Fault diagnosis system and method for parallel DC power supply of 110kV transformer substation

Abstract

The invention provides a fault diagnosis system and method for a parallel DC power supply of a 110kV transformer substation, and belongs to the technical field of DC power supplies of transformer substations. The fault diagnosis system comprises a data acquisition module used for acquiring multi-dimensional data and aligning timestamps; the data processing module is used for generating a fault identification feature vector; the fault diagnosis module is used for constructing a digital twinborn architecture and diagnosing a model by using an attention mechanism to obtain a fault judgment result; the decision making module is used for making differentiated topology adjustment strategies; the topology execution module is used for controlling the switch elements to act according to a preset time sequence; and the optimization module is used for feeding back the monitoring result to the fault diagnosis module and the decision making module for closed-loop optimization and feedback iteration. According to the invention, single-point fault self-healing is realized, fault units are quickly isolated, the operation and maintenance cost and fault recovery time are reduced, the fault is accurately positioned, the service life of the power supply is prolonged, and the differentiated power supply requirements of different types of loads are met.

Description

Technical Field

[0001] This invention relates to the field of DC power supply technology for substations, and specifically to a fault diagnosis system and method for a parallel DC power supply in a 110kV substation. Background Technology

[0002] DC power supply systems are responsible for providing stable power to secondary and communication equipment, circuit breaker opening and closing, signal circuits, UPS systems and other equipment in substations. Their safety and reliability are directly related to the stable operation of the power grid. Therefore, DC power supply is figuratively called the "heart" of substations.

[0003] In existing technologies, traditional substations connected to parallel DC power supplies often suffer from the fatal flaw of single-point faults leading to system outages. Transformer damage or complete substation shutdowns caused by DC system faults result in significant economic losses. Furthermore, traditional substations use a single electrical quantity monitoring method, which easily leads to high false alarm rates and low location accuracy in fault diagnosis, causing safety hazards and affecting equipment lifespan. In addition, traditional substations use a one-size-fits-all power supply mode, which cannot adapt to the differentiated needs of different types of loads within the substation, making it difficult to balance power supply reliability and economy. Summary of the Invention

[0004] The purpose of this invention is to provide a fault diagnosis system and method for parallel DC power supplies in 110kV substations, in order to solve the problems in the prior art where single-point faults in parallel DC power supplies in substations lead to system outages, high false alarm rates and low positioning accuracy of single electrical quantity monitoring, and inability to adapt to the differentiated needs of different types of loads.

[0005] To achieve the above objectives, this invention provides a fault diagnosis system for a 110kV substation parallel DC power supply. The fault diagnosis system includes: a data acquisition module for collecting multi-dimensional data and aligning timestamps based on sensors deployed on the parallel DC power supply in the substation; a data processing module for denoising and standardizing the multi-dimensional data, extracting features, and generating a fault identification feature vector; a fault diagnosis module for constructing a digital twin architecture based on the fault identification feature vector, and using an attention mechanism diagnostic model to analyze the fault identification feature vector to obtain a fault determination result; a decision-making module for formulating differentiated topology adjustment strategies based on the fault determination result, and using a digital twin for simulation verification of decision security; a topology execution module for controlling the operation of switching elements according to a preset timing sequence based on the topology adjustment instructions in the topology adjustment strategy, thereby reconstructing the series-parallel connection of the cells, modules, and busbars of the parallel DC power supply in the substation; and an optimization module for monitoring the operating status and adjustment effect of the parallel DC power supply in the substation, and feeding back the monitoring results to the fault diagnosis module and the decision-making module for closed-loop optimization and feedback iteration.

[0006] Optionally, the multi-dimensional data includes electrical quantity data of the cells, parallel modules, busbars and loads of the parallel DC power supply in the substation, as well as mechanical quantity data related to the structure of the parallel DC power supply in the substation, environmental quantity data around the structure and load characteristic quantity data.

[0007] Optionally, the step of denoising and standardizing the multi-dimensional data and extracting features to generate a fault identification feature vector includes: using a differentiated denoising method to filter high-frequency spike noise in the electrical quantity data, separating the operating noise and fault impact signal in the mechanical quantity data, smoothing the instantaneous fluctuations in the environmental quantity data, and removing abnormal spikes in the load characteristic quantity data; standardizing the denoised multi-dimensional data and normalizing it to a unified interval to obtain preprocessed data; and extracting time-domain, frequency-domain, and spatiotemporal correlation features based on the preprocessed data to generate a fault identification feature vector.

[0008] Optionally, the step of constructing a digital twin architecture based on the fault identification feature vector and analyzing the fault identification feature vector using an attention mechanism diagnostic model to obtain a fault determination result includes: using a digital twin architecture to establish a three-dimensional virtual model corresponding to the physical quantities of the parallel DC power supply in the substation, and synchronizing the physical quantities with the parameters of the three-dimensional virtual model; using an attention mechanism diagnostic model, through training with historical fault data and simulated fault samples, identifying various typical faults in the fault identification feature vector, and obtaining a fault determination result.

[0009] Optionally, the fault determination result includes: when the fault level is minor, it corresponds to a single-point attenuation that does not affect the operation of the parallel DC power supply in the substation; when the fault level is medium, it corresponds to a local functional abnormality of the parallel DC power supply in the substation; when the fault level is severe, it corresponds to a fault state that endangers the safety of the parallel DC power supply in the substation.

[0010] Optionally, the step of formulating a differentiated topology adjustment strategy based on the fault determination result includes: constructing multi-level decision rules based on the fault level; formulating a topology adaptation strategy corresponding to the load type according to the load type in the fault determination result; and obtaining a differentiated topology adjustment strategy based on the multi-level decision rules and the topology adaptation strategy.

[0011] Optionally, the multi-level decision rules include: when the fault level is minor, only the load distribution strategy is adjusted to transfer the load of the faulty unit to the healthy unit to avoid further degradation of the faulty unit; when the fault level is moderate, local topology reconfiguration is triggered to isolate the faulty module or cell and maintain the output parameters in compliance through the series-parallel reconstruction of the healthy unit; when the fault level is severe, global topology switching is triggered to isolate the faulty area and switch to the backup topology architecture to prioritize the power supply of critical loads.

[0012] Optionally, the step of controlling the operation of switching elements according to a preset timing sequence based on the topology adjustment instructions in the topology adjustment strategy includes: using solid-state relays as switching elements to construct a multi-level flexible topology switch matrix structure, configuring corresponding switching elements for the cells, modules, and bus sections of the parallel DC power supply in the substation, and configuring a spare switch for each switching element; using a dual-core controller to generate switch drive signals and monitor the status of switching elements to perform topology switching control and drive; and taking multiple protection measures during the topology switching control and drive process to ensure safe and uninterrupted switching of the parallel DC power supply in the substation.

[0013] Optionally, multiple protection measures are adopted during the topology switching control and drive process, including: configuring arc suppression elements at both ends of the multi-level flexible topology switch matrix structure to suppress the arc generated during the switching process, so as to protect the switch elements and the bus; connecting energy storage elements in parallel at the load input end of the parallel DC power supply in the substation, so that when the topology switching generates instantaneous voltage fluctuations, the energy storage elements quickly discharge to compensate, so as to maintain the stability of the load end voltage; monitoring the on / off status of each switch element, and if it is detected that the switch element does not act according to the topology adjustment command, the backup switch is triggered and the fault information is reported.

[0014] On the other hand, this invention provides a fault diagnosis method for a 110kV substation parallel DC power supply. The fault diagnosis method includes: collecting multi-dimensional data and aligning timestamps based on sensors deployed on the parallel DC power supply in the substation; denoising and standardizing the multi-dimensional data, and extracting features to generate a fault identification feature vector; constructing a digital twin architecture based on the fault identification feature vector, and using an attention mechanism diagnostic model to analyze the fault identification feature vector to obtain a fault determination result; formulating a differentiated topology adjustment strategy based on the fault determination result, and using a digital twin to simulate and verify the safety of the decision; controlling the operation of switching elements according to a preset timing sequence based on the topology adjustment instructions in the topology adjustment strategy to reconstruct the series-parallel connection of the cells, modules, and buses of the parallel DC power supply in the substation; monitoring the operating status and adjustment effect of the parallel DC power supply in the substation, and feeding the monitoring results back to the fault diagnosis module and the decision-making module for closed-loop optimization and feedback iteration.

[0015] The above technical solutions, through the construction of a multi-level flexible topology switch matrix and the formulation of topology adjustment strategies, solve the pain point of traditional parallel DC power supplies in substations where a single failure leads to total loss. This achieves 100% self-healing of single-point faults, enabling rapid isolation of faulty units and reorganization of healthy units, ensuring uninterrupted power supply to critical loads, and reducing operation and maintenance costs and fault recovery time. Furthermore, by constructing a digital twin and attention mechanism diagnostic model, fault diagnosis accuracy is improved, enabling precise fault location, early fault warning capabilities, and extending the overall lifespan of the power supply. Finally, by establishing dynamic adaptation rules between load demand and topology, the system meets the differentiated power supply requirements of different types of loads, avoiding equipment damage or performance degradation due to improper load adaptation, optimizing system energy distribution efficiency, and balancing power supply reliability and economy.

[0016] Other features and advantages of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating the fault diagnosis system for a parallel DC power supply in a 110kV substation according to the present invention. Figure 1 ; Figure 2 This is a schematic diagram of the process for generating fault identification feature vectors in this invention; Figure 3 This is a schematic diagram of the process for obtaining fault determination results in this invention; Figure 4This is a flowchart illustrating the process of formulating differentiated topology adjustment strategies in this invention; Figure 5 This is a flowchart illustrating the process of controlling the operation of switching elements according to a preset timing sequence in this invention. Figure 6 This is a flowchart illustrating the multiple protection measures employed in this invention; Figure 7 This is a flowchart illustrating the fault diagnosis system for a parallel DC power supply in a 110kV substation according to the present invention. Figure 2 . Detailed Implementation

[0018] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0019] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0020] Please refer to Figure 1 This invention provides a fault diagnosis system for a 110kV substation parallel DC power supply. The system includes: a data acquisition module for collecting multi-dimensional data and aligning timestamps based on sensors deployed on the parallel DC power supply; a data processing module for denoising and standardizing the multi-dimensional data, extracting features, and generating fault identification feature vectors; a fault diagnosis module for constructing a digital twin architecture based on the fault identification feature vectors and using an attention mechanism diagnostic model to analyze the fault identification feature vectors and obtain fault determination results; a decision-making module for formulating differentiated topology adjustment strategies based on the fault determination results and using a digital twin for simulation verification of decision security; a topology execution module for controlling the operation of switching elements according to a preset timing sequence based on the topology adjustment instructions in the topology adjustment strategy, thereby reconstructing the series-parallel connection of the cells, modules, and busbars of the parallel DC power supply in the substation; and an optimization module for monitoring the operating status and adjustment effect of the parallel DC power supply in the substation and feeding back the monitoring results to the fault diagnosis module and the decision-making module for closed-loop optimization and feedback iteration.

[0021] Combination Figure 7 In this embodiment of the invention, the data acquisition module can be used to collect multi-dimensional data and align timestamps based on sensors deployed in the parallel DC power supply of the substation.

[0022] In a preferred embodiment of the present invention, the multi-dimensional data includes electrical quantity data of the cells, parallel modules, busbars and loads of the parallel DC power supply in the substation, as well as mechanical quantity data related to the structure of the parallel DC power supply in the substation, environmental quantity data around the structure and load characteristic quantity data.

[0023] In a preferred embodiment of the present invention, electrical quantity data may include voltage, current, internal resistance, insulation resistance, ripple characteristics, circulating current status, etc., directly reflecting the health status of the power transmission and conversion process; mechanical quantity data may include module housing vibration, busbar strain, connection node tightness, cell expansion degree, switch action response, etc., capturing early signals of aging or failure of the equipment's mechanical structure; environmental quantity data may include temperature and humidity, local temperature, electromagnetic interference intensity, dust concentration, etc., to understand the impact of environmental factors on system operation; load characteristic quantity data may include parameters such as identifying the electrical characteristics, power fluctuation patterns, and instantaneous power demand of different loads, clarifying the differences in power supply requirements of various loads.

[0024] For example, when monitoring the insulation condition of busbars, online insulation resistance monitoring sensors can be installed at busbar sections. Simultaneously, distributed sensing units can be embedded within the busbar itself. These sensors capture real-time changes in busbar insulation resistance and local strain. Combined with data from temperature and humidity sensors within the substation's parallel DC power supply equipment compartment, a complete data chain for the busbar's operating status can be formed. When slight aging occurs in the busbar insulation layer, the insulation resistance data will show a slow downward trend. Coupled with environmental data showing localized temperature increases, potential insulation faults can be detected early.

[0025] In this embodiment of the invention, the data processing module can be used to denoise and standardize multi-dimensional data, extract features, and generate fault identification feature vectors.

[0026] Please refer to Figure 2 In a preferred embodiment of the present invention, multi-dimensional data is denoised and standardized, and features are extracted to generate a fault identification feature vector, which may include: Step S110: Use differentiated noise reduction methods to filter high-frequency spike noise in electrical quantity data, separate operating noise and fault impact signals in mechanical quantity data, smooth instantaneous fluctuations in environmental quantity data, and remove abnormal spikes in load characteristic quantity data.

[0027] Step S120: Standardize the denoised multi-dimensional data and normalize it to a unified range to obtain preprocessed data.

[0028] In a preferred embodiment of the present invention, since the dimensions and numerical ranges of different physical quantities vary greatly, the data can be standardized to eliminate the influence of dimensions, and then normalized to a unified range to improve the computational efficiency and accuracy of subsequent algorithms and avoid diagnostic bias caused by numerical differences.

[0029] Step S130: Based on the preprocessed data, extract time-domain, frequency-domain, and spatiotemporal correlation features to generate a fault identification feature vector.

[0030] In a preferred embodiment of the present invention, time-domain features reflect the statistical regularity of data, frequency-domain features capture frequency distribution characteristics through signal conversion, and spatiotemporal correlation features uncover the intrinsic relationship between physical quantities at different locations and time points to construct a comprehensive fault identification feature vector.

[0031] For example, when processing module vibration data, the vibration signals generated during module operation include normal operation noise and potential fault impact signals. After separating the two types of signals using a targeted noise reduction algorithm, the effective value, peak factor, and other time-domain features of the vibration signals are extracted. Frequency-domain features such as characteristic frequencies and frequency band energy proportions are extracted through signal transformation. Simultaneously, combined with the current change data corresponding to the module, the temporal synchronization between vibration and current fluctuations is analyzed to form a feature vector that can distinguish between normal operation, module loosening, and internal component failure.

[0032] In this embodiment of the invention, the fault diagnosis module can be used to construct a digital twin architecture based on the fault identification feature vector, and use an attention mechanism diagnostic model to analyze the fault identification feature vector to obtain the fault determination result.

[0033] Please refer to Figure 3 In a preferred embodiment of the present invention, a digital twin architecture is constructed based on the fault identification feature vector, and an attention mechanism diagnostic model is used to analyze the fault identification feature vector to obtain the fault determination result, which may include: Step S210: Using a digital twin architecture, establish a three-dimensional virtual model corresponding to the physical quantities of the parallel DC power supply in the substation, and synchronize the physical quantities with the parameters of the three-dimensional virtual model.

[0034] In a preferred embodiment of the present invention, the normal operating parameter range under different working conditions can be calibrated by simulation and measured data to form a three-dimensional baseline database of "working condition-parameter-health status". The three-dimensional virtual model is dynamically updated with the collected data to ensure consistency with the physical entity status.

[0035] Step S220: Using the attention mechanism diagnostic model, through training with historical fault data and simulated fault samples, identify various typical faults in the fault identification feature vector and obtain the fault judgment result.

[0036] In a preferred embodiment of the present invention, the attention mechanism diagnostic model can automatically focus on key feature information for fault identification, assign higher weights to fault-sensitive physical quantities, reduce the influence of interference-type physical quantities, and has the ability to identify various typical faults. Furthermore, it can continuously iterate and optimize through real-time data collection to adapt to the aging characteristics of equipment.

[0037] In a preferred embodiment of the present invention, the fault determination result includes: when the fault level is minor, it corresponds to a single-point attenuation that does not affect the operation of the parallel DC power supply in the substation; when the fault level is moderate, it corresponds to a local functional abnormality of the parallel DC power supply in the substation; when the fault level is severe, it corresponds to a fault state that endangers the safety of the parallel DC power supply in the substation. Furthermore, by combining the spatial deployment information of the sensing units, precise positioning at the cell level, module level, and bus level can be achieved, while simultaneously clarifying the fault occurrence time and impact range.

[0038] For example, when a battery cell experiences a short-circuit fault, the data acquisition module captures electrical data such as a sudden drop in voltage and an abnormal increase in current, as well as mechanical data such as cell expansion and sudden vibration changes in surrounding modules. This data, after preprocessing, forms a feature vector that is input into an attention-based diagnostic model. The model uses this attention mechanism to focus on abnormal voltage and current changes, combined with cell expansion data, to quickly determine if it is a short-circuit fault, indicating its severity. Based on the data acquisition module's identification number, the model accurately locates the faulty cell and simultaneously outputs an assessment of the fault's impact on surrounding modules and the overall power supply.

[0039] In this embodiment of the invention, the decision-making module can be used to formulate differentiated topology adjustment strategies based on fault determination results, and to use a digital twin to simulate and verify the safety of the decision.

[0040] Please refer to Figure 4 In a preferred embodiment of the present invention, a differentiated topology adjustment strategy is formulated based on the fault determination result, including: Step S310: Construct multi-level decision rules based on fault levels.

[0041] In a preferred embodiment of the present invention, the multi-level decision-making rules include: when the fault level is minor, only the load distribution strategy is adjusted to transfer the load of the faulty unit to the healthy unit to avoid further degradation of the faulty unit; when the fault level is moderate, local topology reconfiguration is triggered to isolate the faulty module or cell and maintain the output parameters in compliance through the series-parallel reconstruction of the healthy unit; when the fault level is severe, global topology switching is triggered to isolate the faulty area and switch to the backup topology architecture to prioritize the power supply of critical loads.

[0042] Step S320: Based on the load type in the fault determination result, formulate a topology adaptation strategy corresponding to the load type.

[0043] In a preferred embodiment of the present invention, the load type can be identified first, distinguishing between instantaneous high-power loads, stable loads, and sensitive loads, as different types of loads have different requirements for power supply stability, ripple control, and power response speed. Corresponding topology adaptation strategies are formulated for different load types, such as adjusting the series and parallel connection of modules and allocating dedicated power supply modules, to meet the differentiated needs of various loads. Simultaneously, based on the system health status and load fluctuation predictions, redundant capacity is reserved and dynamically allocated to avoid power shortages caused by sudden load changes.

[0044] Step S330: Based on multi-level decision rules and topology adaptation strategies, obtain differentiated topology adjustment strategies.

[0045] In a preferred embodiment of the present invention, after the topology adjustment strategy is generated, it is first simulated in the virtual model of the digital twin to simulate the power supply stability, current distribution balance, and fault isolation effectiveness after topology adjustment. The topology adjustment command is only issued after verifying the absence of risks such as overvoltage and overcurrent. A dual-backup decision-making mechanism is set up to avoid execution failures caused by the failure of the decision-making module. When fault handling conflicts with load adaptation, the principles of safety priority and critical load priority are followed.

[0046] For example, when the diagnostic system identifies an excessive ripple in the output of a parallel module, indicating a moderate fault, the decision-making module first simulates the system's operating state after fault isolation using a digital twin. Considering that this module originally supplied power to secondary equipment (stable load), the topology adjustment strategy is to isolate the faulty module and change the connection of two adjacent healthy modules from series to parallel to supplement the output capacity of the faulty module. Simultaneously, the load distribution ratio of these two healthy modules is adjusted to ensure that the power supply voltage stability and ripple parameters of the secondary equipment meet the requirements. After simulation verification confirms no risk of voltage drop, a topology adjustment command is issued.

[0047] In this embodiment of the invention, the topology execution module can be used to control the operation of switching elements according to a preset timing sequence based on the topology adjustment instructions in the topology adjustment strategy, so as to realize the series-parallel reconfiguration of the cells, modules and buses of the parallel DC power supply in the substation.

[0048] Please refer to Figure 5 In a preferred embodiment of the present invention, controlling the operation of switching elements according to a preset timing sequence based on the topology adjustment instruction in the topology adjustment strategy includes: Step S410: Using solid-state relays as switching elements, a multi-level flexible topology switch matrix structure is constructed to configure corresponding switching elements for the cells, modules, and bus sections of the parallel DC power supply in the substation, and to configure a spare switch for each switching element.

[0049] In a preferred embodiment of the invention, solid-state relays with fast response speed and strong voltage and current resistance can be used as switching elements at the core to construct a three-level matrix structure of "cell-level switch array - module-level switch array - bus-level switch array". Each cell, module, and bus segment is configured with a corresponding switching element to achieve independent access, isolation, and connection mode switching functions. Each switching element is equipped with a backup switch to ensure rapid replacement in case of switch failure through hot backup, avoiding topology execution failure.

[0050] Step S420: A dual-core controller is used to generate switch drive signals and monitor the status of switch elements to perform topology switching control and drive.

[0051] In a preferred embodiment of the invention, a dual-core controller can be used to generate the switch drive signal and monitor the switch status. The drive circuit uses optocoupler isolation design to ensure reliable signal transmission and integrates overvoltage and overcurrent protection functions. When the electrical parameters at both ends of the switch exceed the safe range, the drive signal is immediately cut off. Strict switching timing sequences are established to avoid bus short circuits or load power outages during switching, ensuring continuous power supply.

[0052] Step S430: During the topology switching control and drive process, multiple protection measures are taken to ensure the safe and uninterrupted switching process of the parallel DC power supply in the substation.

[0053] Please refer to Figure 6 In a preferred embodiment of the present invention, step S430 may include: Step S4301: Arrive suppression elements are configured at both ends of the multi-level flexible topology switch matrix structure to suppress the arc generated during the switching process, so as to protect the switch elements and the bus.

[0054] Step S4302: Connect an energy storage element in parallel at the load input terminal of the parallel DC power supply in the substation. When the topology switching causes instantaneous voltage fluctuations, the energy storage element discharges quickly to compensate and maintain the voltage stability at the load terminal.

[0055] Step S4303: Monitor the on / off status of each switching element. If a switching element is detected to be not operating according to the topology adjustment command, trigger the backup switch to switch and report the fault information.

[0056] For example, when three health modules need to be switched from series to parallel to meet the power supply requirements of circuit breaker opening and closing, i.e., instantaneous high-power loads, the flexible topology switch matrix operates according to a preset timing sequence: first, the parallel branch switches between the three modules are turned on; after the parallel branches stabilize, the original series branch switches are then turned off. During the switching process, the energy storage element at the load input compensates for instantaneous voltage fluctuations, the arc suppression element prevents arc damage to the equipment caused by the switching action, and the controller monitors the operating status of all switches in real time to ensure that the three modules successfully switch to parallel mode, thereby improving the instantaneous power output capability.

[0057] In this embodiment of the invention, the optimization module can be used to monitor the operating status and adjustment effect of the parallel DC power supply in the substation, and feed the monitoring results back to the fault diagnosis module and the decision-making module for closed-loop optimization and feedback iteration.

[0058] In a preferred embodiment of the present invention, after topology switching, the data acquisition module continuously collects system output parameters, topology status, and load response data. It focuses on monitoring the stability of output voltage and current, the load distribution balance of healthy units, the isolation effectiveness of faulty units, and the operating status of critical loads to comprehensively evaluate the topology adjustment effect.

[0059] In a preferred embodiment of the invention, the adjustment effect is evaluated based on monitoring data. If problems such as substandard power supply parameters or unbalanced load distribution exist, the information is fed back to the decision-making module. The decision-making module then re-optimizes the topology adjustment strategy until the system returns to its optimal operating state. Simultaneously, fault data, diagnostic results, topology adjustment strategies, and adjustment effects are stored to form a case library, providing data support for subsequent optimization.

[0060] In a preferred embodiment of the present invention, the weight parameters and decision rules of the attention mechanism diagnostic model can be optimized periodically based on case library data to improve the accuracy of fault diagnosis and the rationality of topology adjustment; the redundancy configuration and switching strategy of the topology switch matrix can be optimized through digital twin simulation according to the equipment's operating years and load changes; and maintenance suggestions can be generated periodically based on historical fault data and equipment health status assessment, combining self-healing operation with preventive maintenance.

[0061] For example, after topology adjustments, if monitoring detects that a certain healthy module's load percentage is too high, exceeding the balanced load distribution range, this data is transmitted to the decision-making module. The decision-making module, considering the operational status of other healthy modules, re-formulates the load distribution strategy, fine-tuning the topology to transfer the excessive load to other healthy modules with lower load percentages. Simultaneously, the case study, adjustment plan, and effects of this load imbalance are stored in a case library. Subsequent optimization of decision rules will focus on refining the load distribution balance judgment criteria and adjustment algorithms to prevent similar problems from recurring.

[0062] This invention also provides a fault diagnosis method for a parallel DC power supply in a 110kV substation, the fault diagnosis method comprising: Step S1: Collect multi-dimensional data and align the timestamps based on the sensors deployed in the parallel DC power supply of the substation; Step S2: Denoise and standardize the multi-dimensional data, and extract features to generate fault identification feature vectors; Step S3: Based on the fault identification feature vector, construct a digital twin architecture and use an attention mechanism diagnostic model to analyze the fault identification feature vector to obtain the fault determination result; Step S4: Based on the fault determination results, formulate differentiated topology adjustment strategies and use digital twins to simulate and verify the safety of the decisions; Step S5: According to the topology adjustment instructions in the topology adjustment strategy, control the operation of the switching elements according to the preset timing sequence to realize the series-parallel reconfiguration of the cells, modules and busbars of the parallel DC power supply in the substation; Step S6: Monitor the operating status and adjustment effect of the parallel DC power supply in the substation, and feed the monitoring results back to the fault diagnosis module and decision-making module for closed-loop optimization and feedback iteration.

[0063] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0064] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0065] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0066] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0067] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0068] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0069] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0070] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0071] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A fault diagnosis system for a parallel DC power supply in a 110kV substation, characterized in that, The fault diagnosis system includes: The data acquisition module is used to collect multi-dimensional data and align timestamps based on sensors deployed in the parallel DC power supply of the substation. The data processing module is used to denoise and standardize the multi-dimensional data, extract features, and generate fault identification feature vectors. The fault diagnosis module is used to construct a digital twin architecture based on the fault identification feature vector, and to analyze the fault identification feature vector using an attention mechanism diagnostic model to obtain the fault determination result. The decision-making module is used to formulate differentiated topology adjustment strategies based on the fault determination results, and to use a digital twin to simulate and verify the safety of the decision. The topology execution module is used to control the operation of switching elements according to a preset timing sequence based on the topology adjustment instructions in the topology adjustment strategy, so as to realize the series-parallel reconfiguration of the cells, modules and buses of the parallel DC power supply in the substation. The optimization module is used to monitor the operating status and adjustment effect of the parallel DC power supply in the substation, and feeds back the monitoring results to the fault diagnosis module and the decision-making module for closed-loop optimization and feedback iteration.

2. The fault diagnosis system according to claim 1, characterized in that, The multi-dimensional data includes electrical quantity data of the cells, parallel modules, busbars and loads of the parallel DC power supply in the substation, as well as mechanical quantity data related to the structure of the parallel DC power supply in the substation, environmental quantity data around the structure and load characteristic quantity data.

3. The fault diagnosis system according to claim 2, characterized in that, The step of denoising and standardizing the multi-dimensional data, and extracting features to generate a fault identification feature vector includes: A differentiated noise reduction method is used to filter high-frequency spike noise in the electrical quantity data, separate the operating noise and fault impact signal in the mechanical quantity data, smooth the instantaneous fluctuations in the environmental quantity data, and remove abnormal spikes in the load characteristic quantity data. The denoised multi-dimensional data is standardized and normalized to a unified range to obtain preprocessed data. Based on the preprocessed data, time-domain, frequency-domain, and spatiotemporal correlation features are extracted to generate a fault identification feature vector.

4. The fault diagnosis system according to claim 1, characterized in that, The process of constructing a digital twin architecture based on the fault identification feature vector and analyzing the fault identification feature vector using an attention mechanism diagnostic model to obtain a fault determination result includes: A digital twin architecture is adopted to establish a three-dimensional virtual model corresponding to the physical quantities of the parallel DC power supply in the substation, and to achieve synchronization between the physical quantities and the parameters of the three-dimensional virtual model. By using an attention mechanism diagnostic model and training it with historical fault data and simulated fault samples, the model identifies various typical faults in the fault identification feature vector and obtains the fault determination result.

5. The fault diagnosis system according to claim 4, characterized in that, The fault determination results include: When the fault level is minor, it corresponds to a single-point attenuation that does not affect the operation of the parallel DC power supply in the substation. When the fault level is medium, it corresponds to a partial functional abnormality of the parallel DC power supply in the power station. When the fault level is severe, it corresponds to a fault state that endangers the safety of the parallel DC power supply in the substation.

6. The fault diagnosis system according to claim 5, characterized in that, The step of formulating differentiated topology adjustment strategies based on the fault determination results includes: Based on the aforementioned fault levels, multi-level decision rules are constructed; Based on the load type in the fault determination result, formulate a topology adaptation strategy corresponding to the load type; Based on the multi-level decision rules and topology adaptation strategy, a differentiated topology adjustment strategy is obtained.

7. The fault diagnosis system according to claim 6, characterized in that, The multi-level decision-making rules include: When the fault level is minor, only the load distribution strategy is adjusted to transfer the load of the faulty unit to the healthy unit to prevent the faulty unit from further deteriorating. When the fault level is medium, a local topology reconfiguration is triggered to isolate the faulty module or cell and maintain the output parameters in compliance through the series-parallel reconfiguration of healthy units. When the fault level is severe, a global topology switch is triggered to isolate the faulty area and switch to the backup topology architecture, prioritizing power supply to critical loads.

8. The fault diagnosis system according to claim 1, characterized in that, The step of controlling the switching element to operate according to a preset timing sequence based on the topology adjustment instruction in the topology adjustment strategy includes: Solid-state relays are used as switching elements to construct a multi-level flexible topology switch matrix structure. Corresponding switching elements are configured for the cells, modules, and bus sections of the parallel DC power supply in the substation, and a spare switch is configured for each switching element. A dual-core controller is used to generate switch drive signals and monitor the status of switching elements to perform topology switching control and drive. Multiple protection measures are adopted during topology switching control and drive to ensure safe and uninterrupted switching of parallel DC power supplies in substations.

9. The fault diagnosis system according to claim 8, characterized in that, Multiple protection measures are adopted during the topology switching control and driving process, including: Arc suppression elements are configured at both ends of the multi-level flexible topology switch matrix structure to suppress the arc generated during switching, thereby protecting the switch elements and the bus. In the parallel DC power supply of the substation, an energy storage element is connected in parallel at the load input end. When the topology switching causes instantaneous voltage fluctuations, the energy storage element discharges quickly to compensate and maintain the voltage stability at the load end. Monitor the on / off status of each switching element. If a switching element is detected to have failed to operate according to the topology adjustment command, trigger the backup switch to switch and report the fault information.

10. A fault diagnosis method for a parallel DC power supply in a 110kV substation, characterized in that, The fault diagnosis method includes: Based on sensors deployed in the parallel DC power supply of the substation, multi-dimensional data is collected and timestamps are aligned. The multi-dimensional data is denoised and standardized, and features are extracted to generate a fault identification feature vector. Based on the fault identification feature vector, a digital twin architecture is constructed, and an attention mechanism diagnostic model is used to analyze the fault identification feature vector to obtain the fault determination result. Based on the fault determination results, differentiated topology adjustment strategies are formulated, and digital twins are used for simulation to verify the safety of the decisions. According to the topology adjustment instructions in the topology adjustment strategy, the switching elements are controlled to operate according to a preset timing sequence in order to realize the series-parallel reconfiguration of the cells, modules and busbars of the parallel DC power supply in the substation. Monitor the operating status and adjustment effect of the parallel DC power supply in the substation, and feed the monitoring results back to the fault diagnosis module and decision-making module for closed-loop optimization and feedback iteration.

Images