CN-122017401-A - Power distribution cabinet fault positioning method and system

CN122017401ACN 122017401 ACN122017401 ACN 122017401ACN-122017401-A

Abstract

The invention relates to the technical field of power distribution cabinets, in particular to a power distribution cabinet fault positioning method and system. The method comprises the steps of synchronously collecting sensing data of a plurality of monitoring points deployed in a power distribution cabinet according to an electrical topological relation to generate a multi-source monitoring sequence, extracting and fusing target features from the multi-source monitoring sequence based on feature templates corresponding to monitoring point association components to generate a comprehensive feature vector, dynamically fusing and analyzing the comprehensive feature vector and current operation working condition parameters through a preset diagnosis model to generate a diagnosis data set, calculating the probability that each component is a fault source by combining a fault positioning knowledge graph corresponding to the power distribution cabinet with the diagnosis data set to generate a component fault probability data set, and judging the component exceeding a preset fault threshold as a fault component and generating corresponding fault alarm data. According to the invention, through collaborative innovation of multi-source data synchronous acquisition, feature template accurate fusion, working condition dynamic adaptation analysis and knowledge graph auxiliary positioning, the accuracy and reliability of fault positioning are effectively improved.

Inventors

ZHONG ZEPENG
Ye Qiongjin
WU XIZHENG

Assignees

海南美亚电能有限公司

Dates

Publication Date: 20260512
Application Date: 20260116

Claims (10)

1. The utility model provides a power distribution cabinet fault location method which is characterized in that the method comprises the following steps: Synchronously collecting sensing data of a plurality of monitoring points deployed in the power distribution cabinet according to an electrical topological relation to generate a multi-source monitoring sequence; Extracting and fusing target features from the multi-source monitoring sequence based on a feature template corresponding to the monitoring point association component to generate a comprehensive feature vector; Carrying out dynamic fusion analysis by adopting the comprehensive feature vector and the current operation condition parameters through a preset diagnosis model to generate a diagnosis data set; based on the fault positioning knowledge graph corresponding to the power distribution cabinet and the diagnosis data set, calculating the probability that each part in the power distribution cabinet is a fault source, and generating a part fault probability data set; And judging the parts exceeding the preset fault threshold value in the part fault probability data set as fault parts, and generating fault alarm data corresponding to the fault parts.
2. The method for locating a fault in a power distribution cabinet according to claim 1, wherein the step of synchronously collecting sensing data of a plurality of monitoring points deployed in the power distribution cabinet according to an electrical topological relation to generate a multi-source monitoring sequence comprises the steps of: Based on an electrical wiring diagram of the power distribution cabinet, taking physical nodes representing key electrical connection states as monitoring points; And synchronously acquiring temperature, partial discharge and vibration data at the monitoring points to generate a time-aligned multi-source monitoring sequence.
3. The method for locating a fault in a power distribution cabinet according to claim 1, wherein the step of extracting and fusing target features from the multi-source monitoring sequence based on the feature templates corresponding to the monitoring point association components to generate a comprehensive feature vector comprises the steps of: matching key stress types of all physical components in the power distribution cabinet from a preset component stress rule base according to the electrical topological structure of the power distribution cabinet, and determining a typical fault mode associated with the key stress types based on a historical fault database; Extracting the characteristics corresponding to each typical fault mode from the multi-source monitoring sequence based on a preset fault mechanism and characteristic mapping relation table, and generating target characteristics corresponding to the typical fault modes; Based on the occurrence probability corresponding to the typical fault mode in the historical fault database, calculating a fusion weight coefficient of the target feature by combining the historical characterization contribution degree corresponding to the target feature; And weighting calculation is carried out by adopting all the target features and the corresponding fusion weight coefficients, so as to generate a comprehensive feature vector.
4. The method for positioning a fault of a power distribution cabinet according to claim 1, wherein the preset diagnosis model includes a plurality of sub-diagnosis models optimized for different fault types or different component areas, respectively, and the step of generating a diagnosis data set by performing dynamic fusion analysis on the preset diagnosis model by using the integrated feature vector and the current operation condition parameters includes: inputting current operation condition parameters into a pre-trained weight distribution function, and calculating to obtain dynamic decision weights of the sub-diagnostic models; Forward reasoning is carried out on the comprehensive feature vector through each sub-diagnosis model, and candidate fault types, component areas and initial confidence degrees corresponding to the sub-diagnosis models are generated; Multiplying the initial confidence coefficient with a corresponding dynamic decision weight respectively to generate a weighted confidence coefficient; probability normalization is carried out on the weighted confidence coefficient by adopting a softmax function, and normalized confidence probability is generated; Selecting the normalized confidence probability higher than a preset report threshold, and generating a target confidence probability; and constructing a diagnosis data set by adopting the candidate fault type, the component area and the target confidence probability corresponding to the target confidence probability.
5. The method for positioning faults of a power distribution cabinet according to claim 1, wherein the step of calculating the probability that each component in the power distribution cabinet is a fault source based on the fault positioning knowledge graph corresponding to the power distribution cabinet and the diagnosis data set and generating a component fault probability data set comprises the following steps: each physical component in the power distribution cabinet is taken as a node, an electrical connection relation among the components is taken as an edge, prior fault probability based on historical statistics is configured for each node, conditional probability based on historical fault propagation statistics is configured for each edge, and a fault positioning knowledge graph is constructed and obtained; mapping the diagnosis data set as observation evidence to corresponding nodes and fault types in the knowledge graph respectively; Based on the observation evidence, the prior fault probability of each node and the conditional probability of each side, carrying out iterative probability propagation calculation on the topological structure of the knowledge graph, and updating to obtain posterior probability of each node serving as a fault source; And constructing a component fault probability data set by adopting each node and the updated posterior probability thereof.
6. The method for locating a fault in a power distribution cabinet according to claim 5, wherein the step of determining the component exceeding the preset fault threshold in the component fault probability data set as a faulty component and generating fault alarm data corresponding to the faulty component includes: Selecting a part with posterior probability exceeding a first preset threshold value in the part fault probability data set as a first-level fault part; selecting a part with posterior probability exceeding a second preset threshold value but not exceeding the first preset threshold value in the part fault probability data set as a secondary fault part, wherein the first preset threshold value is higher than the second preset threshold value; Extracting typical fault reasons and maintenance measures of similar components corresponding to the primary fault components from a historical maintenance database, and generating high-priority alarm information; Determining a fault probability change trend feature corresponding to the secondary fault component based on component fault probability data generated by the secondary fault component in a preset history period; Formatting the fault probability change trend characteristics to generate trend research and judgment conclusion comprising trend description and risk level; Adopting the trend research conclusion, the component identification corresponding to the secondary fault component and the current fault probability thereof to construct observation level early warning information; and summarizing the high-priority alarm information and the observation level early-warning information to generate fault alarm data.
7. A power distribution cabinet fault location system, comprising: the multi-source synchronous acquisition module is used for synchronously acquiring sensing data of a plurality of monitoring points deployed in the power distribution cabinet according to the electrical topological relation to generate a multi-source monitoring sequence; The mechanism feature fusion module is used for extracting and fusing target features from the multi-source monitoring sequence based on the feature templates corresponding to the monitoring point association components to generate comprehensive feature vectors; the working condition self-adaptive diagnosis module is used for carrying out dynamic fusion analysis by adopting the comprehensive feature vector and the current operation working condition parameters through a preset diagnosis model to generate a diagnosis data set; The probability positioning reasoning module is used for calculating the probability of each component in the power distribution cabinet as a fault source based on the fault positioning knowledge graph corresponding to the power distribution cabinet and the diagnosis data set, and generating a component fault probability data set; and the hierarchical intelligent alarm module is used for judging the parts exceeding the preset fault threshold value in the part fault probability data set as fault parts and generating fault alarm data corresponding to the fault parts.
8. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of the method for locating a fault in a power distribution cabinet as claimed in any one of claims 1 to 6.
9. A computer readable storage medium having stored thereon a computer program, wherein the computer program when executed implements the power distribution cabinet fault localization method of any of claims 1-6.
10. A computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions, wherein the program instructions, when executed by a computer, cause the computer to perform the method of fault localization of a power distribution cabinet as claimed in any one of claims 1 to 6.

Description

Power distribution cabinet fault positioning method and system Technical Field The invention relates to the technical field of power distribution cabinets, in particular to a power distribution cabinet fault positioning method and system. Background Along with the continuous expansion of the scale of the urban power grid and the continuous improvement of the intelligent level, the power distribution cabinet is used as a key node for electric energy distribution and control, and the operation reliability of the power distribution cabinet directly influences the power supply quality. At present, the power distribution cabinet is widely deployed in various substations, industrial and mining enterprises and commercial buildings, and has large quantity and wide distribution. In order to ensure the power supply continuity, the internal faults of the power distribution cabinet are rapidly and accurately positioned, and the method is a key for realizing timely early warning and active operation and maintenance. The existing fault positioning method mainly relies on that a single sensor, such as a temperature sensor or a partial discharge sensor, is arranged at a key position in a cabinet, and specific physical quantity data is collected and uploaded to a monitoring center. When the data exceeds a preset fixed threshold, the system triggers an alarm and generally indicates an abnormal region. However, in actual operation, the above positioning method has a significant defect, resulting in lower accuracy of fault positioning. Firstly, it is difficult to comprehensively and early capture composite faults or hidden faults depending on single characteristic parameters, and erroneous judgment or missed judgment is easy to cause. And secondly, the fixed static threshold cannot be adapted to the aging curve, load fluctuation and environmental change of the equipment, and the alarm sensitivity and accuracy are difficult to be compatible under the complex working condition. And the existing method lacks of intelligent analysis on the relevance of the fault type and the position, generally only can position the abnormality to a certain cabinet body or a rough section, and cannot accurately position the minimum replaceable units such as a specific breaker, a busbar connection point or a cable terminal and the like, so that difficulty is brought to the field investigation of operation and maintenance personnel, and the power failure time is prolonged. Disclosure of Invention The invention provides a power distribution cabinet fault positioning method and system, which solve the technical problem that the existing power distribution cabinet fault positioning method lacks comprehensive sensing and intelligent analysis capabilities for multidimensional time-varying fault characteristics, so that the positioning accuracy is low. The invention provides a fault positioning method for a power distribution cabinet, which comprises the following steps: Synchronously collecting sensing data of a plurality of monitoring points deployed in the power distribution cabinet according to an electrical topological relation to generate a multi-source monitoring sequence; Extracting and fusing target features from the multi-source monitoring sequence based on a feature template corresponding to the monitoring point association component to generate a comprehensive feature vector; Carrying out dynamic fusion analysis by adopting the comprehensive feature vector and the current operation condition parameters through a preset diagnosis model to generate a diagnosis data set; based on the fault positioning knowledge graph corresponding to the power distribution cabinet and the diagnosis data set, calculating the probability that each part in the power distribution cabinet is a fault source, and generating a part fault probability data set; And judging the parts exceeding the preset fault threshold value in the part fault probability data set as fault parts, and generating fault alarm data corresponding to the fault parts. Optionally, the step of synchronously collecting sensing data of a plurality of monitoring points deployed in the power distribution cabinet according to an electrical topological relation and generating a multi-source monitoring sequence includes: Based on an electrical wiring diagram of the power distribution cabinet, taking physical nodes representing key electrical connection states as monitoring points; And synchronously acquiring temperature, partial discharge and vibration data at the monitoring points to generate a time-aligned multi-source monitoring sequence. Optionally, the step of extracting and fusing target features from the multi-source monitoring sequence based on the feature templates corresponding to the monitoring point association components to generate a comprehensive feature vector includes: matching key stress types of all physical components in the power distribution cabinet from a preset component stress rule