CN-122017657-A - Fault diagnosis system and method for parallel direct-current power supply of 110kV transformer substation

CN122017657ACN 122017657 ACN122017657 ACN 122017657ACN-122017657-A

Abstract

The invention provides a fault diagnosis system and method for a parallel direct current power supply of a 110kV transformer substation, and belongs to the technical field of direct current power supplies of transformer substations. The fault diagnosis system comprises a data acquisition module, a data processing module, a fault diagnosis module, a decision making module, a topology executing module and an optimizing module, wherein the data acquisition module is used for acquiring multi-dimensional data and aligning time stamps, the data processing module is used for generating a fault identification feature vector, the fault diagnosis module is used for constructing a digital twin architecture and diagnosing a model to a fault judgment result by using an attention mechanism, the decision making module is used for making a differentiated topology adjustment strategy, the topology executing module is used for controlling the action of a switching element according to a preset time sequence, and the optimizing module is used for feeding back a monitoring result to the fault diagnosis module and the decision making module to conduct closed-loop optimization and feedback iteration. The invention realizes the self-healing of single-point faults, quickly isolates fault units, reduces operation and maintenance cost and fault recovery time, accurately locates faults, prolongs the service life of a power supply and meets the differentiated power supply requirements of different types of loads.

Inventors

Xie Yuanjin
Ding Zuan
ZHOU BO
TAN LIANG
YAN YANLI
ZHANG LONG
LI JIAJUAN
Fu Yangcun
LI LIPING
ZHANG YU
DENG KAI
HU YUEMIN
LIU CHONGLI
HAO SHUAI
YAO LIN
WANG CHAO

Assignees

国网四川省电力公司宜宾供电公司

Dates

Publication Date: 20260512
Application Date: 20251202

Claims (10)

1. A fault diagnosis system for a parallel type direct current power supply of a 110kV transformer substation, characterized in that the fault diagnosis system comprises: the data acquisition module is used for acquiring multidimensional data and aligning time stamps according to the sensors deployed on the parallel type direct current power supply of the transformer substation; the data processing module is used for carrying out noise reduction and standardization processing on the multidimensional data and carrying out feature extraction to generate a fault identification feature vector; the fault diagnosis module is used for constructing a digital twin architecture according to the fault identification feature vector, analyzing the fault identification feature vector by using an attention mechanism diagnosis model, and obtaining a fault judgment result; the decision making module is used for making a differentiated topology adjustment strategy based on the fault judging result and adopting a digital twin body to simulate and verify decision security; The topology execution module is used for controlling the action of the switching element according to a preset time sequence according to a topology adjustment instruction in the topology adjustment strategy so as to reconstruct the series-parallel connection mode of the battery cells, the modules and the buses of the parallel type direct current power supply of the transformer substation; the optimization module is used for monitoring the running state and the adjusting effect of the parallel direct current power supply of the transformer substation, feeding back the monitoring result to the fault diagnosis module and the decision making module, and carrying out closed loop optimization and feedback iteration.
2. The fault diagnosis system according to claim 1, wherein the multi-dimensional data includes electrical quantity data of a cell, a parallel module, a bus and a load of the substation parallel type direct current power supply, and further includes mechanical quantity data related to a structure of the substation parallel type direct current power supply, environmental quantity data around the structure and load characteristic quantity data.
3. The fault diagnosis system according to claim 2, wherein the noise reduction and normalization process of the multi-dimensional data and feature extraction are performed to generate a fault recognition feature vector, comprising: Filtering high-frequency peak noise of the electric quantity data by adopting a differential noise reduction method, separating operation noise and fault impact signals of the mechanical quantity data, smoothing instantaneous fluctuation of the environment quantity data, and removing abnormal peaks of the load characteristic quantity data; Carrying out standardization processing on the multidimensional data after noise reduction, and carrying out standardization mapping to a unified interval to obtain preprocessed data; And extracting time domain, frequency domain and time space associated features based on the preprocessing data, and generating a fault identification feature vector.
4. The fault diagnosis system according to claim 1, wherein the constructing a digital twin architecture according to the fault identification feature vector, and analyzing the fault identification feature vector by using an attention mechanism diagnosis model, to obtain a fault determination result, includes: a digital twin architecture is adopted, a three-dimensional virtual model corresponding to the physical quantity of the parallel direct current power supply of the transformer substation is established, and the parameter synchronization of the physical quantity and the three-dimensional virtual model is realized; And identifying various typical faults in the fault identification feature vector by utilizing an attention mechanism diagnosis model and through historical fault data and simulation fault sample training, and obtaining a fault judgment result.
5. The fault diagnosis system according to claim 4, wherein the fault determination result includes: When the fault level is slight, corresponding single-point attenuation which does not affect the operation of the parallel direct current power supply of the transformer substation; When the fault level is medium, the local function of the parallel direct current power supply of the corresponding transformer substation is abnormal; And when the fault level is serious, the fault state endangering the safety of the parallel direct current power supply of the transformer substation is correspondingly achieved.
6. The fault diagnosis system according to claim 5, wherein the formulating a differentiated topology adjustment strategy based on the fault determination result comprises: constructing a multi-level decision rule based on the fault level; according to the load type in the fault judging result, formulating a topology adaptation strategy corresponding to the load type; And obtaining a differentiated topology adjustment strategy based on the multi-level decision rule and the topology adaptation strategy.
7. The fault diagnosis system of claim 6, wherein the multi-level decision rule comprises: when the fault level is slight, only adjusting the load distribution strategy, and transferring the load of the fault unit to the healthy unit, so as to avoid further attenuation of the fault unit; when the fault level is medium, triggering local topology recombination, isolating a fault module or a battery cell, and maintaining the output parameters to reach standards through serial-parallel reconstruction of the health units; when the fault level is serious, triggering global topology switching, isolating a fault area and switching to a standby topology architecture, and preferentially guaranteeing power supply of a key load.
8. The fault diagnosis system according to claim 1, wherein the controlling the switching element to operate according to a preset timing sequence according to the topology adjustment instruction in the topology adjustment policy includes: A solid-state relay is used as a switching element, a multistage flexible topology switching matrix structure is constructed, corresponding switching elements are configured for the battery cores, modules and buses of the parallel direct current power supply of the transformer substation in a sectionalized mode, and a standby switch is configured for each switching element; Generating a switch driving signal and monitoring the state of a switch element by adopting a dual-core controller, and performing topology switching control and driving; In the topology switching control and driving process, multiple protection measures are adopted to ensure the safety and uninterrupted switching process of the parallel direct current power supply of the transformer substation.
9. The fault diagnosis system according to claim 8, wherein the taking of multiple protection measures during the topology switching control and driving process comprises: Arc suppression elements are arranged at two ends of the multistage flexible topological switch matrix structure to suppress arcs generated in the switching process so as to protect the switch elements and the bus; The method comprises the steps that an energy storage element is connected in parallel with a load input end of a transformer substation parallel direct current power supply, and when instantaneous voltage fluctuation is generated by topology switching, the energy storage element rapidly discharges and compensates to maintain voltage stability of the load end; and monitoring the on-off state of each switching element, and triggering the standby switch to switch and reporting fault information if detecting that the switching elements do not act according to the topology adjustment instruction.
10. The fault diagnosis method for the parallel direct current power supply of the 110kV transformer substation is characterized by comprising the following steps of: according to a sensor deployed on a parallel type DC power supply of a transformer substation, acquiring multidimensional data and aligning a time stamp; The multi-dimensional data is subjected to noise reduction and standardization processing, and feature extraction is carried out to generate a fault identification feature vector; Constructing a digital twin architecture according to the fault identification feature vector, and analyzing the fault identification feature vector by using an attention mechanism diagnosis model to obtain a fault judgment result; Based on the fault judging result, a differentiated topology adjustment strategy is formulated, and digital twin is adopted to simulate and verify decision security; According to a topology adjustment instruction in the topology adjustment strategy, controlling the action of a switching element according to a preset time sequence so as to reconstruct the series-parallel connection mode of a battery core, a module and a bus of the parallel type direct current power supply of the transformer substation; And monitoring the running state and the adjusting effect of the parallel direct current power supply of the transformer substation, feeding back the monitoring result to the fault diagnosis module and the decision making module, and performing closed loop optimization and feedback iteration.

Description

Fault diagnosis system and method for parallel direct-current power supply of 110kV transformer substation Technical Field The invention relates to the technical field of direct current power supplies of substations, in particular to a fault diagnosis system and method for a parallel direct current power supply of a 110kV substation. Background The direct current power supply system is burdened with the task of providing a stable operation power supply for secondary and communication equipment, breaker opening and closing, signal loops, UPS systems and other equipment in the transformer substation, and the safety and reliability of the direct current power supply system are directly related to the stable operation of a power grid, so that the direct current power supply is also visually called as a heart of the transformer substation. In the prior art, the conventional parallel direct current power supply of the transformer substation often has the fatal defect of system outage caused by single-point faults, and the transformer is damaged or the transformer substation is completely stopped due to the fault of the direct current system, so that great economic loss is caused; in addition, the traditional transformer substation adopts a one-cut power supply mode, so that the differential requirements of different types of loads in the transformer substation cannot be adapted, and the power supply reliability and the economy are inconvenient to consider. Disclosure of Invention The invention aims to provide a fault diagnosis system and method for a parallel direct current power supply of a 110kV transformer substation, which are used for solving the problems that in the prior art, single-point faults of the parallel direct current power supply of the transformer substation cause system outage, single-electric quantity monitoring has high false alarm rate and low positioning accuracy, and the system cannot adapt to the differential requirements of different types of loads. The invention provides a fault diagnosis system of a 110kV transformer substation parallel type direct current power supply, which comprises a data acquisition module, a data processing module, an optimization module, a decision making module, a topology execution module and a closed loop iteration and feedback optimization module, wherein the data acquisition module is used for acquiring multidimensional data and aligning a time stamp according to a sensor deployed on the transformer substation parallel type direct current power supply, the data processing module is used for denoising and standardizing the multidimensional data and extracting features to generate a fault identification feature vector, the fault diagnosis module is used for constructing a digital twin architecture according to the fault identification feature vector and analyzing the fault identification feature vector by using an attention mechanism diagnosis model to obtain a fault judgment result, the decision making module is used for making a differential topology adjustment strategy and adopting a digital twin body to simulate and verify decision security, the topology execution module is used for controlling actions of a switching element according to a topology adjustment instruction in the topology adjustment strategy so as to realize serial-parallel mode reconstruction of a cell, a module and a bus of the transformer substation parallel type direct current power supply, and the optimization module is used for monitoring the running state and adjusting the parallel type direct current power supply and feeding back the fault judgment result to the fault diagnosis module. Optionally, the multidimensional data comprise electrical quantity data of a battery core, a parallel module, a bus and a load of the transformer substation parallel direct-current power supply, and also comprise mechanical quantity data related to a transformer substation parallel direct-current power supply structure, environmental quantity data around the structure and load characteristic quantity data. The method comprises the steps of filtering high-frequency peak noise of electric quantity data by a differential noise reduction method, separating operation noise and fault impact signals of mechanical quantity data, smoothing instantaneous fluctuation of environment quantity data, removing abnormal peaks of load characteristic quantity data, carrying out standardization processing on the multidimensional data after noise reduction, carrying out normalized mapping to a unified interval to obtain preprocessing data, and extracting time domain, frequency domain and time space correlation characteristics based on the preprocessing data to generate the fault recognition characteristic vector. Optionally, constructing a digital twin architecture according to the fault identification feature vector, analyzing the fault identification feature vector by using an attention mechanism diagnosis model to obtai