Correlation analysis method for defect and failure risk conversion of power distribution equipment

Abstract

A method for analyzing the conversion association between defects and fault risks of power distribution equipment is proposed. According to the corresponding relationship between historical defect discovery time and fault discovery time and the corresponding relationship between historical defective equipment and faulty equipment, an association rule between defects and faults is established so that the defect detail data has corresponding historical fault data, thereby forming a defect and fault association library. The present invention uses an improved FP-growth algorithm to perform association mining on historical defect and fault text records, finds the correlation between defects and faults, screens out defects that are likely to cause faults, solves the problem of predicting faults caused by defects, and provides a new idea for subsequent operation and maintenance of power distribution equipment.

Description

Technical Field

[0001] The present invention relates to a technology in the field of power distribution network control, and specifically to a method for analyzing the conversion association between defects and failure risks of power distribution equipment. Background Art

[0002] In the operation and maintenance management of power distribution equipment, a large amount of text data is recorded, including inspection information, defect records, fault records, and fault elimination reports. Currently, in the process of mining association relationships, most of the historical text records of defects and faults are analyzed based on a manually set minimum support, and data greater than this support is considered a frequent item. However, in the actual application of power distribution equipment, items appear at different frequencies and have different importance, so it is unreasonable to use a unique support setting. Existing association rule mining algorithms are mostly based on the Aprori algorithm and the FP-growth algorithm, and cannot be directly used for defect and fault mining of power distribution equipment. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention proposes a method for analyzing the conversion association between defects and fault risks of distribution equipment. The improved FP-growth algorithm is used to perform association mining on historical defect and fault text records, find the correlation between defects and faults, screen out defects that are likely to cause faults, solve the problem of predicting faults caused by defects, and provide new ideas for the subsequent operation and maintenance of distribution equipment.

[0004] The present invention is achieved through the following technical solutions:

[0005] The present invention relates to a method for analyzing the conversion association between defects and fault risks of power distribution equipment. According to the correspondence between the historical defect discovery time and the fault discovery time and the correspondence between the historical defect equipment and the fault equipment, by establishing the association rules between defects and faults, the defect detail data has the corresponding historical fault data, and then a defect and fault association library is formed, which specifically includes:

[0006] Step 1: Extract defect feature information and fault feature information from the original defect text database and the fault text database. Obtain the defect information required for association, including defect device ID, defect discovery time, defect type, and defect level; obtain the fault information required for association, including fault device ID, fault discovery time, and fault type.

[0007] Step 2: Based on the equipment and occurrence time of defects and faults, data preprocessing is performed to obtain detailed information on defects and faults that have occurred on the same equipment, forming a historical defect fault library D. The occurrence time of defects and faults is compared, requiring that the defect occurrence time is earlier than the fault occurrence time, and considering the defect level, different defect levels are set with different defect and fault association time differences. Defects and fault contents that meet the conditions are retained.

[0008] Step 3: Use the FP-growth algorithm to mine the associations between items in the historical defect fault database D. Set the support and confidence indicators, mine frequent item sets from the input database through FP-tree, select items that meet the minimum support and confidence, establish association rules, and set different minimum supports for different defect levels.

[0009] Step 4: Complete the correlation analysis of defects and faults of distribution equipment through association rules to form a defect and fault association library.

[0010] Technical Effects

[0011] The present invention uses the association algorithm of data mining to find the intrinsic connection between defects and faults. Taking into account that different defect levels have different elimination times, different association time differences are set when defects are associated with faults. Taking into account that different types of defects have different importance in the entire network, they cannot be simply regarded as equal relationships during the mining process, so different support levels are set for different defect types to conduct in-depth mining of important information. Compared with the prior art, the present invention can easily realize the automatic generation and dynamic update of defects and faults, and can adapt to the complex and diverse defects and faults of today's distribution equipment and the requirements of dynamic changes. And it can provide defect processing information for distribution network operators and enhance the safety and reliability of the distribution network. By mining the defect and fault text data of distribution equipment, fault warnings can be made for real-time defects based on the defect and fault association library, and grid staff can be guided to eliminate defects that are prone to cause faults in a timely manner, prevent problems before they occur, and effectively improve the power supply reliability of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a flow chart of the present invention;

[0013] Figure 2 It is a flow chart of the association between defects and failures of power distribution equipment in the embodiment;

[0014] Figure 3 The present invention is a flow chart of preprocessing the defects and fault association of power distribution equipment in an embodiment. DETAILED DESCRIPTION

[0015] like Figure 1As shown, this embodiment relates to a method for analyzing the correlation between defects and failure risks of power distribution equipment, including:

[0016] Step 1: Extract defect feature information and fault feature information from the original defect text database and the fault text database to obtain the defect device ID, defect discovery time, defect type and defect level;

[0017] The original defect text database contains a large amount of redundant information, which needs to be screened and processed before the defects and faults are associated. The results after the screening process before the final feature information extraction are shown in Table 1.

[0018] The screening process described herein refers to: performing redundant data cleaning, that is, identifying and removing invalid or missing information from the original data to ensure the accuracy and effectiveness of subsequent analysis.

[0019] The association processing refers to identifying and establishing the relationship between defects and faults according to the designed FP-growth algorithm.

[0020] Table 1 Feature extraction of defect and fault text data

[0021]

[0022]

[0023] Step 2: Based on the equipment and occurrence time of defects and faults, detailed information on defects and faults that have occurred on the same equipment is obtained through data preprocessing to form a historical defect and fault library.

[0024] The data preprocessing refers to: extracting the defective device ID in the defect text data and the faulty device ID in the fault text data, and determining the defects and faults that have occurred on a distribution device based on the unique distribution equipment ID; comparing the occurrence time of the defects and faults, and screening out the defects that occur earlier than the fault time; and considering the defect level, setting different defect and fault association time differences for different defect levels, that is, the difference between the corresponding times associated with the defect data items and the fault data items, retaining the defects and fault contents that meet the conditions, and forming a historical defect and fault library.

[0025] Step 3: Figure 2 As shown, the FP-growth algorithm is used to mine the association relationship between the items in the historical defect fault library, specifically including: setting support and confidence indicators, mining frequent item sets for the input database through FP-tree, and screening out items that meet the minimum support and confidence to establish association rules.

[0026] The association rule, i.e., an implication expression in the form of X→Y, where X is the antecedent and Y is the consequence, specifically includes:

[0027] 1) The ratio of the number of transactions containing the itemset to the total number of transactions, i.e. the support of the itemset Where: Σ(X,Y) is the number of times X and Y appear at the same time; n is the total number of transactions in the database.

[0028] The minimum support refers to: different defect levels correspond to different minimum support, as shown in Table 2.

[0029] Table 2 Minimum support setting

[0030]

[0031] In this embodiment, the minimum confidence level is set to 0.6.

[0032] 2) The probability that Y exists when a given event contains X, i.e., the confidence level Where: Σ(X) is the total number of times transaction X occurs.

[0033] Step 4: Complete the correlation analysis of defects and faults of the distribution equipment through association rules to form a defect and fault association library.

[0034] After specific practical experiments, based on the actual defect text data and fault text data of the municipal power grid, feature information was extracted from the original data text, and the following defect feature information and fault feature information were extracted, as shown in Table 3 and Table 4 for some examples. Then use "DEVICE_ID" to determine the defects and faults that have occurred on the same device. If it is the same device, further compare the occurrence time of the defect and the fault, requiring the defect to occur earlier than the fault. After forming the historical defect and fault association library shown in Table 5, the FP-growth algorithm is used for association analysis. Frequent items are selected according to the number of times the defect and the fault occur at the same time. When the defect and the fault are associated multiple times and meet the set support and confidence, it is considered that this defect may cause this fault. Finally, the association result is output and the defect and fault association library shown in Figure 6 is established.

[0035] Table 3 Defect characteristic information

[0036]

[0037] Table 4 Fault characteristic information

[0038]

[0039] Table 5 Historical defect fault association library D

[0040]

[0041]

[0042] Table 6 Defect and fault association library

[0043] Defect type name => possible failure Insulator damage => medium voltage insulator cracking Arrester damage => medium voltage arrester breakdown Distribution transformer oil leakage => medium voltage oil immersed transformer oil leakage

[0044] Compared with the prior art, the present invention sets multiple support degrees according to the actual situation of the defects of the power distribution equipment, which is more in line with the actual situation. Taking full account of the actual situation of the power distribution equipment, multiple association conditions are set in the process of associating defects and faults to ensure the rationality and accuracy of the association results. It is not only applicable to power distribution equipment, but also to the defect and fault association analysis of other equipment.

[0045] The above-mentioned specific implementation can be partially adjusted in different ways by those skilled in the art without departing from the principle and purpose of the present invention. The protection scope of the present invention shall be based on the claims and shall not be limited by the above-mentioned specific implementation. Each implementation scheme within its scope shall be subject to the constraints of the present invention.

Claims

1. A method for analyzing the conversion association between defects and failure risks of power distribution equipment, characterized in that: According to the correspondence between the historical defect discovery time and the fault discovery time, and the correspondence between the historical defect equipment and the fault equipment, by establishing the association rules between defects and faults, the defect detail data has the corresponding historical fault data, thereby forming a defect and fault association library; The analysis method specifically comprises: Step 1: Extract defect feature information and fault feature information from the original defect text database and the fault text database to obtain the defect device ID, defect discovery time, defect type and defect level; Step 2: Based on the equipment and occurrence time of defects and faults, detailed information of defects and faults that have occurred on the same equipment is obtained through data preprocessing to form a historical defect and fault library; Step 3: Use the FP-growth algorithm to mine the associations between items in the historical defect fault database, specifically including: setting support and confidence indicators, mining frequent item sets from the input database through FP-tree, and selecting items that meet the minimum support and confidence to establish association rules; Step 4: Complete the correlation analysis of defects and faults of the power distribution equipment through association rules to form a defect and fault association library; The data preprocessing refers to: extracting the defective device ID in the defect text data and the faulty device ID in the fault text data, and determining the defects and faults that have occurred on a power distribution device based on the unique power distribution device ID; comparing the occurrence time of the defects and faults, and screening out the defects that occurred earlier than the faults; and considering the defect level, setting different defect and fault association time differences for different defect levels, that is, the difference between the corresponding time associated with the defect data item and the fault data item, retaining the defect and fault content that meets the conditions, and forming a historical defect and fault library; The association rule, i.e., an implication expression in the form of X→Y, where X is the antecedent and Y is the consequence, specifically includes: 1) The ratio of the number of transactions containing the itemset to the total number of transactions, i.e. the support of the itemset ,in: is the number of times X and Y appear at the same time; n is the total number of transactions in the database; 2) The probability that Y exists when a given event contains X, i.e., the confidence level ,in: is the total number of occurrences of transaction X; The minimum support levels include: support levels less than or equal to 1.5% for emergency, support levels greater than 1.5% and less than or equal to 5% for major, and support levels greater than 5% and less than or equal to 10% for general.

2. The method for analyzing the conversion association between defects and failure risks of power distribution equipment according to claim 1 is characterized in that: The original defect text database is processed after screening, and then the defects and faults are associated; The screening process mentioned above refers to: cleaning redundant data, that is, identifying and removing invalid or missing information from the original data to ensure the accuracy and effectiveness of subsequent analysis; The association processing refers to identifying and establishing the relationship between defects and faults according to the designed FP-growth algorithm.

Images