoring data of type B, the monitoring data of type B obtained within the preset time window is stored in sequence U B ;

[0053] TU is used to represent all types of sequences of monitoring data except B, and sequences of monitoring data of each type in TU obtained within a preset time window are obtained respectively;

[0054] Calculate U B and U i The correlation value between the monitoring data; U i represents the sequence of monitoring data of type i obtained within a preset time window, i∈TU;

[0055] Use relv max Indicates the maximum value of the correlation value obtained, judging relv max Is it greater than the set correlation value threshold? If so, a non-correlation filtering algorithm is used to filter U B The monitoring data in U is filtered. If not, the relevant filtering algorithm is used to filter the monitoring data in U B The monitoring data in is filtered.

[0056] In a third aspect, the present invention provides a computer storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the above-mentioned distribution network equipment aging life prediction method is implemented.

[0057] Beneficial effects:

[0058] The present invention improves the existing method for predicting the aging life of distribution network equipment. When the correlation value meets the requirements, the monitoring data can be filtered based on the correlation value when the filtered monitoring data undergoes a large change. In this way, the probability of the monitoring data currently being filtered being considered as noise and being erroneously over-filtered when the state of the distribution network equipment undergoes a sudden change, can be effectively reduced. This can effectively improve the accuracy of the filtering results, thereby improving the accuracy of the remaining life prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0060] Figure 1 A schematic diagram of the method for predicting the aging life of distribution network equipment according to the present invention.

[0061] Figure 2 A schematic diagram of the aging life prediction system for distribution network equipment of the present invention. DETAILED DESCRIPTION

[0062] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0063] like Figure 1 In one embodiment shown, the present invention provides a method for predicting the aging life of distribution network equipment, comprising:

[0064] S1, monitor the distribution network equipment and periodically obtain various types of monitoring data of the distribution network equipment, and the collection period of various types of monitoring data is consistent;

[0065] S2, filtering each type of monitoring data of the distribution network equipment within a preset time window to obtain multiple types of filtered monitoring data;

[0066] S3, respectively obtaining data features of each type of filtered monitoring data;

[0067] S4, input the data features of all types of monitoring data into the pre-trained deep learning model for prediction, and obtain the remaining life of the distribution network equipment;

[0068] Among them, S2 includes:

[0069] S21, for monitoring data of type B, storing the monitoring data of type B obtained within a preset time window into sequence U B ;

[0070] S22, using TU to represent all types of sequences of monitoring data except B, and obtaining sequences of monitoring data of each type in TU obtained within a preset time window;

[0071] S23, calculate U B and U i The correlation value between the monitoring data; U i represents the sequence of monitoring data of type i obtained within a preset time window, i∈TU;

[0072] S24, use relv max Indicates the maximum value of the correlation value obtained, judging relv max Is it greater than the set correlation value threshold? If so, a non-correlation filtering algorithm is used to filter U B The monitoring data in U is filtered. If not, the relevant filtering algorithm is used to filter the monitoring data in U B The monitoring data in is filtered.

[0073] When the correlation value meets the requirements, the above scheme can filter the monitoring data based on the correlation value when the filtered monitoring data undergoes a large change. In this way, it can effectively reduce the probability of the monitoring data currently being filtered being considered as noise when the state of the distribution network equipment changes suddenly, resulting in a large change in the monitoring data, thereby effectively improving the accuracy of the filtering results and thus improving the accuracy of the remaining life prediction results.

[0074] Furthermore, the correlation value threshold may be set to 0.9 and the collection period may be 1 hour.

[0075] Preferably, the types of monitoring data include output voltage, output current, vibration frequency, temperature, operation noise, partial discharge voltage and partial discharge current.

[0076] Furthermore, the output voltage and the output current can be obtained by setting a sensor on the power output line of the distribution network equipment. The vibration frequency can be obtained by a vibration sensor set on the surface of the distribution network equipment. The temperature can be the temperature of the surface of the distribution network equipment. The operating noise can be obtained by a sound monitoring sensor set near the distribution network equipment. The partial discharge voltage and the partial discharge current can be obtained by setting corresponding sensors on the surface of the distribution network equipment.

[0077] Preferably, the preset time window is [t star -Tw,t star ], t star It indicates the time when the remaining life of the distribution network equipment starts to be predicted, and Tw indicates the preset time window length.

[0078] Furthermore, the preset time window length may be one week.

[0079] Preferably, the data features include time domain features and frequency domain features;

[0080] Time domain features include mean, variance, peak-to-peak value, and zero-crossing rate;

[0081] Frequency domain features include spectrum center, spectrum peak and spectrum width.

[0082] The peak-to-peak value is the difference between the maximum and minimum values. The zero-crossing rate is the ratio of the number of times the monitoring data crosses the zero axis to the total number of monitoring times. The spectrum center is the weighted average frequency of the spectrum energy and can be used to describe the overall frequency range of the monitoring data. The spectrum peak refers to the peak frequency component in the spectrum. The spectrum width refers to the width of the peak in the spectrum.

[0083] Preferably, the pre-trained deep learning model includes an LSTM model or a GRU model.

[0084] Specifically, the training process of the deep learning model is as follows:

[0085] The first step is to obtain monitoring data that affects the life of distribution network equipment;

[0086] The second step is to pre-process the monitoring data;

[0087] The third step is to extract the data features of the preprocessed monitoring data;

[0088] The fourth step is to divide the preprocessed monitoring data into training set, validation set and test set;

[0089] In the fifth step, the training set is used to tune the key parameters of the model (such as learning rate, hidden layer size, etc.). Grid search, random search or Bayesian optimization can be used.

[0090] The sixth step is to verify the model effect through the validation set to avoid overfitting.

[0091] In the seventh step, the validation set is used to evaluate the prediction effect of the model.

[0092] Preferably, calculate U B and U i The correlation values ​​between the monitoring data include:

[0093] The correlation value is calculated using the following formula:

[0094]

[0095] relv B,i Indicates U B and U i The correlation value between the monitoring data of type B, N represents the total number of monitoring data of type B obtained within the preset time window; value B,n Indicates the nth monitoring data of type B obtained within the preset time window, value i,n It indicates the nth monitoring data of type i obtained within the preset time window.

[0096] Specifically, in the present invention, the upper limit of the correlation value is 1, so when U B and U i The higher the correlation between the monitoring data, the larger the correlation value, so that in the subsequent filtering process, another type of data can be introduced to select the filtering algorithm to improve the accuracy of the filtering effect.

[0097] Preferably, a non-correlated filtering algorithm is used to filter U B The monitoring data in is filtered, including:

[0098] Adaptive filtering algorithm is used to B Filter the monitoring data in;

[0099] Adaptive filtering algorithms include:

[0100] For U B Monitoring data in B , for d B The filtering formula is:

[0101]

[0102] sd B Indicates d B The corresponding filtered monitoring data, M represents the total number of monitoring data in the sequence seq, d j represents the jth monitoring data in seq, η1 represents the first control value; η2 represents the second control value;

[0103] The process of obtaining seq is as follows:

[0104] Will U B In the example, the time of acquisition is d B Before, and the acquisition time is the same as d B The sequence formed by sorting the first M monitoring data with the smallest absolute value of the difference in acquisition time in order from early to late in acquisition time is taken as seq, and M is the range control value.

[0105] When filtering the monitoring data, the present invention adopts an adaptive filtering algorithm. B and d j The difference in value and position in seq determines d j The corresponding weight, when d B and d j The larger the difference in values, the j The later the position in seq, the greater the j The larger the weight of , the greater the impact on the filtering result. Therefore, we can keep the original numerical change trend and adjust d B Effective filtering is helpful to retain more effective information for the subsequent data feature extraction process.

[0106] Furthermore, the first control value is Bthr indicates the upper limit of the value range of monitoring data of type B.

[0107] For example, for a distribution network device of a certain model, it has a corresponding output current range, and the range generally includes an upper limit value and a lower limit value.

[0108] Furthermore, the second control value is

[0109] Furthermore, for U B Monitoring data B , if the acquisition time is d B The number of other monitoring data before is less than M, and the mean filtering method is used to filter d B Perform filtering.

[0110] During filtering, the time d is obtained by calculating B The average value of other monitoring data before is used as d B The corresponding filtered monitoring data.

[0111] Since the adaptive filtering algorithm of the present invention needs to refer to more previous monitoring data, when the amount of previous monitoring data is insufficient, mean filtering is used for filtering, and mean filtering does not need to specifically set the amount of data involved in filtering. Subsequent data can be filtered based on a small amount of previous data.

[0112] Preferably, the calculation formula of the range control value is:

[0113]

[0114] K represents the number of monitoring data in the sequence keq, keq is U B The acquisition time in d B Before and with the acquisition time d B The sequence consisting of the first K monitoring data with the smallest absolute value of the acquisition time difference, d k represents the kth monitoring data in keq, dm represents the maximum value of the monitoring data in keq, zd B Indicates that the acquisition time is d B Before and get the time and d B The monitoring data with the smallest absolute value of the acquisition time difference, xd B Indicates the acquisition time is in zd B Before and get the time and d B The monitoring data with the smallest absolute value of the difference in acquisition time, leth represents the preset number, and λ represents the range control weight.

[0115] When calculating the range control value, the present invention calculates the fluctuation of the monitoring data in the sequence keq within a larger range and the severity of the change of the monitoring data within a smaller range, so that when the fluctuation of the monitoring data in keq is greater and the severity of the change of the monitoring data is greater, the range control value is greater, and the intensity of filtering is improved to achieve more effective filtering, and when the fluctuation of the monitoring data in keq is smaller and the severity of the change of the monitoring data is smaller, the range control value is smaller, so as to reduce the number of monitoring data involved in filtering and improve the efficiency of filtering. Therefore, a comprehensive consideration is made between the efficiency of filtering and the effect of filtering, and the effectiveness of the filtering result can be guaranteed while the efficiency of filtering is improved as much as possible.

[0116] Furthermore, the value of K is 5.

[0117] Furthermore, the preset number is 20.

[0118] Furthermore, the range control weight is 0.6.

[0119] Preferably, a correlation filtering algorithm is used to filter U B Filter the data in, including:

[0120] For U B Monitoring data in B , will get the time at d B Before and get the time and d B The monitoring data with the smallest absolute value of the acquisition time difference is expressed as zd B ; Bthr is used to represent the upper limit of the value range of monitoring data of type B;

[0121] like Adaptive filtering algorithm is used to filter d B Filtering is performed, thrs is the preset ratio;

[0122] like C represents the type of monitoring data corresponding to the maximum value of the correlation value; U represents C represents a sequence of monitoring data of type C obtained within a preset time window;

[0123] Calculate the judgment value:

[0124]

[0125] d C and C Indicates U C Get time and d B and C For the same monitoring data, jdval is the judgment value;

[0126] If jdval is less than or equal to the preset judgment value threshold, the adaptive filtering algorithm is used to filter d B Perform filtering;

[0127] If jdval is greater than the preset judgment value threshold, the following method is used to B To filter:

[0128]

[0129] β represents the filter weight, mid B Represents the average value of the monitoring data in seq.

[0130] When there is another type of monitoring data that meets the requirements, the present invention does not filter directly based on the other type of monitoring data, because if the change range of the monitoring data is small, such a filtering method has limited impact on the filtering result, but will affect the overall filtering efficiency due to the increase in calculation items. Therefore, the present invention only determines the algorithm used for filtering adaptively by calculating the judgment value when the change range of the detection data is large, which can further ensure the overall filtering efficiency. When the judgment value is less than or equal to the judgment value threshold, the present invention still uses an adaptive filtering algorithm for filtering, because the smaller the judgment value, the closer the fluctuation ratio of the two types of monitoring data (that is, the ratio of the absolute value of the difference between two monitoring data with adjacent acquisition times to the upper limit value of the monitoring data), d B The greater the probability of being noise. When the judgment value is large, the greater the difference in the fluctuation ratio of the two types of monitoring data, d B The greater the probability of not belonging to noise, the greater the probability of not belonging to noise. At this time, the present invention improves the adaptive filtering and introduces the average value of the monitoring data. Since the average value will be affected by the monitoring data with a large variation, the value of the filtered monitoring data can be made larger, thereby effectively reducing d without increasing the parameter type and the filtering formula with a large difference in formula structure. B The probability of over-filtering enables the filtering result to more truly reflect the change trend of the data of type B, and can also reduce the application difficulty of the present invention because no excessive types of parameters are added for filtering.

[0131] Furthermore, the value of thrs is

[0132] Furthermore, the judgment value threshold is

[0133] Furthermore, the filter weight is 0.5.

[0134] like Figure 2 , the present invention provides a distribution network equipment aging life prediction system, including an acquisition module, a filtering module, a feature extraction module and a prediction module;

[0135] The acquisition module is used to monitor the distribution network equipment and periodically obtain various types of monitoring data of the distribution network equipment. The acquisition period of various types of monitoring data is consistent;

[0136] The filtering module is used to filter each type of monitoring data of the distribution network equipment within a preset time window to obtain multiple types of filtered monitoring data;

[0137] The feature extraction module is used to obtain the data features of each type of filtered monitoring data respectively;

[0138] The prediction module is used to input the data features of all types of monitoring data into the pre-trained deep learning model for prediction and obtain the remaining life of the distribution network equipment;

[0139] Wherein, each type of monitoring data of the distribution network equipment within the preset time window is filtered respectively to obtain multiple types of filtered monitoring data, including:

[0140] For monitoring data of type B, the monitoring data of type B obtained within the preset time window is stored in sequence U B ;

[0141] TU is used to represent all types of sequences of monitoring data except B, and sequences of monitoring data of each type in TU obtained within a preset time window are obtained respectively;

[0142] Calculate U B and U i The correlation value between the monitoring data; U i represents the sequence of monitoring data of type i obtained within a preset time window, i∈TU;

[0143] Use relv max Indicates the maximum value of the correlation value obtained, judging relv max Is it greater than the set correlation value threshold? If so, a non-correlation filtering algorithm is used to filter U B The monitoring data in U is filtered. If not, the relevant filtering algorithm is used to filter the monitoring data in U B The monitoring data in is filtered.

[0144] The present invention also provides a computer storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned distribution network equipment aging life prediction method is implemented.

[0145] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for predicting the aging life of distribution network equipment, characterized in that: include: S1, monitor the distribution network equipment and periodically obtain various types of monitoring data of the distribution network equipment, and the collection period of various types of monitoring data is consistent; S2, filtering each type of monitoring data of the distribution network equipment within a preset time window respectively to obtain multiple types of filtered monitoring data; S3, respectively obtaining data features of each type of filtered monitoring data; S4, input the data features of all types of monitoring data into the pre-trained deep learning model for prediction, and obtain the remaining life of the distribution network equipment; Among them, S2 includes: S21, for monitoring data of type B, storing the monitoring data of type B obtained within a preset time window into sequence U B ; S22, using TU to represent all types of sequences of monitoring data except B, and obtaining sequences of monitoring data of each type in TU obtained within a preset time window; S23, calculate U B and U i The correlation value between the monitoring data; U i represents the sequence of monitoring data of type i obtained within a preset time window, i∈TU; S24, use relv max Indicates the maximum value of the correlation value obtained, judging relv max Is it greater than the set correlation value threshold? If so, a non-correlation filtering algorithm is used to filter U B The monitoring data in U is filtered. If not, the relevant filtering algorithm is used to filter the monitoring data in U B The monitoring data in is filtered.

2. The method for predicting the aging life of distribution network equipment according to claim 1, characterized in that: The types of monitoring data include output voltage, output current, vibration frequency, temperature, operating noise, partial discharge voltage and partial discharge current.

3. The method for predicting the aging life of distribution network equipment according to claim 1, characterized in that: The preset time window is [t star -Tw,t star ], t star It indicates the time when the remaining life of the distribution network equipment starts to be predicted, and Tw indicates the preset time window length.

4. The method for predicting the aging life of distribution network equipment according to claim 1, characterized in that: Data features include time domain features and frequency domain features; Time domain features include mean, variance, peak-to-peak value, and zero-crossing rate; Frequency domain features include spectrum center, spectrum peak and spectrum width.

5. The method for predicting the aging life of distribution network equipment according to claim 1, characterized in that: Pre-trained deep learning models include LSTM models or GRU models.

6. The method for predicting the aging life of distribution network equipment according to claim 1, characterized in that: Calculate U B and U i The correlation values ​​between the monitoring data include: The correlation value is calculated using the following formula: relv B,i Indicates U B and U i The correlation value between the monitoring data of type B, N represents the total number of monitoring data of type B obtained within the preset time window; value B,n Indicates the nth monitoring data of type B obtained within the preset time window, value i,n It indicates the nth monitoring data of type i obtained within the preset time window.

7. The method for predicting the aging life of distribution network equipment according to claim 1, characterized in that: The non-correlation filtering algorithm is used to B The monitoring data in is filtered, including: Adaptive filtering algorithm is used to B Filter the monitoring data in; Adaptive filtering algorithms include: For U B Monitoring data in B , for d B The filtering formula is: sd B Indicates d B The corresponding filtered monitoring data, M represents the total number of monitoring data in the sequence seq, d j represents the jth monitoring data in seq, η1 represents the first control value; η2 represents the second control value; The process of obtaining seq is as follows: Will U B In the example, the time of acquisition is d B Before, and the acquisition time is the same as d B The sequence formed by sorting the first M monitoring data with the smallest absolute value of the difference in acquisition time in order from early to late in acquisition time is taken as seq, and M is the range control value.

8. The method for predicting the aging life of distribution network equipment according to claim 7, characterized in that: The correlation filtering algorithm is used to B Filter the data in, including: For U B Monitoring data in B , will get the time at d B Before and get the time and d B The monitoring data with the smallest absolute value of the acquisition time difference is expressed as zd B ; Bthr is used to represent the upper limit of the value range of monitoring data of type B; like Adaptive filtering algorithm is used to filter d B Filtering is performed, thrs is the preset ratio; like C represents the type of monitoring data corresponding to the maximum value of the correlation value; U represents C represents a sequence of monitoring data of type C obtained within a preset time window; Calculate the judgment value: d C and C Indicates U C Get time and d B and C For the same monitoring data, jdval is the judgment value; If jdval is less than or equal to the preset judgment value threshold, the adaptive filtering algorithm is used to filter d B Perform filtering; If jdval is greater than the preset judgment value threshold, the following method is used to B To filter: β represents the filter weight, mid B Represents the average value of the monitoring data in seq.

9. Distribution network equipment aging life prediction system, characterized by: It includes acquisition module, filtering module, feature extraction module and prediction module; The acquisition module is used to monitor the distribution network equipment and periodically obtain various types of monitoring data of the distribution network equipment. The acquisition period of various types of monitoring data is consistent; The filtering module is used to filter each type of monitoring data of the distribution network equipment within a preset time window to obtain multiple types of filtered monitoring data; The feature extraction module is used to obtain the data features of each type of filtered monitoring data respectively; The prediction module is used to input the data features of all types of monitoring data into the pre-trained deep learning model for prediction and obtain the remaining life of the distribution network equipment; Wherein, each type of monitoring data of the distribution network equipment within the preset time window is filtered respectively to obtain multiple types of filtered monitoring data, including: For monitoring data of type B, the monitoring data of type B obtained within the preset time window is stored in sequence U B ; TU is used to represent all types of sequences of monitoring data except B, and sequences of monitoring data of each type in TU obtained within a preset time window are obtained respectively; Calculate U B and U i The correlation value between the monitoring data; U i represents the sequence of monitoring data of type i obtained within a preset time window, i∈TU; Use relv max Indicates the maximum value of the correlation value obtained, judging relv max Is it greater than the set correlation value threshold? If so, a non-correlation filtering algorithm is used to filter U B The monitoring data in U is filtered. If not, the relevant filtering algorithm is used to filter the monitoring data in U B The monitoring data in is filtered.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the aging life of distribution network equipment according to any one of claims 1 to 8 is implemented.

Images