A method for power distribution network fault prediction

Abstract

This invention discloses a method for predicting faults in power distribution networks. It addresses the technical problems of failing to properly analyze abnormal data and failing to make reasonable predictions based on fault causes. This invention analyzes historical fault data, taking voltage as an example of a fault cause. By classifying voltage into normal and fault voltages, it further determines whether there is an overlap between the two. For cases where there is overlap, it analyzes the overall operating status of the power distribution network by acquiring real-time voltage changes, thereby achieving the prediction purpose. For cases where there is no overlap, it analyzes the overall operating status of the power distribution network by acquiring real-time voltage values ​​and comparing them with corresponding interval values. Through data analysis, abnormal situations can be detected in a timely manner, allowing for early warning and handling, thus reducing the occurrence of faults.

Description

Technical Field

[0001] This invention relates to the field of power distribution network fault prediction technology, specifically a method for power distribution network fault prediction. Background Technology

[0002] Distribution networks are an important component of power systems. With the rapid development of smart grids and the large-scale, unpredictable access of distributed power sources, distribution network fault information is becoming increasingly complex, making accurate and rapid fault analysis increasingly difficult.

[0003] According to the patent application number CN201810866754.3, the method of the patent includes: extracting the time lead, data density, downsampled waveform and local waveform from the fault recording data in the distribution network section to be predicted; inputting the time lead, data density, downsampled waveform and local waveform into the fault prediction model to obtain the prediction result, wherein the fault prediction model includes a deep convolutional neural network, a masking unit and a long short-term memory network unit.

[0004] The aforementioned patent aims to predict faults by analyzing fault recording data. However, it does not adequately address abnormal data in the distribution network. Factors such as voltage, current, and power directly affect the normal operation of the distribution network during operation. Therefore, reasonable data analysis is necessary to determine the cause of the fault and to provide timely prediction and early warning based on the cause. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for predicting power distribution network faults, which solves the problems of not conducting reasonable analysis of abnormal data and not making reasonable predictions based on the causes of faults.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for predicting faults in a power distribution network, which specifically includes the following steps:

[0007] Step 1: First, obtain the historical fault data of the distribution network, and the time period of the historical fault data is T. The historical fault data includes: number of faults, fault causes, and faulty equipment. Then, based on the fault causes, determine whether there are equipment with multiple faults and obtain the judgment result. The judgment result includes the presence result and the absence result. Specifically, equipment with multiple faults is defined as equipment that has experienced faults two or more times.

[0008] Step 2: Next, obtain the non-existent result and obtain the cause of the fault. The specific causes of the fault may include voltage, current and power. At the same time, obtain the number of faults corresponding to voltage, current and power, and select the fault cause with the most faults for analysis. In this application, the fault with voltage has the most faults for analysis.

[0009] Step 3: Obtain all voltage data from historical data, and determine the normal voltage and fault voltage based on the voltage data. Simultaneously, determine the intersection between the normal voltage and fault voltage to obtain the intersection judgment result. Then, obtain the voltage fault comparison result based on the intersection judgment result. The analysis method for voltage changes is as follows:

[0010] S1: Label all voltage data and denote them as Ui, where i = 1, 2, ..., j, and j represents the fault number. The specific value of j is determined according to the actual situation. Then, classify the voltage data into normal voltage and fault voltage according to the cause of the fault. The normal voltage is denoted as Uzk, k = 1, 2, ..., o, and the fault voltage is denoted as Ugh, h = 1, 2, ..., f. Specifically, if the cause of the fault is an abnormal voltage leading to an overall fault, the corresponding voltage is marked as a fault voltage; otherwise, if it is not caused by voltage, the corresponding voltage is marked as a normal voltage.

[0011] S2: Obtain all normal voltages Uzk and fault voltages Ugh and determine if there is any intersection between them. The intersection result includes: intersection result and no intersection result. Here, intersection means that voltages with the same value may be normal voltages or fault voltages. If both normal voltages and fault voltages have the same value, then there is an intersection; otherwise, if neither normal voltage nor fault voltage has the same value, then there is no intersection.

[0012] S3: When the intersection judgment result is that there is no intersection result, the maximum and minimum values ​​of the normal voltage are obtained, and the normal voltage range is generated using the minimum and maximum values. Similarly, the fault voltage range is obtained. The normal voltage range and fault voltage range obtained here are used to compare the generated real-time voltage to judge the overall operating status of the distribution network, thereby achieving the purpose of prediction. At the same time, no intersection also means that the range value obtained by the maximum and minimum values ​​of the normal voltage does not include any voltage value in the fault voltage, which means that the two ranges obtained are two non-interfering ranges. For example, the normal voltage range is [37, 45] and the fault voltage range is [50, 60].

[0013] S4: When the intersection judgment result indicates that an intersection exists, obtain the voltages with the same value and use them as the intersection voltages. Then, using the time when the intersection voltage is generated as the endpoint, obtain the changes in normal voltage and fault voltage within the time t before the endpoint. It is assumed that only one intersection voltage is generated here.

[0014] Step 4: Based on the real-time data of the distribution network obtained by high-precision sensors (specifically, voltmeters), the real-time voltage of the distribution network is measured. Then, the real-time voltage is matched with the voltage comparison results to obtain the distribution network prediction results. The specific matching method is as follows:

[0015] P1: When the obtained results do not have any intersection, the real-time voltage is compared with the normal voltage range and the fault voltage range. The specific comparison method is as follows:

[0016] The maximum value of the normal voltage range and the minimum value of the fault voltage range are obtained and denoted as Uzmax and Ugmin, respectively. An early warning signal is generated when the real-time voltage is greater than Uzmax or greater than Ugmin. This means that an early warning signal is generated when either condition is met, and the real-time voltage is considered normal only when it is within the normal voltage range.

[0017] P2: When the obtained result is "existing," the real-time voltage is acquired and its change within time t is judged. If the real-time voltage changes steadily, it indicates normal operation, and the distribution network is continuously monitored. If the real-time voltage changes upward, it indicates an anomaly, and an early warning signal is generated. The real-time voltages obtained here represent cases where there is overlap. For cases where there is overlap, early warning is issued by judging the voltage change.

[0018] This invention provides a method for predicting faults in a power distribution network. Compared with existing technologies, it has the following advantages:

[0019] This invention analyzes historical fault data, taking voltage as an example of fault causes. By classifying voltage into normal voltage and fault voltage, it further determines whether there is any overlap between the two. For cases where there is overlap, the real-time voltage changes are obtained to analyze the overall operating status of the distribution network, thereby achieving the purpose of prediction. For cases where there is no overlap, the real-time voltage value is obtained and compared with the corresponding interval value to analyze the overall operating status of the distribution network, thereby achieving the purpose of prediction. Data-driven analysis can improve the accuracy of overall prediction, promptly detect abnormalities, provide early warnings and handle them, and reduce the occurrence of faults. Attached Figure Description

[0020] Figure 1 This is a diagram of the method of the present invention;

[0021] Figure 2 This is a flowchart of the present invention;

[0022] Figure 3 This is a schematic diagram of the intersection of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Example 1, please refer to Figures 1 to 3 This application provides a method for predicting faults in a distribution network, which specifically includes the following steps:

[0025] Step 1: First, obtain the historical fault data of the distribution network, and the time period of the historical fault data is T. The historical fault data includes: number of faults, fault causes, and faulty equipment. Then, based on the fault causes, determine whether there are equipment with multiple faults and obtain the judgment result. The judgment result includes the presence result and the absence result. Specifically, equipment with multiple faults is defined as equipment that has experienced faults two or more times.

[0026] Step 2: Next, obtain the non-existent result and obtain the cause of the fault. The specific causes of the fault may include voltage, current and power. At the same time, obtain the number of faults corresponding to voltage, current and power, and select the fault cause with the most faults for analysis. In this application, the fault with voltage has the most faults for analysis.

[0027] Step 3: Obtain all voltage data from historical data, and determine the normal voltage and fault voltage based on the voltage data. Simultaneously, determine the intersection between the normal voltage and fault voltage to obtain the intersection judgment result. Then, obtain the voltage fault comparison result based on the intersection judgment result. The analysis method for voltage changes is as follows:

[0028] S1: Label all voltage data and denote them as Ui, where i = 1, 2, ..., j, and j represents the fault number. The specific value of j is determined according to the actual situation. Then, classify the voltage data into normal voltage and fault voltage according to the cause of the fault. The normal voltage is denoted as Uzk, k = 1, 2, ..., o, and the fault voltage is denoted as Ugh, h = 1, 2, ..., f. Specifically, if the cause of the fault is an abnormal voltage leading to an overall fault, the corresponding voltage is marked as a fault voltage; otherwise, if it is not caused by voltage, the corresponding voltage is marked as a normal voltage.

[0029] S2: Obtain all normal voltages Uzk and fault voltages Ugh and determine if there is any intersection between them. The intersection result includes: intersection result and no intersection result. Here, intersection means that voltages with the same value may be normal voltages or fault voltages. If both normal voltages and fault voltages have the same value, then there is an intersection; otherwise, if neither normal voltage nor fault voltage has the same value, then there is no intersection.

[0030] S3: When the intersection judgment result is that there is no intersection result, the maximum and minimum values ​​of the normal voltage are obtained, and the normal voltage range is generated using the minimum and maximum values. Similarly, the fault voltage range is obtained. The normal voltage range and fault voltage range obtained here are used to compare the generated real-time voltage to judge the overall operating status of the distribution network, thereby achieving the purpose of prediction. At the same time, no intersection also means that the range value obtained by the maximum and minimum values ​​of the normal voltage does not include any voltage value in the fault voltage, which means that the two ranges obtained are two non-interfering ranges. For example, the normal voltage range is [37, 45] and the fault voltage range is [50, 60].

[0031] S4: When the intersection judgment result indicates that an intersection exists, obtain the voltages with the same value and use them as the intersection voltages. Then, using the time when the intersection voltage is generated as the endpoint, obtain the changes in normal voltage and fault voltage within the time t before the endpoint. It is assumed that only one intersection voltage is generated here.

[0032] Based on the analysis of actual conditions, the changes in normal voltage and fault voltage are obtained by calculating the change in voltage value within a time period t. The changes include stability and increase. According to the analysis of actual conditions, the changes in normal voltage are continuously stable within a time period t, while the changes in fault voltage are increasing. Therefore, the difference between the two is used to make judgments and predictions.

[0033] Step 4: Based on the real-time data of the distribution network obtained by high-precision sensors (specifically, voltmeters), the real-time voltage of the distribution network is measured. Then, the real-time voltage is matched with the voltage comparison results to obtain the distribution network prediction results. The specific matching method is as follows:

[0034] P1: When the obtained results do not have any intersection, the real-time voltage is compared with the normal voltage range and the fault voltage range. The specific comparison method is as follows:

[0035] The maximum value of the normal voltage range and the minimum value of the fault voltage range are obtained and denoted as Uzmax and Ugmin, respectively. An early warning signal is generated when the real-time voltage is greater than Uzmax or greater than Ugmin. This means that an early warning signal is generated when either condition is met, and the real-time voltage is considered normal only when it is within the normal voltage range.

[0036] P2: When the obtained result is "existing," the real-time voltage is acquired and its change within time t is judged. If the real-time voltage changes steadily, it indicates normal operation, and the distribution network is continuously monitored. If the real-time voltage changes upward, it indicates an anomaly, and an early warning signal is generated. The real-time voltages obtained here represent cases where there is overlap. For cases where there is overlap, early warning is issued by judging the voltage change.

[0037] Example 2 is implemented based on Example 1. The difference between this example and Example 1 is that this example analyzes the result of the judgment being that the result exists.

[0038] The process involves obtaining the existing results and identifying the causes of the faults. Specific causes can include voltage, current, and power. It also retrieves faulty devices from historical fault causes, where faulty devices refer to those corresponding to the existing results, specifically those devices that exhibited the fault cause multiple times. The fault causes of these devices are then categorized to obtain classified fault causes. Taking voltage as an example from the classified fault causes, the specific analysis method is as follows:

[0039] A1: Obtain all fault voltages and denote them as U1a, where a = 1, 2, ..., b. At the same time, obtain the minimum fault voltage and denote it as U1min. Analyze the current and power corresponding to the minimum voltage. The minimum fault voltage here represents the voltage value under abnormal working conditions. Based on the actual analysis, the minimum voltage also exceeds the voltage under normal working conditions.

[0040] A2: If the current and power corresponding to the minimum voltage are normal, it indicates that the voltage is abnormal. The minimum voltage is used as the standard for judging the abnormality. At the same time, the real-time voltage is obtained and the two are compared. If the real-time voltage does not exceed U1min, it indicates that it is normal and it is continuously monitored. It does not exceed the case of U1min. Conversely, if the real-time voltage exceeds U1min, it indicates that it is abnormal and an early warning signal is generated. It exceeds the case of U1min.

[0041] A3: If the current and power corresponding to the minimum voltage are abnormal, and the existence of the abnormality indicates that either of them is abnormal, then the abnormal current and power are analyzed, and the overall status is analyzed based on the real-time current and real-time power obtained. If both are normal, continuous monitoring is carried out; if either of them is abnormal, an early warning signal is generated.

[0042] Example 3, as Example 3 of the present invention, focuses on combining the implementation processes of Example 1 and Example 2.

[0043] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0044] 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 of power distribution network fault prediction, characterized by, The method specifically includes the following steps: Step 1: Based on the number of failures, the cause of failures, and the faulty equipment, determine whether the equipment has experienced multiple failures. The determination result includes whether the failure exists or not. Step 2: Obtain the fault causes corresponding to the non-existent results, analyze the cases where the fault cause is voltage, and classify the voltage into normal voltage and fault voltage; Step 3: Determine the intersection result based on whether there is an intersection between the normal voltage and the fault voltage. The intersection result includes the result of intersection and the result of no intersection. Step 4: For results where there is no overlap, obtain the normal voltage range and the fault voltage range, and compare the real-time voltage with the two to obtain the distribution network prediction results; Step 5: For results with overlap, obtain the corresponding voltage changes for both, and simultaneously obtain the real-time voltage changes. Then, match the real-time voltage changes with the changes of both to obtain the distribution network prediction results. Step 6: Obtain the fault causes corresponding to the existing results, analyze the cases where the fault cause is voltage, obtain the real-time voltage and compare it with the voltage corresponding to the fault cause to obtain the distribution network prediction results.

2. The method for predicting faults in a distribution network according to claim 1, characterized in that, The specific analysis method for the intersection judgment result in step three is as follows: S1: Label all voltage data and denote them as Ui, where i = 1, 2, ..., j. Classify the voltage data into normal voltage Uzk and fault voltage Ugh according to the cause of the fault, where k = 1, 2, ..., o and h = 1, 2, ..., f. S2: When the normal voltage Uzk and the fault voltage Ugh have the same voltage value, an intersection result is generated; otherwise, a non-intersection result is generated.

3. The method for predicting faults in a distribution network according to claim 1, characterized in that, The method for obtaining the distribution network prediction result in step four for results where there is no intersection is as follows: P1: Obtain the maximum and minimum values ​​corresponding to the normal voltage and fault voltage, and generate the normal voltage range and fault voltage range respectively using the minimum and maximum values; P2: Obtain the maximum value of the normal voltage range Uzmax and the minimum value of the fault voltage range Ugmin. When either the real-time voltage is greater than Uzmax or the real-time voltage is greater than Ugmin, an early warning signal is generated.

4. The method for predicting faults in a distribution network according to claim 1, characterized in that, The method for obtaining the distribution network prediction result in step five, where there are overlapping results, is as follows: A1: When the intersection judgment result is that there is an intersection result, the voltage with the same value is obtained and used as the intersection voltage. Then, with the time when the intersection voltage is generated as the end point, the changes in normal voltage and fault voltage within t time before the end point are obtained. The changes are: stable and rising. A2: Obtain the real-time voltage and determine the changes within time t. If the real-time voltage changes steadily, it indicates normal operation, and the distribution network is continuously monitored. If the real-time voltage changes upward, it indicates an anomaly, and an early warning signal is generated.

5. The method for predicting faults in a distribution network according to claim 1, characterized in that, The method for analyzing the existing results to obtain the distribution network prediction results in step six is ​​as follows: The system obtains the existing results and identifies the causes of the faults, which specifically include voltage, current, and power. It also retrieves the faulty devices from historical fault causes and categorizes the fault causes of these devices to obtain classified fault causes.

6. The method for predicting faults in a distribution network according to claim 5, characterized in that, The voltage analysis method based on the classification of fault causes is as follows: All fault voltages are labeled as U1a, where a = 1, 2, ..., b. The minimum voltage among the fault voltages is obtained and labeled as U1min. The current and power corresponding to the minimum voltage are analyzed to determine whether the current and power are normal, thus identifying abnormal and normal voltages.

7. The method for predicting distribution network faults according to claim 6, characterized in that, When voltage anomalies are detected during analysis, the method for obtaining distribution network prediction results is as follows: The minimum voltage corresponding to the fault voltage is used as the anomaly judgment standard. At the same time, the real-time voltage is obtained and the two are compared. If the real-time voltage does not exceed U1min, it is considered normal and it is continuously monitored. Conversely, if the real-time voltage exceeds U1min, it is considered abnormal and an early warning signal is generated.

8. The method for predicting faults in a distribution network according to claim 6, characterized in that, The method for obtaining distribution network prediction results when the analysis shows that the voltage is normal is as follows: The system acquires abnormal current and power, and analyzes the overall status based on the acquired real-time current and power. If both are normal, continuous monitoring is performed; if either is abnormal, an early warning signal is generated.

Images