A Smart Management Method for Power Distribution Systems Based on Big Data

Abstract

This invention discloses an intelligent management method for power distribution systems based on big data, relating to the field of big data technology, which improves the accuracy and timeliness of power grid fault diagnosis. This invention sets several core points for historical operational data collected under the same sub-cycle of different maintenance cycles, obtains the confidence level of each core point, and then obtains standard operational data intervals for various types of operational data based on the confidence level distribution. Abnormal historical data fragments under various fault item name combinations are extracted from the historical operational data through these standard operational data intervals. Abnormal data vector combinations are then established based on these abnormal historical data fragments under different fault item name combinations. The existence of anomalies in real-time operational data is determined by using the standard operational data intervals. The abnormal data vector combinations are matched based on the types and quantities of real-time operational data that are determined to be abnormal, thereby identifying the current fault items in the power distribution area.

Description

Technical Field

[0001] This invention relates to the field of big data technology, specifically a smart management method for power distribution systems based on big data. Background Technology

[0002] In modern power systems, distribution and consumption areas form a complex and vast network. Distribution areas are responsible for the distribution and transmission of electricity, while consumption areas are where the electricity is actually used. With the continuous growth of electricity demand and the increasing complexity of power systems, the operation and maintenance management of distribution areas faces enormous challenges.

[0003] Traditional power distribution area operation and maintenance methods mainly rely on manual periodic inspections and experience-based judgment. This approach is not only inefficient but also makes it difficult to comprehensively and promptly identify potential faults during the operation of the power distribution area. Due to the lack of effective analysis and utilization of historical operational data, it is difficult to accurately grasp the operational status and patterns of the power distribution area. Furthermore, when faced with complex and ever-changing fault conditions, traditional methods often fail to quickly and accurately identify faults, leading to untimely fault handling and impacting the stability and reliability of power supply.

[0004] Furthermore, the geographical distribution of different power distribution zones varies significantly, and the relationships between power consumption zones and power distribution zones are also quite complex. How to rationally connect these zones based on their geographical distribution and conduct scientific operation and maintenance management of the power distribution zones is a pressing issue in the current power system operation and maintenance field. Moreover, the analysis and processing of power distribution zone operation data lacks systematic and standardized methods, making it difficult to effectively extract valuable information from massive amounts of historical operation data to support fault diagnosis and operation and maintenance decisions. Therefore, this paper proposes a big data-based intelligent management method for power distribution systems. Summary of the Invention

[0005] To address the aforementioned technical problems, the present invention aims to provide an intelligent management method for power distribution systems based on big data.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A big data-based intelligent management method for power distribution systems includes the following steps:

[0008] Step 1: Set up the power distribution area and the power consumption area. Based on the geographical distribution, link the power consumption area and the power distribution area together, set multiple operation and maintenance cycles for each power distribution area, and then collect the historical operation data of each power distribution area.

[0009] Step 2: Divide the operation and maintenance cycle into operation and maintenance sub-cycles, and set several core points for historical operation data collected under the same operation and maintenance sub-cycle but different operation and maintenance cycles. Obtain the confidence level of each core point, and then obtain the standard operation data range of various types of operation data based on the confidence level distribution.

[0010] Step 3: Set up multiple fault item names, extract abnormal historical data fragments under various fault item name combinations from historical operating data through standard operating data ranges, and then establish abnormal data vector combinations based on abnormal historical data fragments under different fault item name combinations.

[0011] Step 4: Obtain real-time operating data of the power distribution area, and determine whether there are any anomalies in the real-time operating data through the standard operating data range. Based on the types and quantities of real-time operating data that are determined to be abnormal, match the abnormal data vector combinations to determine the current faults in the power distribution area.

[0012] Furthermore, the process of setting multiple operation and maintenance cycles for each power distribution area includes:

[0013] Set up m electronic distribution areas and n electronic consumption areas, and assign numbers to each electronic distribution area and electronic consumption area respectively;

[0014] Set the association range radius, with the location of each electronic use area as the center, and associate the electronic distribution area with the electronic use area within the association range radius. Do not operate on the electronic distribution area outside the association range radius.

[0015] Data acquisition units are set up for each electronic distribution area and electronic consumption area, and the data acquisition units are composed of several types of sensors;

[0016] Establish a two-dimensional coordinate system and obtain the historical electricity consumption record curves of each electronic consumption area. With a seven-day maintenance cycle, map the historical electricity consumption record curves of each electronic consumption area under one maintenance cycle onto the same two-dimensional coordinate system.

[0017] Set electricity consumption thresholds and probability thresholds, and then mark the continuous intervals in the historical electricity consumption record curve where the electricity consumption is greater than or equal to the electricity consumption threshold as high electricity consumption intervals, and mark the continuous intervals where the electricity consumption is less than the electricity consumption threshold as low electricity consumption intervals.

[0018] Several maintenance time nodes are set according to the maintenance cycle duration. The number of times the same maintenance time node is located in the high power consumption range or low power consumption range under different cycles is counted. If the proportion of the number of times located in the high power consumption range is greater than or equal to the probability threshold, the maintenance time node is recorded as a high power consumption time point. If the proportion of the number of times located in the low power consumption range is greater than or equal to the probability threshold, the maintenance time node is recorded as a low power consumption time point. Otherwise, it is recorded as a pending power consumption time point.

[0019] Furthermore, the process of collecting historical operating data for each power distribution area includes:

[0020] Connect adjacent and similar electricity consumption times, divide the operation and maintenance cycle of each electricity consumption area into several operation and maintenance sub-cycles, and accumulate and average the curve segments of historical electricity consumption records in each operation and maintenance sub-cycle to obtain the expected electricity consumption of the electricity consumption area in each operation and maintenance sub-cycle.

[0021] Based on the time segment corresponding to the operation and maintenance sub-cycle of each electronic consumption area, set the corresponding data acquisition interval for the data acquisition unit of the electronic distribution area associated with the electronic consumption area;

[0022] Then, each time a data acquisition interval is passed, each data acquisition unit collects the historical or real-time operating data of the distribution area where it is located, and marks the historical or real-time operating data with the corresponding number according to the distribution area associated with the data acquisition interval.

[0023] Furthermore, the process of setting core locations based on historical operational data includes:

[0024] All historical operating data between a pair of power distribution zones and power consumption zones are selected sequentially, and multiple energy interaction intervals are set.

[0025] The number of operation and maintenance sub-cycles between the distribution area and the user area is obtained based on their numbers, and k data distribution spaces are set up, where k is an integer greater than 0.

[0026] The energy interaction intervals and corresponding operation and maintenance sub-cycles of the historical energy interaction volume are the same, but the data collected in different historical operation data of the operation and maintenance cycle are input into the same data distribution space.

[0027] For any type of historical operational data in the data distribution space, the historical operational data is divided into several historical data points by the operation and maintenance time nodes. Then, the historical data points under the same operation and maintenance time node are numerically distributed, and several historical data points with equal numerical intervals are selected from the numerical distribution results in ascending order and recorded as core points.

[0028] Obtain the confidence level f for each core point, where the formula for calculating the confidence level f is:

[0029] ;

[0030] in Let represent the confidence level of the z-th historical data point, denoted as the core point, h be the bandwidth parameter greater than 0, and j be the total number of historical data points. and K represents the historical data point z-th, denoted as the core point. () represents the kernel function, where z and j are positive integers greater than 0, and z is less than or equal to j;

[0031] Furthermore, the process of obtaining the standard operating data ranges for various types of operating data includes:

[0032] Set a confidence threshold, distribute the confidence scores f of each core point normally, select the portion of the normal distribution results that are greater than or equal to the confidence threshold, and sort the selected portion according to the corresponding value of the core point.

[0033] Then, select the parts with continuous size relationship from the sorting results and record them as the standard value intervals of the corresponding operation and maintenance time nodes. Connect the standard value intervals under each operation and maintenance time node in sequence to obtain the standard operating data intervals between the corresponding power distribution area and the power consumption area under the corresponding energy interaction interval and operation and maintenance sub-cycle.

[0034] The process of repeatedly acquiring standard operating data ranges is used to acquire various standard operating data ranges for each distribution and consumption area under different energy interaction ranges and operation and maintenance sub-cycles.

[0035] Furthermore, the process of extracting abnormal historical data fragments under various combinations of fault item names includes:

[0036] Each electronic distribution area is set up with a fault item set, which records several fault item names;

[0037] By comparing the standard operating data ranges of the power distribution area and the power consumption area under different energy interaction intervals and operation and maintenance sub-cycles, the historical operating data of the power distribution area under each operation and maintenance cycle is compared. If it is determined that there are data segments in the historical operating data that are not within the corresponding standard operating data range, the corresponding data segments are marked as abnormal historical data segments; otherwise, no operation is performed.

[0038] Furthermore, the process of establishing the combination of abnormal data vectors includes:

[0039] Obtain the name of the fault item corresponding to each abnormal historical data segment;

[0040] First, select the abnormal historical data segments associated with each electronic distribution area that correspond to only one type of fault name, and map the selected abnormal historical data segments onto the same two-dimensional coordinate system.

[0041] The abnormal historical data segment is divided into several abnormal data points by the operation and maintenance time nodes, and abnormal data vectors are established in chronological order, with the first abnormal data point on the abnormal historical data segment as the starting position and the next abnormal data point as the ending position.

[0042] The abnormal data vectors between the same pair of operation and maintenance time nodes are vector-added to obtain the abnormal change trend vector between the two corresponding operation and maintenance time nodes. The abnormal change trend vectors between each pair of operation and maintenance time nodes are connected in sequence to obtain the abnormal change trend vector of the corresponding type of operation data.

[0043] Then select abnormal historical data segments with two, three, ..., up to all types of fault item names, and establish corresponding abnormal data vector combinations.

[0044] Furthermore, the process of determining the current faults in the power distribution area based on the types and quantities of abnormal real-time operational data includes:

[0045] When the operation and maintenance sub-cycle begins, each data acquisition unit acquires various real-time operating data of its respective power distribution area and retrieves the standard operating data range according to the power distribution area and the power consumption area number.

[0046] Determine if there is any real-time running data that is not within its corresponding standard running data range. If it is determined that there is, then all real-time running data is within its corresponding standard running data range, and no operation is performed.

[0047] If it is determined that there is real-time running data that is not within its corresponding standard running data range, then the process of generating abnormal data vectors is adopted to generate real-time abnormal data vectors for three consecutive operation and maintenance time nodes.

[0048] First, match the corresponding abnormal data vector combinations according to the types and quantities of real-time abnormal data vectors, and then obtain the overlap between the real-time abnormal data vectors and the abnormal data vector combinations of the same type of data vectors.

[0049] Set an overlap threshold, select the abnormal data vector combination with the largest overlap that corresponds to the fault item name combination, record it as the current fault item name of the corresponding power distribution area, and send the current fault item name to the relevant maintenance personnel for maintenance.

[0050] Otherwise, according to step three, a new fault item name is set in the fault item set, and the corresponding abnormal data vector combination for that fault item name is generated;

[0051] The process of repeatedly checking for any abnormalities in the power distribution area at each maintenance time point continues until the end of the current maintenance sub-cycle.

[0052] Compared with the prior art, the beneficial effects of the present invention are:

[0053] 1. This invention divides the operation and maintenance cycle into sub-cycles, sets core locations, and obtains confidence levels to obtain a standard operating data range. By utilizing this standard operating data range, abnormal data segments are extracted from historical operating data, and abnormal data vector combinations are established. After acquiring real-time operating data from the power distribution area, it can accurately determine whether there are anomalies in the real-time operating data and match abnormal data vector combinations based on the anomalies, thereby accurately identifying the current faults in the power distribution area and improving the accuracy and timeliness of fault diagnosis.

[0054] 2. By combining the geographical distribution of power distribution areas and power consumption areas, the various areas are interconnected, which effectively adapts to the complex network structure of the power system and helps to fully understand the operating status of the power system, thereby improving the control and management capabilities of the entire power network system. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention.

[0056] Figure 1 This is a flowchart of a method for intelligent management of power distribution systems based on big data, as described in this invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0058] like Figure 1 As shown, a smart management method for power distribution systems based on big data includes the following steps:

[0059] Step 1: Set up the power distribution area and the power consumption area. Based on the geographical distribution, link the power consumption area and the power distribution area together, set multiple operation and maintenance cycles for each power distribution area, and then collect the historical operation data of each power distribution area.

[0060] Step 2: Divide the operation and maintenance cycle into operation and maintenance sub-cycles, and set several core points for historical operation data collected under the same operation and maintenance sub-cycle but different operation and maintenance cycles. Obtain the confidence level of each core point, and then obtain the standard operation data range of various types of operation data based on the confidence level distribution.

[0061] Step 3: Set up multiple fault item names, extract abnormal historical data fragments under various fault item name combinations from historical operating data through standard operating data ranges, and then establish abnormal data vector combinations based on abnormal historical data fragments under different fault item name combinations.

[0062] Step 4: Obtain real-time operating data of the power distribution area, and determine whether there are any anomalies in the real-time operating data through the standard operating data range. Based on the types and quantities of real-time operating data that are determined to be abnormal, match the abnormal data vector combinations to determine the current faults in the power distribution area.

[0063] Furthermore, step one is achieved through the following steps:

[0064] Step 101: Set up m distribution areas and n usage areas, and assign numbers a1, a2, ..., a3 to each distribution area and usage area. m b1, b2, ..., b n , where m and n are natural numbers greater than 0;

[0065] Set the association range radius, with the location of each electronic use area as the center, and associate the electronic distribution area with the electronic use area within the association range radius. Do not operate on the electronic distribution area outside the association range radius.

[0066] It should be noted that there may be multiple electronic distribution areas that are simultaneously associated with electronic usage areas.

[0067] Step 102: Set up data acquisition units for each electronic distribution area and electronic consumption area. The data acquisition unit consists of several types of sensors, such as voltage sensors, temperature sensors, power sensors, etc.

[0068] Establish a two-dimensional coordinate system and obtain the historical electricity consumption record curves of each electronic consumption area. With a seven-day maintenance cycle (excluding holidays), map the historical electricity consumption record curves of each electronic consumption area under one maintenance cycle onto the same two-dimensional coordinate system.

[0069] Set electricity consumption thresholds and probability thresholds, and then mark the continuous intervals in the historical electricity consumption record curve where the electricity consumption is greater than or equal to the electricity consumption threshold as high electricity consumption intervals, and mark the continuous intervals where the electricity consumption is less than the electricity consumption threshold as low electricity consumption intervals.

[0070] Several maintenance time nodes are set according to the maintenance cycle duration. The number of times the same maintenance time node is located in the high power consumption range or low power consumption range under different cycles is counted. If the proportion of the number of times located in the high power consumption range is greater than or equal to the probability threshold, the maintenance time node is recorded as a high power consumption time point. If the proportion of the number of times located in the low power consumption range is greater than or equal to the probability threshold, the maintenance time node is recorded as a low power consumption time point. Otherwise, it is recorded as a pending power consumption time point.

[0071] By connecting adjacent and similar electricity consumption times and dividing the operation and maintenance cycle of each electricity consumption area into several operation and maintenance sub-cycles, and accumulating and averaging the curve segments of historical electricity consumption records in each operation and maintenance sub-cycle, the expected electricity consumption of the electricity consumption area in each operation and maintenance sub-cycle can be obtained.

[0072] Step 103: Based on the time segment corresponding to the operation and maintenance sub-cycle of each electronic consumption area, set the corresponding data acquisition interval for the data acquisition unit of the electronic distribution area associated with the electronic consumption area;

[0073] It should be noted that the time segments corresponding to the operation and maintenance sub-cycles of different electronic distribution areas are different, and depending on the number of electronic distribution areas associated with the electronic distribution area, the data acquisition units of the electronic distribution area have multiple non-interfering data acquisition intervals in the same time interval.

[0074] Then, each time a data acquisition interval is passed, each data acquisition unit collects the historical or real-time operating data of the distribution area where it is located, and marks the historical or real-time operating data with the corresponding number according to the distribution area associated with the data acquisition interval.

[0075] The types of operational data include voltage data, current data, and energy interaction quantities.

[0076] When using this method, refer to steps 101 to 103:

[0077] By setting up power distribution zones and power consumption zones, and linking them together based on their geographical distribution, the interrelationships between these zones can be clearly and accurately displayed, facilitating a comprehensive understanding of the overall layout of power supply and consumption. Simultaneously, setting multiple operation and maintenance cycles for each power distribution zone and collecting historical operational data provides a rich and comprehensive data foundation for subsequent data analysis and fault diagnosis.

[0078] Furthermore, step two is achieved through the following steps:

[0079] Step 201: Select all historical operating data between a pair of power distribution areas and power consumption areas in sequence, and set multiple energy interaction ranges, such as zero to five thousand kilowatt-hours and five thousand to seven thousand kilowatt-hours.

[0080] The number of operation and maintenance sub-cycles between the distribution area and the user area is obtained based on their numbers, and k data distribution spaces are set up, where k is an integer greater than 0.

[0081] The energy interaction intervals and corresponding operation and maintenance sub-cycles of the historical energy interaction volume are the same, but the data collected in different historical operation data of the operation and maintenance cycle are input into the same data distribution space.

[0082] For any type of historical operational data in the data distribution space, the historical operational data is divided into several historical data points by the operation and maintenance time nodes. Then, the historical data points under the same operation and maintenance time node are numerically distributed, and several historical data points with equal numerical intervals are selected from the numerical distribution results in ascending order and recorded as core points.

[0083] Step 202: Obtain the confidence level f for each core point, where the formula for calculating the confidence level f is:

[0084] ;

[0085] in Let represent the confidence level of the z-th historical data point, denoted as the core point, h be the bandwidth parameter greater than 0, and j be the total number of historical data points. and K represents the historical data point z-th, denoted as the core point. () represents the kernel function, where z and j are positive integers greater than 0, and z is less than or equal to j;

[0086] Set a confidence threshold, distribute the confidence scores f of each core point normally, select the portion of the normal distribution results that are greater than or equal to the confidence threshold, and sort the selected portion according to the corresponding value of the core point.

[0087] Since the historical operating data contained in the data distribution space are all generated in the same pair of distribution and consumption areas within the same energy interaction interval, under different operation and maintenance cycles but corresponding to the same operation and maintenance cycle, the range of various operating data for the corresponding distribution and consumption areas under each energy interaction interval is stable within a fixed value range.

[0088] Step 203: Select the parts with continuous size relationship from the sorting results and record them as the standard value intervals of the corresponding operation and maintenance time nodes. Then connect the standard value intervals under each operation and maintenance time node in sequence to obtain the standard operating data intervals between the corresponding power distribution area and the power consumption area under the corresponding energy interaction interval and operation and maintenance sub-cycle.

[0089] The process of repeatedly acquiring standard operating data ranges is used to acquire various standard operating data ranges for each distribution and consumption area under different energy interaction ranges and operation and maintenance sub-cycles.

[0090] When using this method, refer to steps 201 to 203:

[0091] The operation and maintenance cycle is divided into sub-cycles. Core data points are established for historical operational data collected within the same sub-cycle across different maintenance cycles. Confidence levels are obtained, and standard operational data ranges are determined for various types of operational data. This accurately reflects the normal fluctuation range of power operation data while fully considering the data's changing characteristics at different time scales. The determination of standard operational data ranges provides a scientific and reliable basis for subsequent assessments of whether real-time operational data is abnormal, helping to promptly identify potential risks in power operation.

[0092] Furthermore, step three is achieved through the following steps:

[0093] Step 301: Set up a fault item set for each electronic distribution area. The fault item set records several fault item names, such as line overload fault, short circuit fault, equipment fault, etc.

[0094] By comparing the standard operating data ranges of the power distribution area and the power consumption area under different energy interaction intervals and operation and maintenance sub-cycles, the historical operating data of the power distribution area under each operation and maintenance cycle is compared. If it is determined that there are data segments in the historical operating data that are not within the corresponding standard operating data range, the corresponding data segments are marked as abnormal historical data segments; otherwise, no operation is performed.

[0095] Obtain the name of the fault item corresponding to each abnormal historical data segment. It should be noted that there may be one or more fault item names corresponding to each abnormal historical data segment.

[0096] Step 302: First, select the abnormal historical data segments associated with each electronic distribution area that correspond to only one type of fault name, and map the selected abnormal historical data segments onto the same two-dimensional coordinate system.

[0097] The abnormal historical data segment is divided into several abnormal data points by the operation and maintenance time nodes, and abnormal data vectors are established in chronological order, with the first abnormal data point on the abnormal historical data segment as the starting position and the next abnormal data point as the ending position.

[0098] The abnormal data vectors between the same pair of operation and maintenance time nodes are vector-added to obtain the abnormal change trend vector between the two corresponding operation and maintenance time nodes. The abnormal change trend vectors between each pair of operation and maintenance time nodes are connected in sequence to obtain the abnormal change trend vector of the corresponding type of operation data.

[0099] Then select abnormal historical data segments with two, three, ..., up to all types of fault item names, and establish corresponding abnormal data vector combinations.

[0100] When using this method, refer to steps 301 to 302:

[0101] Multiple fault item names are set, and abnormal historical data fragments under various fault item name combinations are extracted from historical operation data. These fragments are then used to create abnormal data vector combinations, effectively integrating the abnormal data of fault items. Furthermore, by using abnormal historical data fragments under different fault item name combinations, corresponding abnormal data vector combinations are established, clearly displaying the abnormal data characteristics corresponding to different fault items.

[0102] Furthermore, step four is achieved through the following steps:

[0103] Step 401: Before the start of an operation and maintenance sub-cycle, obtain the real-time energy storage of each distribution area, and determine whether the real-time energy storage of the distribution area is greater than or equal to the total expected power consumption of the related electronic areas under the corresponding operation and maintenance sub-cycle.

[0104] If it is determined that the real-time energy storage in the power distribution area is less than the total expected power consumption, then an energy allocation request is sent to the other power distribution areas based on the difference between the total expected power consumption and the real-time energy storage.

[0105] If the real-time energy storage level is equal to the total estimated electricity consumption, no action will be taken.

[0106] If it is determined that the real-time energy storage is greater than the total expected electricity consumption, then the principle of proximity is adopted, and energy less than or equal to the difference between the real-time energy storage and the total expected electricity consumption is sent to the nearest power distribution area with an energy dispatch request.

[0107] Step 402: When the operation and maintenance sub-cycle begins, each data acquisition unit acquires various real-time operating data of its respective power distribution area and retrieves the standard operating data range according to the power distribution area and the power consumption area number.

[0108] Determine if there is any real-time running data that is not within its corresponding standard running data range. If it is determined that there is, then all real-time running data is within its corresponding standard running data range, and no operation is performed.

[0109] If it is determined that there is real-time running data that is not within its corresponding standard running data range, then the process of generating abnormal data vectors is adopted to generate real-time abnormal data vectors for three consecutive operation and maintenance time nodes.

[0110] First, match the corresponding abnormal data vector combinations based on the types and quantities of real-time abnormal data vectors, and then obtain the overlap between the real-time abnormal data vectors and the abnormal data vector combinations of the same type of data vectors.

[0111] Step 403: Set the overlap threshold, select the abnormal data vector combination with the overlap greater than the overlap threshold and the largest overlap, and record it as the current fault item name of the corresponding power distribution area, and send the current fault item name to the relevant maintenance personnel for maintenance.

[0112] Otherwise, according to step three, a new fault item name is set in the fault item set, and the corresponding abnormal data vector combination for that fault item name is generated;

[0113] The process of repeatedly checking for any abnormalities in the power distribution area at each maintenance time point continues until the end of the current maintenance sub-cycle.

[0114] When using this method, refer to steps 401 and 403:

[0115] First, based on the estimated power consumption of each electronic distribution area in each operation and maintenance sub-cycle, and the real-time energy storage of each electronic distribution area, the energy of each electronic distribution area is redistributed to effectively avoid the problem of insufficient power supply later. Then, the real-time operation data of the electronic distribution area is obtained, and the standard operation data range is used to determine whether there are any anomalies in the real-time operation data. This enables the rapid matching of anomaly data vector combinations based on the type and quantity of anomaly data, and accurate identification of the current faults in the electronic distribution area, effectively shortening the response time for fault detection and handling.

[0116] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for intelligent management of power distribution systems based on big data, characterized in that, Includes the following steps: Step 1: Set up the power distribution area and the power consumption area. Based on the geographical distribution, link the power consumption area and the power distribution area together, set multiple operation and maintenance cycles for each power distribution area, and then collect the historical operation data of each power distribution area. The process of setting multiple operation and maintenance cycles for each power distribution area includes: Each electronic distribution area and electronic consumption area is assigned a number and an associated range radius is set. Taking the location of each electronic consumption area as the center, the electronic distribution area and the electronic consumption area within the associated range radius are associated. A data acquisition unit is set for each electronic distribution area and electronic consumption area. The data acquisition unit consists of several types of sensors. Obtain historical electricity consumption curves for each electricity consumption area, and divide the historical electricity consumption curves into several high electricity consumption intervals and low electricity consumption intervals, with a seven-day maintenance cycle. Several maintenance time nodes are set according to the maintenance cycle duration. The number of times the same maintenance time node is located in the high power consumption range or low power consumption range under different cycles is counted. If the proportion of the number of times located in the high power consumption range is greater than or equal to the probability threshold, the maintenance time node is recorded as a high power consumption time point. If the proportion of the number of times located in the low power consumption range is greater than or equal to the probability threshold, the maintenance time node is recorded as a low power consumption time point. Otherwise, it is recorded as a pending power consumption time point. Step 2: Divide the operation and maintenance cycle into operation and maintenance sub-cycles, and set several core points for historical operation data collected under the same operation and maintenance sub-cycle but different operation and maintenance cycles. Obtain the confidence level of each core point, and then obtain the standard operation data range of various types of operation data based on the confidence level distribution. The process of setting several core data points for historical operational data collected under the same operational sub-cycle but in different operational cycles, and obtaining the confidence level of each core data point includes: All historical operating data between a pair of distribution and consumption areas are selected sequentially, and multiple energy interaction intervals are set. Based on the numbers of the distribution and consumption areas, the number of operation and maintenance sub-cycles between them is obtained, and k data distribution spaces are set, where k is an integer greater than 0. The historical energy interaction volume falls within the same energy interaction interval and corresponds to the same operation and maintenance sub-cycle, but the operation and maintenance cycle data collection is different. All data from historical operation data are input into the same data distribution space; For any type of historical operational data in the data distribution space, the historical operational data is divided into several historical data points by the operation and maintenance time nodes. Then, the historical data points under the same operation and maintenance time node are numerically distributed. In the numerical distribution results, several historical data points with equal numerical intervals are selected from small to large and recorded as core points. The confidence level of each core point is obtained. Step 3: Set up multiple fault item names, extract abnormal historical data fragments under various fault item name combinations from historical operating data through standard operating data ranges, and then establish abnormal data vector combinations based on abnormal historical data fragments under different fault item name combinations. Step 4: Obtain real-time operating data of the power distribution area, and determine whether there are any anomalies in the real-time operating data through the standard operating data range. Based on the types and quantities of real-time operating data that are determined to be abnormal, match the abnormal data vector combinations to determine the current faults in the power distribution area.

2. The intelligent management method for power distribution systems based on big data according to claim 1, characterized in that, The process of collecting historical operating data for each power distribution area includes: Connect adjacent and similar electricity consumption times, divide the operation and maintenance cycle of each electricity consumption area into several operation and maintenance sub-cycles, and accumulate and average the curve segments of historical electricity consumption records in each operation and maintenance sub-cycle to obtain the expected electricity consumption of the electricity consumption area in each operation and maintenance sub-cycle. Based on the time segment of the operation and maintenance sub-cycle of each electronic consumption area, the data acquisition units of the associated electronic distribution area are set with corresponding data acquisition intervals. Then, each data acquisition unit collects the historical operation data of its respective electronic distribution area after each data acquisition interval.

3. The intelligent management method for power distribution systems based on big data according to claim 1, characterized in that, The process of obtaining the standard runtime data range for various types of runtime data includes: Set a confidence threshold, distribute the confidence scores of each core point normally, select the portion of the normal distribution results that is greater than or equal to the confidence threshold, and sort the selected portion according to the corresponding value of the core point. Select the parts with continuous size relationships from the sorting results and record them as the standard value intervals of the corresponding operation and maintenance time nodes. Then connect the standard value intervals under each operation and maintenance time node in sequence to obtain the standard operating data intervals between the corresponding power distribution area and the power consumption area under the corresponding energy interaction interval and operation and maintenance sub-cycle.

4. The intelligent management method for power distribution systems based on big data according to claim 1, characterized in that, The process of extracting abnormal historical data segments under various combinations of fault item names includes: Each electronic distribution area is set up with a fault item set, which records several fault item names; By comparing the standard operating data ranges of the power distribution area and the power consumption area under different energy interaction intervals and operation and maintenance sub-cycles, the historical operating data of the power distribution area under each operation and maintenance cycle is compared. If it is determined that there are data segments in the historical operating data that are not within the corresponding standard operating data range, the corresponding data segments are marked as abnormal historical data segments; otherwise, no operation is performed.

5. The intelligent management method for power distribution systems based on big data according to claim 1, characterized in that, The process of creating anomaly data vector combinations includes: Obtain the fault item name corresponding to each abnormal historical data segment, select the abnormal historical data segments associated with each power distribution area that correspond to only one fault item name, and map the selected abnormal historical data segments onto the same two-dimensional coordinate system. The abnormal historical data segment is divided into several abnormal data points by the operation and maintenance time nodes, and abnormal data vectors are established in chronological order, with the first abnormal data point on the abnormal historical data segment as the starting position and the next abnormal data point as the ending position. The abnormal data vectors between the same pair of operation and maintenance time nodes are vector-added to obtain the abnormal change trend vector between the two corresponding operation and maintenance time nodes. The abnormal change trend vectors between each pair of operation and maintenance time nodes are connected in sequence to obtain the abnormal change trend vector of the corresponding type of operation data. Next, select abnormal historical data segments with two, three, ..., or all types of fault names, and establish corresponding abnormal data vector combinations.

6. The intelligent management method for power distribution systems based on big data according to claim 1, characterized in that, The process of determining the current faults in the power distribution area based on the types and quantities of abnormal real-time operating data includes: When the operation and maintenance sub-cycle begins, each data acquisition unit acquires various real-time operating data of its respective power distribution area and retrieves the standard operating data range according to the power distribution area and the power consumption area number. Determine whether there is any real-time operational data that is outside its corresponding standard operational data range, and generate real-time abnormal data vectors for three consecutive operation and maintenance time nodes based on the determination result; Match the corresponding abnormal data vector combination according to the type and quantity of real-time abnormal data vectors, and then obtain the overlap between the real-time abnormal data vector and the data vector combination of the same type of data vector. Set an overlap threshold, select the abnormal data vector combination with the largest overlap that corresponds to the fault item name combination, record it as the current fault item name of the corresponding power distribution area, and send the current fault item name to the relevant maintenance personnel for maintenance. Otherwise, set a new fault item name in the fault item set and generate the corresponding abnormal data vector combination for that fault item name.

Images