A method and system for analyzing and processing data of a power distribution system

Abstract

The present application relates to the field of data analysis, and in particular to a method and system for data analysis and processing of a power distribution system, the method comprising the steps of: constructing a time series segment, the time series segment contains multiple samples, the samples contain multiple features representing monitoring parameters; obtaining multiple historical time series segments, labeling the historical time series segments, the labels are fault types; calculating the matching degree of the historical time series segments, taking the historical time series segment with the greatest matching degree as the optimal matching time series segment, and taking the fault type corresponding to the optimal matching time series segment as the fault type of the real-time time series segment. The present application has the effect of accurately extracting the differences in the characteristic performance of different faults.

Description

Technical Field

[0001] The present application relates to the field of data analysis, and in particular to a method and system for analyzing and processing data of a power distribution system. Background Art

[0002] With the popularization of smart meters, sensors, telemetry instruments and automated monitoring equipment, the power distribution system can realize real-time monitoring of electrical equipment. Through these devices, the power distribution system can obtain various operating parameters, such as voltage, current, power, equipment operating status, temperature and other information, and centrally manage and analyze them through the data acquisition system. These data not only cover the current real-time operating status, but also include long-term accumulated historical data, providing rich information for fault diagnosis and prediction.

[0003] The existing technology can realize continuous monitoring of the status of distribution equipment based on the data collected by sensors, and timely detect possible faults, overloads, equipment aging or potential power quality problems based on data analysis models. Through machine learning and deep learning algorithms, the distribution system can extract features and recognize patterns from the collected data to locate the fault point. For example, a method for evaluating the health index of a distribution network disclosed in a Chinese patent application document with publication number CN105512448A includes the following steps: setting a network health index model; evaluating the health index of transformer and distribution equipment; and evaluating the network health index of the distribution network.

[0004] However, in the prior art, most models for identifying faults through distribution system data analysis identify faults based on single features or multi-feature weighted methods in the time domain or frequency domain. They are unable to accurately extract the differences in feature performance of different faults, and cannot identify mixed faults, which has great limitations. Summary of the invention

[0005] In order to solve the technical problem of large limitations in multi-fault identification in a power distribution system, the present application provides a data analysis and processing method and system for a power distribution system.

[0006] In a first aspect, the present application provides a method for analyzing and processing data of a power distribution system, which adopts the following technical solution:

[0007] A method for analyzing and processing data of a power distribution system comprises the following steps: constructing a time series segment, wherein the time series segment comprises a plurality of samples, and the samples comprise a plurality of features representing monitoring parameters; obtaining a plurality of historical time series segments, and labeling the historical time series segments, wherein the labels are fault types; calculating the matching degree of the historical time series segments, and taking the historical time series segment with the greatest matching degree as the optimal matching time series segment, and taking the fault type corresponding to the optimal matching time series segment as the fault type of the real-time time series segment; wherein the method for calculating the matching degree comprises the following steps: performing sequence averaging on historical time series segments of different fault types according to a DBA algorithm to obtain a global feature average sequence, wherein the global feature average sequence comprises a plurality of global characteristics, Calculate the feature difference of each feature relative to the global feature; calculate the importance of the feature according to the feature difference of different fault types; calculate the weight based on the importance, calculate the time domain similarity of the real-time time series segment and the historical time series segment according to the weight and the element values ​​in the real-time time series segment and the historical time series segment, and in response to the time series similarity being greater than a preset threshold, use the historical time series period as the historical time series segment for time domain matching; convert the real-time time series segment and the historical time series segment into frequency domain signals through discrete Fourier transform, calculate the frequency domain similarity between the time domain matching historical time series segment and the real-time time series segment in the frequency domain; take the product of the time domain similarity and the frequency domain similarity of the historical time series segment as the matching degree.

[0008] The beneficial effects are: by using data analysis methods to analyze and process distribution system data, and combining the time-frequency domain characteristics of distribution system data to accurately identify distribution system faults, it can effectively respond to data changes in the distribution system under different operating conditions. Whether it is load fluctuations, equipment aging, or environmental interference, faults can be accurately identified through time-frequency domain feature analysis, allowing the distribution system to maintain stable operation under complex and changeable conditions, improving the robustness and adaptability of the system.

[0009] By calculating the time domain similarity and frequency domain similarity and taking their product as the matching degree, the characteristic differences of the distribution system data in the time domain and frequency domain are comprehensively considered. The characteristic information of mixed faults is often scattered in multiple domains. This method can extract the characteristic information related to mixed faults in different domains through comprehensive analysis in the time and frequency domains, and integrate them to calculate the matching degree. This helps to identify mixed faults and improve the diagnostic ability of complex fault conditions.

[0010] Optionally, the feature difference is calculated as: , where Indicated in Characteristics of the fault type The difference between the global features, Representation characteristics The first in the feature average sequence values, represents the first values, express The rate of change of the ath value in the characteristic average sequence of represents the first The rate of change of the value, Represents the total number of elements in the feature average sequence or the global feature average sequence.

[0011] The beneficial effect is that the formula not only considers the difference between the eigenvalue and the global eigenvalue, but also introduces the difference in the rate of change. This makes the evaluation of the feature difference more refined and can capture the dynamic changes of the eigenvalue in the time series, thereby more accurately identifying different fault types.

[0012] Optionally, the feature difference is calculated as: , where Indicated in Characteristics of the fault type The difference between the global features, Representation characteristics The first in the feature average sequence values, represents the first values, Represents the total number of elements in the feature average sequence or the global feature average sequence.

[0013] The beneficial effect is that this formula simplifies the calculation process of feature differences by directly calculating the absolute difference between the eigenvalue and the global eigenvalue, reducing the calculation complexity. Compared with complex formulas containing the rate of change, this simplified method is more efficient in practical applications, especially when processing large-scale data, and can significantly improve the calculation speed.

[0014] Optionally, the importance is calculated as: , where Representation characteristics In the The importance of each fault type, Indicates the total number of fault types, Representation characteristics In the The difference in the types of faults, Indicates The standard deviation of all feature differences in the fault type.

[0015] The beneficial effects are: the standard deviation is introduced to standardize the feature differences. This makes the feature differences in different fault types comparable, avoiding deviations caused by different dimensions or uneven data distribution. Through standardization, the formula can adapt to different data distribution situations, especially when the data distribution is uneven or there are extreme values, and can more accurately evaluate the importance of the features.

[0016] Optionally, the importance is calculated as: , where Representation characteristics In the The importance of each fault type, Indicates the total number of fault types, Representation characteristics In the The differences in the types of failures.

[0017] The beneficial effect is that by directly calculating the absolute difference between the feature difference and the average difference, the calculation process of the feature importance is simplified and the calculation complexity is reduced.

[0018] Optionally, for any fault type, the calculation formula for time domain similarity is:

[0019] , where for the Types of faults, Real-time timing segment Historical time series The time domain similarity of Representation characteristics The importance of Indicates real-time timing segment or historical time series Medium Features The total number of Represents the characteristics in the real-time time series No. elements, Represents a historical time series Medium Features No. elements, Indicates An exponential function with base .

[0020] Optionally, for any fault type, the calculation formula for time domain similarity is:

[0021] , where for the Types of faults, Real-time timing segment Historical time series The time domain similarity of Representation characteristics The importance of Indicates real-time timing segment or historical time series Medium Features The total number of Represents the characteristics in the real-time time series No. elements, Represents a historical time series Medium Features No. elements, Indicates An exponential function with base .

[0022] Optionally, the frequency domain similarity is calculated as:

[0023] ; In the formula, Real-time timing segment Historical time series The frequency domain similarity of Indicates real-time timing segment or historical time series Medium Features The total number of is the maximum order of the harmonics of the frequency domain signal; Indicates real-time timing segment Medium Features In the frequency domain signal Number of subharmonic occurrences; Represents a historical time series Medium Features In the frequency domain signal Number of subharmonic occurrences; Indicates real-time timing segment Medium Features In the frequency domain signal The amplitude of the subharmonics, Represents a historical time series Medium Features In the frequency domain signal The amplitude of the subharmonics, represents the maximum value function, Indicates An exponential function with base .

[0024] Optionally, the frequency domain similarity is calculated as:

[0025] ; In the formula, Real-time timing segment Historical time series The frequency domain similarity of Indicates real-time timing segment or historical time series Medium Features The total number of is the maximum order of the harmonics of the frequency domain signal; Indicates real-time timing segment Medium Features In the frequency domain signal Number of subharmonic occurrences; Represents a historical time series Medium Features In the frequency domain signal Number of subharmonic occurrences; Indicates real-time timing segment Medium Features In the frequency domain signal The amplitude of the subharmonics, Represents a historical time series Medium Features In the frequency domain signal The amplitude of the subharmonics, represents the maximum value function, represents the minimum function, Indicates An exponential function with base .

[0026] In a second aspect, the present application provides a power distribution system data analysis and processing system, which adopts the following technical solution:

[0027] A distribution system data analysis and processing system comprises: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the distribution system data analysis and processing method described above is implemented.

[0028] The beneficial effect is: the above-mentioned distribution system data analysis and processing method is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a system based on the memory and the processor for easy use.

[0029] This application has the following technical effects:

[0030] By using data analysis methods to analyze and process distribution system data, and combining the time and frequency domain characteristics of distribution system data to accurately identify distribution system faults, the distribution system can maintain stable operation under complex and changeable working conditions, thereby improving the robustness and adaptability of the system.

[0031] The characteristic information of mixed faults is often scattered in multiple domains. Through comprehensive analysis in the time and frequency domains, the characteristic information related to mixed faults can be extracted in different domains and integrated to calculate the matching degree, thereby improving the diagnostic capability of complex fault conditions.

[0032] The importance of features is calculated according to the feature differences of different fault types and incorporated into the calculation of time domain similarity as a weight to highlight the impact of key features on fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present application will become easily understood. In the accompanying drawings, several embodiments of the present application are shown in an exemplary and non-restrictive manner, and the same or corresponding numbers represent the same or corresponding parts.

[0034] Figure 1 It is a method flow chart of a distribution system data analysis and processing method according to an embodiment of the present application.

[0035] Figure 2 It is a method flow chart of step S2 in a method for analyzing and processing power distribution system data in an embodiment of the present application. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0037] It should be understood that when the terms "first", "second", etc. are used in the claims, specification and drawings of the present application, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the specification and claims of the present application indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0038] The present application embodiment discloses a method for analyzing and processing data of a power distribution system, referring to Figure 1 , including steps S1-S2, which are as follows:

[0039] S1: Construct a time series segment, where the time series segment includes multiple samples, and the sample includes multiple features representing monitoring parameters; obtain multiple historical time series segments, and label the historical time series segments, where the label is the fault type.

[0040] Sensors are installed at different parts of the power distribution system to collect real-time data on the power distribution system, such as current, voltage, load, temperature, etc., and the data is preprocessed and labeled. In one embodiment, sensors are used to collect data on current, voltage, load, temperature, humidity, etc. at locations such as transformers, distribution lines, and switchgear in the power distribution system. Wavelet transform is used to denoise the data. Since different sensor data have different dimensions, the data is normalized to obtain a data set for the power distribution system.

[0041] Compared with the transmission network, the fault types of the distribution system are more complex, and there is a certain trend in the occurrence of faults in time. Therefore, a time series segment is constructed, and technicians mark the normal or specific fault type according to the data of each time series segment. It can be expressed as:

[0042] ; In the formula, is the sample size, It is the number of features contained in a sample. Features refer to the above-mentioned current, voltage, load, temperature, humidity and other data. The number of samples can be 10. One time segment corresponds to one fault type, and each fault type contains multiple time segments.

[0043] S2: Calculate the matching degree of the historical time series segment, take the historical time series segment with the greatest matching degree as the optimal matching time series segment, and take the fault type corresponding to the optimal matching time series segment as the fault type of the real-time time series segment.

[0044] Reference Figure 2 The method for calculating the matching degree includes steps S20 to S23, which are as follows:

[0045] S20: According to the DBA algorithm, historical time series segments of different fault types are averaged to obtain a global feature average sequence, which includes multiple global features. The feature difference of each feature relative to the global feature is calculated.

[0046] Since the fault types of the distribution network are relatively complex, multiple faults may occur simultaneously in the same time period or the fault types may change over time, resulting in the phenomenon of misidentification of faults in traditional fault identification based on time domain features. Therefore, this application performs sequence averaging on historical time series segments of different fault types according to the DBA (DTW Barycenter Averaging) algorithm; calculates the difference of each feature relative to the global feature average sequence, calculates the importance of the feature according to the feature difference between different faults, and uses the importance as the weight to calculate the time domain similarity between the real-time time series segment and the historical time series segment. The DBA (DTW Barycenter Averaging) algorithm is a prior art and will not be described in detail here. Because there are many sequences for the same fault type, the purpose of averaging is to simplify the calculation without calculating the feature difference for each sequence.

[0047] In one embodiment, the calculation formula of the feature difference is:

[0048] , where Indicated in Characteristics of the fault type The difference between the global features, Representation characteristics The first in the feature average sequence values, represents the first values, express The first in the feature average sequence The rate of change of the value, represents the first The rate of change of the value, Represents the total number of elements in the feature average sequence or the global feature average sequence.

[0049] The calculation method of the change rate can be the average of the difference between the previous element value and the next element value of the current element value, which can be expressed as follows in mathematical expression: , Representation characteristics The first in the feature average sequence values; Representation characteristics The first in the feature average sequence -1 value. Similarly, we get The value of .

[0050] In other embodiments, the characteristic difference is calculated as: , where Indicated in Characteristics of the fault type The difference between the global features, Representation characteristics The first in the feature average sequence values, represents the first values, It represents the total number of elements in the feature average sequence or the global feature average sequence. By directly calculating the absolute difference between the eigenvalue and the global eigenvalue, the calculation process of the feature difference is simplified and the calculation complexity is reduced.

[0051] Exemplary, feature 1 of time segment 1: ; Feature 2: ; Feature 2 of sequence segment 2: ; Feature 2: ; The average sequence of feature 1 is ; The average sequence of feature 1 is ; The global feature average sequence is: .

[0052] S21: Calculate the importance of the features according to the feature differences of different fault types; calculate the weight based on the importance, calculate the time domain similarity between the real-time time series segment and the historical time series segment according to the weight and the element values ​​in the real-time time series segment and the historical time series segment, and in response to the time series similarity being greater than a preset threshold, use the historical time series period as the historical time series segment for time domain matching.

[0053] In one embodiment, the calculation formula of importance is: , where Representation characteristics In the The importance of each fault type, Indicates the total number of fault types, Representation characteristics In the The difference in the types of faults, Indicates The standard deviation of all feature differences in the fault type.

[0054] In other embodiments, the calculation formula of the importance may also be: , where Representation characteristics In the The importance of each fault type, Indicates the total number of fault types, Representation characteristics In the The differences in the types of failures.

[0055] For any fault type ( Fault type), the calculation formula of time domain similarity is:

[0056] , where Real-time timing segment Historical time series The time domain similarity of Representation characteristics The importance of Indicates real-time timing segment or historical time series Medium Features The total number of Represents the characteristics in the real-time time series No. elements, Represents a historical time series Medium Features No. elements, Indicates An exponential function with base .

[0057] Indicates the first feature The weight of .

[0058] If the time series similarity is greater than a preset threshold, illustratively, the preset threshold is 0.9, the historical time series period is used as the historical time series segment for time domain matching.

[0059] Different features have different manifestations in different features. For example, short-circuit faults are manifested as instantaneous surges in current and sharp drops in voltage. Transformer overload is manifested as continuous increase in current, significant load fluctuations, abnormal temperature rise, and harmonic pollution is manifested as abnormal temperature rise and voltage flicker. For the time domain features of fault data, the traditional recognition method uses the same recognition weight for each feature, which may lead to fault recognition errors or fault recognition as normal, causing accidents.

[0060] The present application divides the importance of each feature in the fault type according to the deviation distance of the feature average sequence. The greater the difference in the performance of a feature in different fault types, the greater the difference in the performance of the feature from other features when identifying the fault type, the greater the importance. Different weights are assigned to different features to achieve accurate matching in the time domain.

[0061] In other embodiments, the calculation formula of the time domain similarity may also be:

[0062] , where for the Types of faults, Real-time timing segment Historical time series The time domain similarity of Representation characteristics The importance of Indicates real-time timing segment or historical time series Medium Features The total number of Represents the characteristics in the real-time time series No. elements, Represents a historical time series Medium Features No. elements, Indicates An exponential function with base .

[0063] S22: The real-time time series segment and the historical time series segment are converted into frequency domain signals through discrete Fourier transform, and the frequency domain similarity between the historical time series segment and the real-time time series segment matched in the time domain is calculated.

[0064] In one embodiment, since the characteristic change trend caused by different faults in the early stage of the fault is small, noise may cause the time domain characteristics to be covered or distorted, and when multiple faults occur at the same time, the faults may interfere with each other and the time domain characteristics cannot accurately extract the characteristics of each fault, resulting in matching errors. Therefore, the present application uses discrete Fourier transform to convert the real-time time segment x and the historical time segment y matched in the time domain into a frequency domain signal, calculates the frequency domain similarity of the real-time time segment and the historical time segment matched in the time domain according to the characteristics of the frequency domain signal, selects the best matching time segment through the time domain similarity and frequency domain similarity, and uses the fault type of the best matching time segment as the fault type of the real-time time segment.

[0065] The calculation formula of frequency domain similarity is:

[0066] ; In the formula, Real-time timing segment Historical time series The frequency domain similarity of Indicates real-time timing segment or historical time series Medium Features The total number of is the maximum order of the harmonics of the frequency domain signal; Indicates real-time timing segment Medium Features In the frequency domain signal Number of subharmonic occurrences; Represents a historical time series Medium Features In the frequency domain signal The number of times the harmonic occurs (the number of times the harmonic occurs is 1 or 0, the 1st harmonic means that the frequency of the harmonic is 1 times that of the fundamental wave, for example, the 3rd harmonic means that the frequency of the harmonic is 3 times that of the fundamental wave).

[0067] Indicates real-time timing segment Medium Features In the frequency domain signal The amplitude of the subharmonics, Represents a historical time series Medium Features In the frequency domain signal The amplitude of the subharmonics, represents the maximum value function, Indicates An exponential function with base .

[0068] Indicates real-time timing segment Historical time series feature The greater the difference, the smaller the frequency domain similarity; the smaller the difference, the greater the frequency domain similarity.

[0069] In other embodiments, the calculation formula of frequency domain similarity is:

[0070] ; In the formula, Real-time timing segment Historical time series The frequency domain similarity of Indicates real-time timing segment or historical time series Medium Features The total number of is the maximum order of the harmonics of the frequency domain signal; Indicates real-time timing segment Medium Features In the frequency domain signal Number of subharmonic occurrences; Represents a historical time series Medium Features In the frequency domain signal Number of subharmonic occurrences; Indicates real-time timing segment Medium Features In the frequency domain signal The amplitude of the subharmonics, Represents a historical time series Medium Features In the frequency domain signal The amplitude of the subharmonics, represents the maximum value function, represents the minimum function, Indicates An exponential function with base .

[0071] S23: The product of the time domain similarity and the frequency domain similarity of the historical time series segment is used as the matching degree.

[0072] The calculation formula of the matching degree can be expressed as: ; In the formula, Real-time timing segment Historical time series The degree of matching, Real-time timing segment Historical time series The frequency domain similarity of Real-time timing segment Historical time series similarity in time domain.

[0073] The historical time segment with the greatest matching degree is used as the optimal matching time segment, and the fault type corresponding to the optimal matching time segment is used as the fault type of the real-time time segment. The fault type is fed back to the technical staff through SMS service and other means.

[0074] An embodiment of the present application also discloses a distribution system data analysis and processing system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a distribution system data analysis and processing method according to the present application is implemented.

[0075] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0076] In the present application, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high bandwidth memory HBM (High Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0077] Although this specification has shown and described a plurality of embodiments of the present application, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, modifications and alternatives without departing from the thought and spirit of the present application. It should be understood that in the process of practicing the present application, various alternatives to the embodiments of the present application described herein may be adopted.

[0078] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for analyzing and processing data of a power distribution system, characterized in that: Includes steps: Construct a time series segment, which contains multiple samples, and each sample contains multiple features representing monitoring parameters; obtain multiple historical time series segments, label the historical time series segments, and label them as fault types; calculate the matching degree of the historical time series segments, and use the historical time series segment with the greatest matching degree as the optimal matching time series segment, and use the fault type corresponding to the optimal matching time series segment as the fault type of the real-time time series segment; The method for calculating the matching degree includes the following steps: According to the DBA algorithm, the historical time series of different fault types are averaged to obtain the global feature average sequence. The global feature average sequence contains multiple global features, and the feature difference of each feature relative to the global feature is calculated; The importance of the features is calculated according to the feature differences of different fault types; the weight is calculated based on the importance, and the time domain similarity of the real-time time series segment and the historical time series segment is calculated according to the weight and the element values ​​in the real-time time series segment and the historical time series segment. In response to the time series similarity being greater than a preset threshold, the historical time series period is used as the historical time series segment for time domain matching; The real-time time series segment and the historical time series segment are converted into frequency domain signals through discrete Fourier transform, and the frequency domain similarity between the historical time series segment and the real-time time series segment matching in the time domain is calculated; The product of the time domain similarity and the frequency domain similarity of the historical time series segment is taken as the matching degree.

2. The power distribution system data analysis and processing method according to claim 1, characterized in that: The formula for calculating the feature difference is: , where Indicated in Characteristics of the fault type The difference between the global features, Representation characteristics The first in the feature average sequence values, represents the first values, express The rate of change of the ath value in the characteristic average sequence of represents the first The rate of change of the value, Represents the total number of elements in the feature average sequence or the global feature average sequence.

3. The power distribution system data analysis and processing method according to claim 1, characterized in that: The formula for calculating the feature difference is: , where Indicated in Characteristics of the fault type The difference between the global features, Representation characteristics The first in the feature average sequence values, represents the first values, Represents the total number of elements in the feature average sequence or the global feature average sequence.

4. The power distribution system data analysis and processing method according to claim 1, characterized in that: The calculation formula for importance is: , where Representation characteristics In the The importance of each fault type, Indicates the total number of fault types, Representation characteristics In the The difference in the types of faults, Indicates The standard deviation of all feature differences among the fault types.

5. The power distribution system data analysis and processing method according to claim 1, characterized in that: The calculation formula for importance is: , where Representation characteristics In the The importance of each fault type, Indicates the total number of fault types, Representation characteristics In the The differences in the types of failures.

6. The power distribution system data analysis and processing method according to claim 1, characterized in that: For any fault type, the calculation formula of time domain similarity is: , where for the Types of faults, Real-time timing segment Historical time series The time domain similarity of Representation characteristics The importance of Indicates real-time timing segment or historical time series Medium Features The total number of Represents the characteristics in the real-time time series No. elements, Represents a historical time series Medium Features No. elements, Indicates An exponential function with base .

7. The power distribution system data analysis and processing method according to claim 1, characterized in that: For any fault type, the calculation formula of time domain similarity is: , where for the Types of faults, Real-time timing segment Historical time series The time domain similarity of Representation characteristics The importance of Indicates real-time timing segment or historical time series Medium Features The total number of Represents the characteristics in the real-time time series No. elements, Represents a historical time series Medium Features No. elements, Indicates An exponential function with base .

8. The power distribution system data analysis and processing method according to claim 1, characterized in that: The calculation formula of frequency domain similarity is: ; In the formula, Real-time timing segment Historical time series The frequency domain similarity of Indicates real-time timing segment or historical time series Medium Features The total number of is the maximum order of the harmonics of the frequency domain signal; Indicates real-time timing segment Medium Features In the frequency domain signal Number of subharmonic occurrences; Represents a historical time series Medium Features In the frequency domain signal Number of subharmonic occurrences; Indicates real-time timing segment Medium Features In the frequency domain signal The amplitude of the subharmonics, Represents a historical time series Medium Features In the frequency domain signal The amplitude of the subharmonics, represents the maximum value function, Indicates An exponential function with base .

9. The power distribution system data analysis and processing method according to claim 1, characterized in that: The calculation formula of frequency domain similarity is: ; In the formula, Real-time timing segment Historical time series The frequency domain similarity of Indicates real-time timing segment or historical time series Medium Features The total number of is the maximum order of the harmonics of the frequency domain signal; Indicates real-time timing segment Medium Features In the frequency domain signal Number of subharmonic occurrences; Represents a historical time series Medium Features In the frequency domain signal Number of subharmonic occurrences; Indicates real-time timing segment Medium Features In the frequency domain signal The amplitude of the subharmonics, Represents a historical time series Medium Features In the frequency domain signal The amplitude of the subharmonics, represents the maximum value function, represents the minimum function, Indicates An exponential function with base .

10. A data analysis and processing system for a power distribution system, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the power distribution system data analysis and processing method according to any one of claims 1 to 9 is implemented.

Images