Intelligent interaction diagnosis method based on substation multi-dimensional information cloud computing power fusion

Abstract

The invention discloses an intelligent interaction diagnosis method based on substation multi-dimensional information cloud computing power fusion, and belongs to the field of fault diagnosis. The method is applied to a cloud computing power platform, and comprises the following steps: firstly, receiving high-frequency waveform data and operation remote signaling data of an auxiliary power supply system of a transformer substation, marking and positioning a start-stop event window according to a load start-stop event, and performing multi-scale time-frequency decomposition on the waveform data in the window to generate a two-dimensional time-frequency map; and comparing the structural similarity of the atlas and the template, and generating and displaying the degradation index of each load. Then receiving a detection instruction of a user on a selected target load, triggering to inject a preset square wave pulse into a branch cable of the load, and collecting reflection waveform data; and finally, the cloud computing power platform obtains defect property information by determining polarity information of the reflection waveform data, and determines and displays final fault diagnosis information in a diagnosis rule knowledge base according to the defect property information and the degradation index, so that the reliability and accuracy of diagnosis can be improved.

Description

Technical Field

[0001] This application belongs to the field of fault diagnosis, and in particular relates to an intelligent interactive diagnostic method based on the fusion of multi-dimensional information cloud computing power of substations. Background Technology

[0002] With the acceleration of power grid intelligence, intelligent substation diagnostic technology based on cloud computing power integration has become an important development direction for realizing intelligent diagnosis and condition assessment of substation equipment. This method leverages the powerful computing and storage capabilities of the cloud to deeply mine data collected on-site and provides decision support for operation and maintenance personnel through a human-machine interface, showing broad application prospects.

[0003] In existing technologies, common substation diagnostic methods involve collecting parameters such as current or voltage waveforms from various devices, performing Fast Fourier Transform on these parameters in the cloud to obtain the signal's spectral distribution, and then comparing the amplitude of specific harmonic components in the spectrum with preset health status thresholds to determine whether the load has functional abnormalities.

[0004] However, existing diagnostic methods focus on the spectral characteristics of equipment under steady-state operation. Their diagnostic conclusions only reflect deviations in electrical parameters and lack the ability to analyze transient operating conditions such as equipment start-up and shutdown. In particular, they cannot provide a comprehensive and accurate diagnosis of the complete operating status of equipment in complex scenarios. Therefore, existing diagnostic methods suffer from insufficient reliability and accuracy in their diagnostic results. Summary of the Invention

[0005] This application provides an intelligent interactive diagnostic method, system, device, and computer storage medium based on the fusion of multi-dimensional information cloud computing power in substations, which can improve the reliability and accuracy of diagnosis.

[0006] Firstly, this application provides an intelligent interactive diagnostic method based on the fusion of multi-dimensional information from substations and cloud computing power, applied to a cloud computing power platform. The method includes: Receive high-frequency waveform data of the main bus of the auxiliary power system sent by the auxiliary power system of the substation, as well as remote signaling data of the operation of multiple loads in the auxiliary power system; Based on the load start-stop event markers of the remote signaling data, the start-stop event window of each load is located in the high-frequency waveform data, and multi-scale time-frequency decomposition is performed on the high-frequency waveform data within the start-stop event window to generate a two-dimensional time-frequency spectrum. The two-dimensional time-frequency spectrum is compared with the preset time-frequency spectrum template for structural similarity to generate the degradation index of each load, and the degradation index of each load is displayed on the interactive interface. The system receives a pulse detection command input by the user to detect the target load selected by the user, sends the pulse detection command to the auxiliary power system, and receives the reflected waveform data of the target load sent by the auxiliary power system. The pulse detection command is used to control the auxiliary power system to inject a preset square wave pulse into the branch cable corresponding to the target load and collect the reflected waveform data. By determining the polarity information of the reflected waveform data, the defect nature information of the branch cable corresponding to the target load can be determined; Based on the degradation index and defect nature information, the fault diagnosis information of the target load is determined in the preset diagnostic rule knowledge base, and the fault diagnosis information of the target load is displayed on the interactive interface.

[0007] In one feasible implementation, a structural similarity comparison is performed between the two-dimensional time-frequency spectrum and a preset time-frequency spectrum template to generate a degradation index for each load, including: The two-dimensional time-frequency spectrum is divided along the frequency axis into an electrical characteristic time-frequency spectrum and a mechanical characteristic time-frequency spectrum. The electrical characteristic time-frequency spectrum covers the power supply fundamental frequency of the auxiliary power supply system and the preset integer multiple harmonic frequency bands, while the mechanical characteristic time-frequency spectrum covers the non-harmonic and high frequency bands other than the electrical characteristic time-frequency spectrum. Extract the electrical template map and mechanical template map corresponding to the electrical feature time-frequency map and mechanical feature time-frequency map from the preset time-frequency map template; Calculate the first similarity score between the electrical feature time-frequency spectrum and the electrical template spectrum, and the second similarity score between the mechanical feature time-frequency spectrum and the mechanical template spectrum, respectively. The first similarity score and the second similarity score are converted into electrical deterioration index and mechanical deterioration index, respectively. The deterioration index is obtained by weighted summation of the electrical deterioration index and the mechanical deterioration index.

[0008] In one feasible implementation, the method further includes: Based on the reflected waveform data, the round-trip time of the preset square wave pulse in the branch cable corresponding to the target load is calculated to determine the defect location information of the branch cable corresponding to the target load. Based on the defect location information and the preset cable network topology data of the auxiliary power system, the target associated load that is in the same branch cable as the target load and is far away from the main bus of the auxiliary power system relative to the target load is identified in the cable network topology data. The range definition information is generated by calculating the absolute value of the difference between the degradation index of the target load and the degradation index of the target associated load, and comparing the absolute value of the difference with a preset threshold.

[0009] In one feasible implementation, based on the load start-stop event markers of the running remote signaling data, the start-stop event window of each load is located in the high-frequency waveform data, and multi-scale time-frequency decomposition is performed on the high-frequency waveform data within the start-stop event window to generate a two-dimensional time-frequency spectrum, including: Determine the start and stop event timestamps for each load in the running remote signaling data; Based on the start and stop event time points, the start and end boundaries of the start and stop event windows for each load are determined according to the preset time offset. In high-frequency waveform data, waveform data segments corresponding to the time period are extracted based on the start and end boundaries; The waveform data segment is divided into multiple time scales, and the frequency distribution is calculated at each time scale. The results are then combined to form a two-dimensional time-frequency spectrum.

[0010] In one feasible implementation, the auxiliary power supply system is controlled to inject a preset square wave pulse into the branch cable corresponding to the target load and to acquire reflected waveform data, including: Send a status query command to the auxiliary power system to obtain the operating status of the target load; When the target load is in standby or light load condition, a synchronous trigger signal is sent to the signal generation unit and data acquisition unit in the auxiliary power system. The synchronous trigger signal is used to drive the signal generation unit to generate an injection signal based on a preset square wave pulse and inject it into the branch cable corresponding to the target load. At the same time, it drives the data acquisition unit to record the mixed waveform including the injection signal and the reflected signal. The reflected waveform data is separated from the mixed waveform by a bandpass filter. The passband range of the bandpass filter covers the fundamental frequency and third harmonic frequency band of the preset square wave pulse.

[0011] In one feasible implementation, by determining the polarity information of the reflected waveform data, the defect nature information of the branch cable corresponding to the target load is determined, including: Zero-point level correction is performed on the reflected waveform data, and the first main reflected pulse is located in the corrected reflected waveform. The peak direction of the primary reflected pulse is determined by calculating the maximum deviation direction of the waveform amplitude of the primary reflected pulse relative to the zero-point level. By comparing the peak direction with the reference polarity of a preset square wave pulse, defect nature information is obtained, which includes high-impedance defects and low-impedance defects.

[0012] In one feasible implementation, based on the degradation index and defect nature information, fault diagnosis information of the target load is determined from a pre-defined diagnostic rule knowledge base, including: Based on the defect nature information, a subset of candidate fault modes is selected from the pre-set diagnostic rule knowledge base; By comparing the degradation index with the degradation index threshold range corresponding to each candidate failure mode, the target failure mode is determined to obtain failure diagnosis information.

[0013] Secondly, this application provides an intelligent interactive diagnostic system based on the fusion of multi-dimensional information and cloud computing power from substations, applied to a cloud computing platform. The system includes: The receiving module is used to receive high-frequency waveform data of the main bus of the auxiliary power system sent by the auxiliary power system of the substation, as well as remote signaling data of the operation of multiple loads in the auxiliary power system. The generation module is used to locate the start and stop event window of each load in the high-frequency waveform data based on the load start and stop event markers of the running remote signaling data, and to perform multi-scale time-frequency decomposition on the high-frequency waveform data within the start and stop event window to generate a two-dimensional time-frequency spectrum. The generation module is also used to compare the structural similarity of the two-dimensional time-frequency spectrum with the preset time-frequency spectrum template, generate the degradation index of each load, and display the degradation index of each load on the interactive interface. The receiving module is also used to receive the pulse detection command input by the user to detect the target load selected by the user, send the pulse detection command to the auxiliary power system, and receive the reflected waveform data of the target load sent by the auxiliary power system. The pulse detection command is used to control the auxiliary power system to inject a preset square wave pulse into the branch cable corresponding to the target load and collect the reflected waveform data. The determination module is used to determine the defect nature information of the branch cable corresponding to the target load by determining the polarity information of the reflected waveform data; The display module is used to determine the fault diagnosis information of the target load from a preset diagnostic rule knowledge base based on the degradation index and defect nature information, and to display the fault diagnosis information of the target load on the interactive interface.

[0014] Thirdly, this application provides an electronic device, which includes: a processor and a memory storing computer program instructions; the processor reads and executes the computer program instructions to implement the intelligent interactive diagnostic method based on the fusion of multi-dimensional information cloud computing power of substations as described in any embodiment of the first aspect.

[0015] Fourthly, this application provides a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the intelligent interactive diagnostic method based on the fusion of multi-dimensional information cloud computing power of substations as described in any embodiment of the first aspect.

[0016] This application discloses an intelligent interactive diagnostic method, system, equipment, and computer storage medium based on the fusion of multi-dimensional information cloud computing power in substations. It generates a two-dimensional time-frequency spectrum that comprehensively characterizes the transient processes of load start-up and shutdown by performing multi-scale time-frequency decomposition on the high-frequency waveform data of the main bus of the substation auxiliary power system. This spectrum is then compared with a preset template to generate a degradation index, overcoming the shortcomings of existing technologies that only focus on steady-state spectrum while ignoring transient information. Through user interaction, physical detection of the target load is triggered, and the degradation index reflecting its functional state is fused with the defect information reflecting the physical state of its power supply cable for diagnosis. This effectively links functional abnormalities with physical defects, forming a complete diagnostic evidence chain and improving the accuracy of fault root cause tracing and the reliability of diagnostic conclusions in complex scenarios.

[0017] Furthermore, by dividing the two-dimensional time-frequency spectrum into electrical characteristic time-frequency spectrum and mechanical characteristic time-frequency spectrum along the frequency axis and comparing their similarity, the comprehensive degradation state can be decomposed into electrical degradation index and mechanical degradation index with different sources. This decoupling of electromechanical state makes the diagnosis no longer limited to judging whether there is an abnormality, but can further distinguish whether the physical root cause of the abnormality is more inclined to electrical circuit problems or mechanical structure problems. Therefore, it can provide deeper diagnostic information and improve the precision of diagnosis of complex electromechanical coupling faults. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating an embodiment of the intelligent interactive diagnostic method based on the fusion of multi-dimensional information and cloud computing power of substations provided in this application. Figure 2 This is a schematic diagram of a two-dimensional time-frequency spectrum provided in one embodiment of this application; Figure 3 This is a schematic diagram of a time-frequency spectrum template provided in one embodiment of this application; Figure 4 This is a schematic diagram of the structure of an intelligent interactive diagnostic system based on the fusion of multi-dimensional information cloud computing power of substations, provided in one embodiment of this application; Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in one embodiment of this application. Detailed Implementation

[0020] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0021] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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 a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0022] In existing technologies, common substation diagnostic methods involve collecting parameters such as current or voltage waveforms from various devices, performing Fast Fourier Transform on these parameters in the cloud to obtain the signal's spectral distribution, and then comparing the amplitude of specific harmonic components in the spectrum with preset health status thresholds to determine whether the load has functional abnormalities.

[0023] However, existing diagnostic methods focus on the spectral characteristics of equipment under steady-state operation. Their diagnostic conclusions only reflect deviations in electrical parameters and lack the ability to analyze transient operating conditions such as equipment start-up and shutdown. In particular, they cannot provide a comprehensive and accurate diagnosis of the complete operating status of equipment in complex scenarios. Therefore, existing diagnostic methods suffer from insufficient reliability and accuracy in their diagnostic results.

[0024] To address the problems of existing technologies, embodiments of this application provide an intelligent interactive diagnostic method, system, device, and computer storage medium based on the fusion of multi-dimensional information cloud computing power from substations. The intelligent interactive diagnostic method based on the fusion of multi-dimensional information cloud computing power from substations provided in this application embodiment will be described below first.

[0025] The overall architecture of this embodiment includes a remotely deployed cloud computing platform and one or more substations as diagnostic targets. Each substation is equipped with an auxiliary power system to provide power to multiple load devices within the substation.

[0026] Figure 1 This illustration shows a flowchart of an intelligent interactive diagnostic method based on the fusion of multi-dimensional information from substations and cloud computing power, according to an embodiment of this application. This method is applied to cloud computing power platforms, such as... Figure 1 As shown, the method includes steps S110 to S160.

[0027] A cloud computing platform refers to a pool of computing and storage resources consisting of high-performance server clusters deployed in a remote data center. An auxiliary power system refers to the critical power supply network within a substation that provides AC and DC power to all control, communication, and power equipment.

[0028] S110: Receives high-frequency waveform data of the main bus of the auxiliary power system sent by the auxiliary power system of the substation, as well as remote signaling data of the operation of multiple loads in the auxiliary power system.

[0029] High-frequency waveform data refers to the sequence of raw sampled values ​​representing instantaneous changes in voltage and current, acquired by high-sampling-rate data acquisition units deployed on the main bus of the auxiliary power supply system. The sampling frequency is generally above the kilohertz level, which can capture details in dynamic processes such as load start-up and shutdown. Operational remote signaling data refers to discrete signals representing the operating status of each load device, acquired by the Supervisory Control and Data Acquisition (SCADA) system of the auxiliary power supply system. For example, operational remote signaling data can be Boolean values ​​used to indicate the running or stopped status of a cooling fan motor. Each piece of operational remote signaling data itself constitutes a load start-up and shutdown event marker.

[0030] The cloud computing platform establishes a communication connection with the substation via a data transmission gateway. The substation's data acquisition unit continuously collects instantaneous voltage and current values ​​from the main bus of the auxiliary power system at a preset high sampling frequency, such as 10,000 times per second, and packages these instantaneous value sequences into high-frequency waveform data packets with precise timestamps. Simultaneously, the cloud computing platform subscribes to the remote signaling services of the substation's SCADA system in real time. If the operating status of any load, such as the main transformer cooling oil pump or the gas-insulated switchgear compartment heater, changes, the SCADA system generates an operational remote signaling data record containing the device identifier (ID), the status change (e.g., from 0 to 1), and a timestamp—essentially a load start / stop event marker.

[0031] For example, taking a large hub substation as an example, at the start of the diagnostic task, the cloud computing platform first initiates a data acquisition command for the auxiliary power supply DC bus and AC bus through its edge computing nodes deployed in the substation. The high-speed acquisition card installed at the bus outlet captures the total voltage and current waveforms superimposed on the electrical behavior of all downstream loads on the bus at a sampling rate of 20kHz, forming high-frequency waveform data. Upon detecting the start of the cooling oil pump of the main transformer, the SCADA system generates a remote signaling data entry as a load start / stop event marker, assuming the content is "Device ID: PUMP-02, Status: 1, Timestamp: 10:00:00.123". After receiving this remote signaling data and the high-frequency waveform data from the same time period, the cloud computing platform timestamps the confirmation data packet.

[0032] S120: Based on the load start-stop event markers of the remote signaling data, locate the start-stop event window of each load in the high-frequency waveform data, and perform multi-scale time-frequency decomposition on the high-frequency waveform data within the start-stop event window to generate a two-dimensional time-frequency spectrum.

[0033] Load start / stop event markers are operational telemetry data records that uniquely identify a load state change, including the timestamp of the event and the direction of the state change. Start / stop event windows are data segments of a specific duration captured on the time axis of high-frequency waveform data, capable of completely covering the entire process from load startup or shutdown to its entry into a new stable operating state.

[0034] First, using the start and stop event time points contained in the load start and stop event markers, the start and end boundaries of the start and stop event window are calculated based on a preset time offset for different types of loads. For example, for a motor load, whose start-up process is relatively long, the preset time offset can be set to 0.1 seconds before the start event time point to 2 seconds after start-up, thus forming a 2.1-second start and stop event window. Then, based on these calculated start and end boundaries, a waveform data segment corresponding to this time period is extracted from the high-frequency waveform data aligned with it. Finally, this waveform data segment undergoes multi-scale time-frequency decomposition. This can be achieved by applying signal processing methods such as continuous wavelet transform or short-time Fourier transform, converting the one-dimensional time-domain waveform data into a two-dimensional matrix of energy distribution in time and frequency, i.e., a two-dimensional time-frequency spectrum. Here, the horizontal axis represents time, the vertical axis represents frequency, and the pixel value represents energy intensity. The two-dimensional time-frequency spectrum is used to visually represent the distribution of signal energy in both time and frequency dimensions.

[0035] Figure 2 A schematic diagram of a two-dimensional time-frequency spectrum provided in one embodiment of this application is shown.

[0036] like Figure 2As shown, in the two-dimensional time-frequency spectrum, the horizontal axis represents time and the vertical axis represents frequency. Pixel values ​​are displayed in grayscale, with dark areas representing high energy density and light areas representing low energy density. The two-dimensional time-frequency spectrum includes three key stages: the stationary stage, which presents extremely low background noise; the startup transition stage, which shows the broadband impact and frequency spread at the moment of startup; and the stable operation stage, which mainly focuses on the 50Hz power frequency and its harmonic components.

[0037] S130: Compare the two-dimensional time-frequency spectrum with the preset time-frequency spectrum template for structural similarity, generate the degradation index of each load, and display the degradation index of each load on the interactive interface.

[0038] A time-frequency graph template refers to a standard two-dimensional time-frequency graph representing the start-up and shutdown of a specific load in a healthy state. It is established based on multiple start-up and shutdown processes of the load during factory testing or historical normal operation. The degradation index is a quantified value within a preset range of 0 to 100, which can represent the degree of deviation of the current health state of the load from the ideal health state. The higher the value, the more severe the degradation.

[0039] A startup time-frequency spectrum template for a healthy motor exhibits the following energy distribution: initially, energy is concentrated in the lower-frequency impact component region, then rapidly shifts towards the power frequency (e.g., 50 Hz) and its odd harmonic bands (e.g., 150 Hz and 250 Hz); energy distribution is very weak in other non-harmonic and high-frequency regions. A preset time-frequency spectrum template corresponding to the current two-dimensional time-frequency spectrum is retrieved from the database for structural similarity comparison. This can be achieved by calculating the Structural Similarity Index (SSIM) or Root Mean Square Error (RMSE) between the two two-dimensional time-frequency spectra, yielding a similarity score that quantifies the structural differences between them. A non-linear mapping function converts the calculated similarity score or deviation into an intuitive degradation index. For example, a similarity score of 1.0 (identical) maps to a degradation index of 0, while a similarity score of 0.5 maps to a degradation index of 80. The generated degradation index can be pushed to the interactive interface and displayed to the user in a list or topology diagram with coloring.

[0040] Figure 3 A schematic diagram of a time-frequency spectrum template provided in one embodiment of this application is shown.

[0041] like Figure 3As shown, the horizontal axis of the time-frequency spectrum template represents time, and the vertical axis represents frequency. Pixel values ​​are displayed in grayscale, with dark areas representing high energy density and light areas representing low energy density. The grayscale distribution of pixel values ​​represents the energy intensity characteristics under healthy conditions, with dark gray indicating areas of concentrated strong energy. The time-frequency spectrum template also includes three key stages: the static stage, the startup transition stage, and the stable operation stage. Compared to a two-dimensional time-frequency spectrum, the startup process is smoother, the frequency transition is more regular, the 50Hz power frequency component dominates, the 150Hz and 250Hz harmonics are clear, the energy distribution in the high-frequency region is uniform, and the background noise is extremely low.

[0042] S140: Receives a pulse detection command input by the user to detect the target load selected by the user, sends the pulse detection command to the auxiliary power system, and receives the reflected waveform data of the target load sent by the auxiliary power system. The pulse detection command is used to control the auxiliary power system to inject a preset square wave pulse into the branch cable corresponding to the target load and collect the reflected waveform data.

[0043] A preset square wave pulse refers to a rectangular electrical signal with a fixed amplitude, pulse width, and repetition frequency, which can be injected into a low-voltage cable and generate reflection. The reflected waveform data refers to the time-domain waveform sequence of the reflected signal generated when the injected preset square wave pulse propagates in the cable and encounters an impedance discontinuity, which is recorded by the data acquisition unit.

[0044] When maintenance personnel click on a load icon with a high degradation index on the interactive interface and select the detection function, the cloud computing platform receives user input including the target load ID and encapsulates it into a pulse detection command. Before sending this command, the cloud computing platform queries the auxiliary power system for the target load's operating status. Only after confirming that it is in standby or light-load safe operating conditions will it officially send the command. Upon receiving the pulse detection command, the control unit of the substation's auxiliary power system drives a signal generation unit to generate a preset square wave pulse, for example, with an amplitude of 5 volts and a pulse width of 100 nanoseconds, and injects it into the branch cable connected to the target load. Simultaneously, a high-speed data acquisition unit, triggered synchronously with the injection, begins recording the signal on the cable, i.e., a waveform sequence including multiple reflections. After internal signal separation processing, such as bandpass filtering to extract the reflected waveform data, the reflected waveform data, the target load ID, and the current timestamp are finally sent back to the cloud computing platform.

[0045] S150: By determining the polarity information of the reflected waveform data, the defect nature information of the branch cable corresponding to the target load is determined.

[0046] Polarity information refers to the direction of the main peak of the reflected pulse signal relative to zero level, i.e., whether it is positive or negative. Defect nature information is a classification description of the physical properties of impedance discontinuities in a cable, distinguishing whether the defect point exhibits a higher or lower impedance than the cable's normal characteristic impedance. For example, defect nature information can include high-impedance defects and low-impedance defects. High-impedance defects typically correspond to cable breaks or poor connector contact, while low-impedance defects typically correspond to damage to the cable insulation, moisture absorption, or short circuits to ground.

[0047] After receiving the reflected waveform data from the target load, a zero-level correction is first performed on the reflected waveform data to eliminate possible DC bias effects. Then, in the corrected waveform, the first main reflected pulse with the most significant energy is located using methods such as peak search. Next, the maximum deviation direction of the waveform amplitude of this main reflected pulse from the corrected zero-level is calculated to determine its peak direction. Finally, this determined peak direction is compared with a preset reference polarity corresponding to the injected square wave pulse. If the two directions are the same, the defect type is determined to be a high-impedance defect; if the directions are opposite, it is determined to be a low-impedance defect.

[0048] S160: Based on the degradation index and defect nature information, determine the fault diagnosis information of the target load in the preset diagnostic rule knowledge base, and display the fault diagnosis information of the target load on the interactive interface.

[0049] The diagnostic rule knowledge base comprises a combination of degradation index ranges and defect nature information, along with a logical mapping relationship between these elements and one or more specific failure modes. Fault diagnostic information refers to the specific diagnostic status of the target load's health condition, indicating whether a problem exists, its root cause, and its specific type.

[0050] The degradation index and defect nature information of the target load are used as input conditions for matching in the diagnostic rule knowledge base. First, based on the defect nature information, all candidate fault modes related to this type of defect nature are filtered in the knowledge base. Then, the actual degradation index of the target load is compared with the preset degradation index threshold range corresponding to each candidate fault mode. Finally, the candidate fault mode whose threshold range covers the current degradation index is selected as the final fault diagnosis information. Finally, the specific fault diagnosis information is presented to the maintenance personnel in a pop-up window on the interactive interface. For example, the diagnostic result for target load PUMP-02: cable insulation is damp, degradation index 65, defect nature is low impedance type.

[0051] This embodiment performs multi-scale time-frequency decomposition on the high-frequency waveform data of the main bus of the substation auxiliary power supply system to generate a two-dimensional time-frequency spectrum that can comprehensively characterize the transient process of load start-up and shutdown. The spectrum is then compared with a preset template to generate a degradation index, overcoming the shortcomings of existing technologies that only focus on steady-state spectrum and ignore transient information. By triggering physical detection of the target load through user interaction, and fusing the degradation index reflecting its functional status with the defect information reflecting the physical status of its power supply cable, functional abnormalities and physical defects can be effectively correlated, thereby forming a complete chain of diagnostic evidence and improving the accuracy of fault root cause tracing and the reliability of diagnostic conclusions in complex scenarios.

[0052] In one feasible implementation, step S130: performing a structural similarity comparison between the two-dimensional time-frequency spectrum and a preset time-frequency spectrum template to generate a degradation index for each load, including: S131: Divide the two-dimensional time-frequency spectrum along the frequency axis into an electrical characteristic time-frequency spectrum and a mechanical characteristic time-frequency spectrum. The electrical characteristic time-frequency spectrum covers the power supply fundamental frequency of the auxiliary power supply system and the preset integer multiple harmonic frequency bands. The mechanical characteristic time-frequency spectrum covers the non-harmonic and high-frequency bands other than the electrical characteristic time-frequency spectrum.

[0053] Electrical characteristic time-frequency spectrum refers to a sub-spectrum that represents the state of the electrical parts of a device, which can include windings and power supplies, etc.

[0054] Electrical faults in equipment typically manifest as abnormalities in power frequency and harmonic components, while mechanical faults generally present as vibrations or friction, manifesting in non-harmonic and higher frequency signal components. The two-dimensional time-frequency spectrum generated by S120 is logically segmented along the frequency axis. Based on preset frequency bands, for example, rows containing integer multiples of harmonic frequency bands such as the power supply fundamental frequency of 50 Hz, third harmonic of 150 Hz, and fifth harmonic of 250 Hz are extracted and combined to form an electrical characteristic time-frequency spectrum. Simultaneously, all other rows of data in the two-dimensional time-frequency spectrum, excluding these harmonic frequency bands, especially the higher frequency portions, are combined to form a mechanical characteristic time-frequency spectrum.

[0055] S132: Extract the electrical template map and mechanical template map corresponding to the electrical feature time-frequency map and mechanical feature time-frequency map from the preset time-frequency map template.

[0056] The electrical template map and the mechanical template map are two sub-template maps obtained by dividing the preset time-frequency map template representing the health status according to the same frequency band division standard as S131.

[0057] Retrieve the health time-frequency spectrum template corresponding to the current two-dimensional time-frequency spectrum from the database. Then, according to the preset harmonic and non-harmonic frequency band division rules, that is, the same frequency band division standard as S131, divide this complete health template spectrum into an electrical template spectrum and a mechanical template spectrum.

[0058] S133: Calculate the first similarity score between the electrical feature time-frequency spectrum and the electrical template spectrum, and the second similarity score between the mechanical feature time-frequency spectrum and the mechanical template spectrum.

[0059] The first similarity score and the second similarity score are two independent quantification values, used to represent the degree of similarity between the electrical and mechanical parts of the current load and an ideal healthy state.

[0060] Two structural similarity (SSIM) quantification calculations were performed based on the time-frequency maps of electrical features and electrical template maps, and on the time-frequency maps of mechanical features and mechanical template maps, respectively. SSIM is a commonly used index to measure the similarity between two images, with output values ​​ranging from 0 to 1, where a value closer to 1 indicates a higher similarity. First, the SSIM between the time-frequency maps of electrical features and electrical template maps was calculated, yielding a value as the first similarity score; let's assume the value is 0.92. Second, the SSIM between the time-frequency maps of mechanical features and mechanical template maps was calculated in the same way, yielding a second value as the second similarity score; let's assume the value is 0.75.

[0061] S134: Convert the first similarity score and the second similarity score into electrical deterioration index and mechanical deterioration index respectively, and obtain the deterioration index by weighted summation based on the electrical deterioration index and the mechanical deterioration index.

[0062] First, the two similarity scores obtained in the previous step, the first similarity score representing the electrical part and the second similarity score representing the mechanical part, are transformed by applying a preset nonlinear mapping function. The nonlinear mapping function is specifically represented as shown in formula (1).

[0063] (1) Where D is the output degradation index, S is the input similarity score, and α is a nonlinear adjustment factor greater than 1, for example, α equals 2. Formula (1) can nonlinearly map the similarity score ranging from 0 to 1 to the degradation index range of 0 to 100. Through this function, the first similarity score of 0.92 and the second similarity score of 0.75 are converted to obtain an electrical degradation index of 15.36 and a mechanical degradation index of 43.75, respectively. Then, based on the preset weight coefficients, the electrical degradation index and the mechanical degradation index are weighted and summed. Specifically, the process of calculating the degradation index is shown in formula (2).

[0064] (2) in, The comprehensive degradation index is a degradation index obtained by weighted summation of the electrical degradation index and the mechanical degradation index. The electrical degradation index, This is the mechanical deterioration index. and These are the preset electrical weights and mechanical weights, for example... It is 0.4. The value is 0.6. Finally, the total degradation index is calculated to be 32.39 using formula (2).

[0065] This embodiment divides the two-dimensional time-frequency spectrum along the frequency axis into electrical characteristic time-frequency spectrum and mechanical characteristic time-frequency spectrum, and performs similarity comparisons on them respectively. This allows the comprehensive degradation state to be decomposed into electrical degradation index and mechanical degradation index with different sources. This decoupling of electromechanical state makes the diagnosis no longer limited to judging whether there is an abnormality, but can further distinguish whether the physical root cause of the abnormality is more inclined to electrical circuit problems or mechanical structure problems. Therefore, it can provide deeper diagnostic information and improve the precision of diagnosis of complex electromechanical coupling faults.

[0066] In one feasible implementation, the method further includes: Based on the reflected waveform data, the round-trip time of the preset square wave pulse in the branch cable corresponding to the target load is calculated to determine the defect location information of the branch cable corresponding to the target load.

[0067] Round-trip time refers to the total time it takes for the injected preset square wave pulse to travel from the injection point to the impedance discontinuity point in the cable, be reflected back, and be captured again by the data acquisition unit. Defect location information is a numerical value representing the physical length of the impedance discontinuity point from the cable injection end; it is a key parameter for spatially locating the fault point.

[0068] The start time of the injected pulse and the arrival time of the first significant reflected pulse are identified from the reflected waveform data. The round-trip time of the pulse is obtained by calculating the time difference between these two times. Then, the corresponding electromagnetic wave propagation speed, such as the speed of light in a specific medium, is retrieved from a database. Finally, by multiplying the round-trip time by the electromagnetic wave propagation speed and dividing by two, the distance from the defect point to the measuring end can be calculated. This distance value is the final determined defect location information.

[0069] Based on the defect location information and the preset cable network topology data of the auxiliary power system, the target associated load is identified in the cable network topology data that is in the same branch cable as the target load and is far away from the main bus of the auxiliary power system relative to the target load.

[0070] Cable network topology data refers to a graph database that represents the physical connections and wiring paths between all cables, junction boxes, and loads in an auxiliary power system.

[0071] First, the pre-stored cable network topology data of the entire auxiliary power system is retrieved. Then, the defect location information is compared with the installation location of the target load to determine whether the defect is upstream or downstream of the target load. Following the cable routing path, starting from the target load location, the graph database of the cable network topology data is used to traverse and search in the direction away from the main busbar, identifying all other load devices on this path and forming a list of target-related loads.

[0072] The range definition information is generated by calculating the absolute value of the difference between the degradation index of the target load and the degradation index of the target associated load, and comparing the absolute value of the difference with a preset threshold.

[0073] Scope definition information can include localized faults and linear faults, indicating whether a defect affects only a single device or the entire branch. The absolute value of the difference between the degradation index of the target load and the degradation index of each associated target load is calculated. These calculated absolute values ​​are then compared to a preset threshold, for example, set to 20. If all absolute values ​​are greater than this threshold, it indicates that only the target load is in poor condition while other associated loads are in good condition, and the scope definition is a localized fault. Conversely, if one or more absolute values ​​are less than the threshold, it indicates that an associated load is in a similarly poor condition as the target load, and the scope definition is a linear fault.

[0074] For example, the round-trip time is calculated to be 0.2 microseconds based on the reflected waveform data, and the defect location is determined to be 20 meters based on the electromagnetic wave propagation speed of the cable. Next, the cable network topology data is retrieved, and the installation location of the target load is found to be 18 meters from the injection point. Based on topology path traversal, the fan located 35 meters downstream is identified as the target associated load. Given that the degradation index of the target load is 65 and the degradation index of the target associated load is 18, the absolute value of the difference between the two degradation indices is calculated, yielding a result of 47. This difference of 47 is compared to a preset threshold of 20. Since 47 is greater than 20, it is determined to be a localized fault, which is reflected in the fault diagnosis report to alert maintenance personnel.

[0075] In one feasible implementation, step S120: Based on the load start-stop event markers of the running remote signaling data, locate the start-stop event window of each load in the high-frequency waveform data, and perform multi-scale time-frequency decomposition on the high-frequency waveform data within the start-stop event window to generate a two-dimensional time-frequency spectrum, including: S121: Determine the start / stop event time points of the load start / stop event markers for each load in the running remote signaling data.

[0076] For each remote signaling data record that serves as a marker for a load start / stop event, the timestamp field is parsed, and the timestamp value is extracted. For example, for the start event of oil pump PUMP-02, the start / stop event time is determined to be 10:00:00.123 from its remote signaling data record "Device ID: PUMP-02, Status: 1, Timestamp: 10:00:00.123".

[0077] S122: Based on the start and stop event time points, determine the start and end boundaries of the start and stop event window for each load according to the preset time offset.

[0078] The time offset is a pair of time parameters that are pre-defined for different types or models of loads. It includes a start offset and an end offset, which are used to define the time range extending forward and backward from the start and stop event time.

[0079] First, based on the currently processed load device ID, the specific time offset corresponding to that load is retrieved from a preset device parameter configuration library. Assume the retrieved start time offset is (-0.1 seconds, +2.0 seconds). Then, the determined start / stop event time point 10:00:00.123 is subtracted by the start offset of 0.1 seconds and added by the end offset of 2.0 seconds, respectively, to calculate the start boundary of the start / stop event window as 10:00:00.023 and the end boundary as 10:00:02.123.

[0080] S123: In high-frequency waveform data, extract waveform data segments corresponding to the time period based on the start and end boundaries.

[0081] Using the start boundary 10:00:00.023 and end boundary 10:00:02.123 of the start / stop event window, the system searches and locates the data within the already time-synchronized high-frequency waveform data. It then extracts all continuous voltage and current samples from the original sampled value sequence whose timestamps fall within this 2.1-second interval, forming an independent waveform data segment. This segment includes continuous voltage and current samples from the moment before startup to the point of complete stable operation of the load device.

[0082] S124: Divide the waveform data segment into multiple time scales, calculate the frequency distribution at each time scale, and combine them to form a two-dimensional time-frequency spectrum.

[0083] A time-frequency analysis method can be used to analyze the waveform data segment extracted in the previous step. For example, an analysis window function with 256 sampling points and a 50% overlap can be used, sliding along the time axis in small steps, starting from the beginning of the waveform data segment. At each sliding window position, a Fast Fourier Transform is performed on the data within the window to calculate the frequency distribution at that moment. By arranging and combining the frequency distribution results of all time positions in chronological order, a two-dimensional energy distribution map is formed. This map is the final two-dimensional time-frequency map representing the target load's startup process.

[0084] In one feasible implementation, step S140, controlling the auxiliary power system to inject a preset square wave pulse into the branch cable corresponding to the target load and acquiring reflected waveform data, includes: S141: Send a status query command to the auxiliary power system to obtain the operating status of the target load.

[0085] Before officially sending the pulse detection command, the cloud computing platform first performs a security check. For example, for the target load oil pump PUMP-02, the device ID of PUMP-02 is encapsulated into a status query command and sent to the auxiliary power system monitoring unit of the substation via the network. After receiving the command, the field monitoring unit queries its local real-time database to obtain the current operating status of PUMP-02 as stopped, and then sends this status information back to the cloud computing platform.

[0086] S142: When the target load is in standby or light load condition, a synchronization trigger signal is sent to the signal generation unit and data acquisition unit in the auxiliary power system. The synchronization trigger signal is used to drive the signal generation unit to generate an injection signal based on a preset square wave pulse and inject it into the branch cable corresponding to the target load. At the same time, the data acquisition unit is driven to record the mixed waveform including the injection signal and the reflected signal.

[0087] The signal generation unit and data acquisition unit are hardware devices deployed at the substation site, responsible for generating detection signals and recording response signals, respectively. The hybrid waveform is a composite waveform that is actually measured on the cable after the detection signal is injected, containing the original injected signal, multiple reflected signals, and background noise.

[0088] Upon receiving feedback that PUMP-02 is in a stopped state and confirming that standby or light-load conditions are met, the substation main control unit immediately generates a synchronization trigger signal and simultaneously sends it to the signal generation unit and data acquisition unit. Upon receiving the signal, the signal generation unit generates a preset square wave pulse as an injection signal, which is then injected into the PUMP-02's cable via a coupling device. The preset square wave pulse amplitude can be 5 volts and the pulse width can be 100 nanoseconds. Simultaneously, the data acquisition unit begins recording the voltage waveform at the injection point at a high sampling rate to form a mixed waveform, until a preset recording duration sufficient to capture the farthest reflection ends.

[0089] S143: The reflected waveform data is separated from the mixed waveform by a bandpass filter. The passband range of the bandpass filter covers the fundamental frequency and third harmonic frequency band of the preset square wave pulse.

[0090] The data acquisition unit processes the recorded mixed waveform. First, based on the injected 100-nanosecond pulse width of the preset square wave, it calculates that the spectral energy is mainly concentrated in a frequency band with a fundamental frequency of 10 MHz. Then, a digital bandpass filter with a passband range of, for example, 5 to 35 MHz is designed, and the mixed waveform is passed through this filter. Since the spectral characteristics of the reflected signal are highly consistent with the injected signal, and the frequencies of signals such as power frequency interference and high-frequency white noise are outside this range, most of the noise is filtered out, resulting in a relatively clean reflected waveform data with improved signal-to-noise ratio. This data, along with the device ID, is then transmitted back to the cloud computing platform.

[0091] In one feasible implementation, step S150: by determining the polarity information of the reflected waveform data, the defect nature information of the branch cable corresponding to the target load is determined, including: S151: Perform zero-level correction on the reflected waveform data and locate the first main reflected pulse in the corrected reflected waveform.

[0092] First, the average amplitude of the received reflected waveform data during the no-signal period is calculated, and this average value is subtracted from all sampling points of the entire waveform data as a DC bias to complete zero-point level correction. Then, a peak search algorithm is applied to the corrected waveform data, and the pulse that appears earliest in time and whose absolute amplitude exceeds a preset noise threshold is identified as the first master reflected pulse. For example, for the received reflected waveform data of the oil pump PUMP-02, a DC bias of +0.05 volts is calculated at the baseline. After correction, the first pulse with an absolute amplitude exceeding the 0.1 volt noise threshold is determined as the master reflected pulse through peak search.

[0093] S152: Determine the peak direction of the primary reflected pulse by calculating the maximum deviation direction of the waveform amplitude of the primary reflected pulse relative to the zero-point level.

[0094] The maximum deviation direction refers to whether the peak of the main reflected pulse points towards a positive or negative amplitude relative to the corrected zero-point level.

[0095] The waveform data segment of the located main reflected pulse is analyzed. All sampling points within this segment are traversed, and the absolute values ​​of its maximum positive and maximum negative amplitudes are identified. By comparing these two values, the maximum deviation direction is determined, thus identifying the peak direction. For example, analyzing the main reflected pulse of the target load, its maximum positive amplitude is calculated to be +0.15 volts, and its maximum negative amplitude is -0.85 volts. Since the absolute value of -0.85 is greater than the absolute value of +0.15, the maximum deviation direction is determined to be negative, and the peak direction is recorded as negative.

[0096] S153: By comparing the peak direction with the reference polarity of the preset square wave pulse, defect nature information is obtained, including high impedance type defects and low impedance type defects.

[0097] The reference polarity corresponding to the preset square wave pulse used in this detection is read from the configuration library, and the peak direction of the main reflected pulse determined in the previous step is logically compared with this reference polarity. If they are the same, the defect nature information is determined to be a high-impedance defect; if they are opposite, it is determined to be a low-impedance defect. For example, the reference polarity read in this detection is positive. Since the peak direction of the target load determined in the previous step is negative, the two are opposite, therefore, the final defect nature information is determined to be a low-impedance defect.

[0098] In one feasible implementation, step S160: Based on the degradation index and defect nature information, determine the fault diagnosis information of the target load in a preset diagnostic rule knowledge base, including: S161: Based on the defect nature information, a subset of candidate fault modes is selected from the preset diagnostic rule knowledge base.

[0099] A diagnostic rule knowledge base is a combination of degradation index ranges and defect nature information, along with logical mapping relationships between them and one or more specific failure modes. Table 1 provides an example of a diagnostic rule knowledge base.

[0100]

[0101] As shown in Table 1, each row in this table represents an independent diagnostic rule. The rule ID is a unique identifier for each rule. The defect nature information column contains keywords for initial screening; faults are broadly categorized into high-impedance and low-impedance types. The fault mode column provides a detailed description of the specific fault cause, which is the final diagnostic result. The degradation index threshold range column sets the quantitative judgment criteria for each fault mode.

[0102] The defect nature information is used as the search keyword to perform a query and match in the diagnostic rule knowledge base shown in Table 1. The knowledge base pre-stores the association between defect nature and fault modes. By searching, all fault modes associated with the current defect nature information are extracted to form a subset of candidate fault modes. For example, when the defect nature information is a low-impedance defect, the subset of candidate fault modes selected by searching Table 1 includes: cable insulation dampness, cable metallic short circuit, and motor winding inter-turn short circuit.

[0103] S162: By comparing the degradation index with the degradation index threshold range corresponding to each candidate failure mode, the target failure mode is determined to obtain failure diagnosis information.

[0104] The degradation index threshold range is a numerical range that is pre-defined for each specific failure mode, as shown in Table 1. When such a failure occurs, its corresponding degradation index usually falls into the range.

[0105] The algorithm iterates through each candidate fault mode in the subset of candidate fault modes selected in the previous step. For each candidate fault mode, its corresponding degradation index threshold range is obtained from the diagnostic rule knowledge base. Then, the calculated actual degradation index of the target load is compared with this threshold range to determine whether it falls within the range. Finally, the candidate fault mode whose threshold range successfully covers the current degradation index is selected as the target fault mode and used as the final fault diagnosis information. For example, for the subset of candidate fault modes selected in the previous step, the corresponding degradation index threshold ranges are obtained from Table 1, which are [40, 70], [90, 100], and [75, 95]. Subsequently, the actual degradation index 65 of the target load oil pump is compared with these three ranges one by one, and it is determined that 65 only falls within the range [40, 70]. The fault mode corresponding to this range is cable insulation dampness. Therefore, the target fault mode of the target load oil pump is finally determined to be cable insulation dampness.

[0106] Based on the same concept, this application provides an intelligent interactive diagnostic system based on the fusion of multi-dimensional information cloud computing power from substations. The following is a detailed description... Figure 4 This application provides a detailed description of the intelligent interactive diagnostic system based on the fusion of multi-dimensional information cloud computing power from substations, as provided in the embodiments of this application.

[0107] Figure 4 This is a structural block diagram of an intelligent interactive diagnostic system based on the fusion of multi-dimensional information cloud computing power from substations, as shown in an embodiment of this application.

[0108] like Figure 4 As shown, this intelligent interactive diagnostic system based on the fusion of multi-dimensional information from substations and cloud computing power is applied to a cloud computing power platform. The system may include: The receiving module 410 is used to receive high-frequency waveform data of the main bus of the auxiliary power system sent by the auxiliary power system of the substation, as well as remote signaling data of multiple loads in the auxiliary power system. The generation module 420 is used to locate the start and stop event window of each load in the high-frequency waveform data according to the load start and stop event markers of the running remote signaling data, and to perform multi-scale time-frequency decomposition on the high-frequency waveform data within the start and stop event window to generate a two-dimensional time-frequency spectrum. The generation module 420 is also used to compare the structural similarity of the two-dimensional time-frequency spectrum with the preset time-frequency spectrum template, generate the degradation index of each load, and display the degradation index of each load on the interactive interface. The receiving module 410 is also used to receive a pulse detection command input by the user to detect the target load selected by the user, send the pulse detection command to the auxiliary power system, and receive the reflected waveform data of the target load sent by the auxiliary power system. The pulse detection command is used to control the auxiliary power system to inject a preset square wave pulse into the branch cable corresponding to the target load and collect the reflected waveform data. The determination module 430 is used to determine the defect nature information of the branch cable corresponding to the target load by determining the polarity information of the reflected waveform data; The display module 440 is used to determine the fault diagnosis information of the target load in a preset diagnostic rule knowledge base based on the degradation index and defect nature information, and to display the fault diagnosis information of the target load on the interactive interface.

[0109] In one embodiment, the generation module 420 is specifically used to divide the two-dimensional time-frequency spectrum along the frequency axis into an electrical characteristic time-frequency spectrum and a mechanical characteristic time-frequency spectrum. The electrical characteristic time-frequency spectrum covers the power supply fundamental frequency of the auxiliary power supply system and a preset integer multiple harmonic frequency band, while the mechanical characteristic time-frequency spectrum covers the non-harmonic and high-frequency bands other than the electrical characteristic time-frequency spectrum. From the preset time-frequency spectrum template, the module extracts the electrical template spectrum and the mechanical template spectrum corresponding to the electrical characteristic time-frequency spectrum and the mechanical characteristic time-frequency spectrum. The module calculates the first similarity score between the electrical characteristic time-frequency spectrum and the electrical template spectrum, and the second similarity score between the mechanical characteristic time-frequency spectrum and the mechanical template spectrum. The module converts the first similarity score and the second similarity score into an electrical degradation index and a mechanical degradation index, respectively, and obtains the degradation index by weighted summation based on the electrical degradation index and the mechanical degradation index.

[0110] In one embodiment, the generation module 420 is further configured to calculate the round-trip time of a preset square wave pulse in the branch cable corresponding to the target load based on the reflected waveform data, and determine the defect location information of the branch cable corresponding to the target load; based on the defect location information and the preset cable network topology data of the auxiliary power system, determine the target associated load in the cable network topology data that is in the same branch cable as the target load and is far away from the main bus of the auxiliary power system relative to the target load; and generate range definition information by calculating the absolute value of the difference between the degradation index of the target load and the degradation index of the target associated load and comparing the absolute value of the difference with a preset threshold.

[0111] In one embodiment, the generation module 420 is specifically used to determine the start and stop event time points of the load start and stop event markers for each load in the running remote signaling data; based on the start and stop event time points, determine the start and end boundaries of the start and stop event window for each load according to a preset time offset; in the high-frequency waveform data, extract waveform data segments of the corresponding time period according to the start and end boundaries; divide the waveform data segments into multiple time scales, calculate the frequency distribution at each time scale, and combine them to form a two-dimensional time-frequency spectrum.

[0112] In one embodiment, the receiving module 410 is further configured to send a status query command to the auxiliary power system to obtain the operating status of the target load; when the target load is in standby or light load condition, it sends a synchronization trigger signal to the signal generation unit and data acquisition unit in the auxiliary power system. The synchronization trigger signal is used to drive the signal generation unit to generate an injection signal based on a preset square wave pulse and inject it into the branch cable corresponding to the target load, while driving the data acquisition unit to record a mixed waveform including the injection signal and the reflected signal; the reflected waveform data is separated from the mixed waveform by a bandpass filter, and the passband range of the bandpass filter covers the fundamental frequency and third harmonic frequency band of the preset square wave pulse.

[0113] In one embodiment, the determining module 430 is specifically used to perform zero-level correction on the reflected waveform data and locate the first main reflected pulse in the corrected reflected waveform; determine the peak direction of the main reflected pulse by calculating the maximum deviation direction of the waveform amplitude of the main reflected pulse relative to the zero-level; and obtain defect nature information by comparing the peak direction with the reference polarity of a preset square wave pulse. The defect nature information includes high-impedance defects and low-impedance defects.

[0114] In one embodiment, the display module 440 is specifically used to filter a subset of candidate fault modes from a preset diagnostic rule knowledge base based on defect nature information; and to determine the target fault mode by comparing the degradation index with the degradation index threshold range corresponding to each candidate fault mode to obtain fault diagnosis information.

[0115] Figure 4 Each module in the system shown has an implementation Figures 1 to 3 The functions of each step in the process and their corresponding technical effects are described in detail here for the sake of brevity.

[0116] Figure 5 A schematic diagram of the hardware structure of an electronic device provided in one embodiment of this application is shown.

[0117] The electronic device may include a processor 510 and a memory 520 storing computer program instructions.

[0118] Specifically, the processor 510 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0119] Memory 520 may include mass storage for data or instructions. For example, and not limitingly, memory 520 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 520 may include removable or non-removable (or fixed) media. Where appropriate, memory 520 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 520 is non-volatile solid-state memory.

[0120] Memory may include read-only memory (ROM), random access memory (RAM), disk storage media devices, optical storage media devices, flash memory devices, and electrical, optical, or other physical / tangible memory storage devices. Therefore, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to the first aspect of this disclosure.

[0121] The processor 510 reads and executes computer program instructions stored in the memory 520 to implement any of the intelligent interactive diagnostic methods based on the fusion of multi-dimensional information cloud computing power of substations in the above embodiments.

[0122] In one example, the electronic device may also include a communication interface 530 and a bus 540. Wherein, such as Figure 5 As shown, the processor 510, memory 520, and communication interface 530 are connected through bus 540 and complete communication with each other.

[0123] The communication interface 530 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0124] Bus 540 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 540 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0125] This electronic device can execute the intelligent interactive diagnostic method based on the fusion of multi-dimensional information cloud computing power from substations as described in this application embodiment, thereby achieving a combination of... Figures 1 to 3 This paper describes an intelligent interactive diagnostic method based on the fusion of multi-dimensional information from substations and cloud computing power.

[0126] Furthermore, in conjunction with the intelligent interactive diagnostic method based on the fusion of multi-dimensional information cloud computing power of substations in the above embodiments, this application embodiment can provide a computer-readable storage medium for implementation. This computer-readable storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the intelligent interactive diagnostic methods based on the fusion of multi-dimensional information cloud computing power of substations in the above embodiments.

[0127] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0128] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0129] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0130] The aspects of this application have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block in 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, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0131] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. An intelligent interactive diagnosis method based on substation multi-dimensional information cloud computing power fusion, applied to a cloud computing power platform, characterized in that, The method includes: Receive high-frequency waveform data of the main bus of the auxiliary power system sent by the auxiliary power system of the substation, as well as remote signaling data of the operation of multiple loads in the auxiliary power system; Based on the load start-stop event markers of the remote signaling data, the start-stop event window of each load is located in the high-frequency waveform data, and the high-frequency waveform data is decomposed into a multi-scale time-frequency spectrum within the start-stop event window. The two-dimensional time-frequency spectrum is compared with a preset time-frequency spectrum template for structural similarity to generate a degradation index for each load, and the degradation index for each load is displayed on the interactive interface. The system receives a pulse detection command input by the user to detect the target load selected by the user, sends the pulse detection command to the auxiliary power system, and receives the reflected waveform data of the target load sent by the auxiliary power system. The pulse detection command is used to control the auxiliary power system to inject a preset square wave pulse into the branch cable corresponding to the target load and collect the reflected waveform data. By determining the polarity information of the reflected waveform data, the defect nature information of the branch cable corresponding to the target load can be determined; Based on the degradation index and the defect nature information, the fault diagnosis information of the target load is determined in the preset diagnostic rule knowledge base, and the fault diagnosis information of the target load is displayed on the interactive interface.

2. The method of claim 1, wherein, The step of comparing the two-dimensional time-frequency spectrum with a preset time-frequency spectrum template for structural similarity to generate a degradation index for each load includes: The two-dimensional time-frequency spectrum is divided along the frequency axis into an electrical characteristic time-frequency spectrum and a mechanical characteristic time-frequency spectrum. The electrical characteristic time-frequency spectrum covers the power supply fundamental frequency and the preset integer multiple harmonic frequency band of the auxiliary power supply system, and the mechanical characteristic time-frequency spectrum covers the non-harmonic and high frequency bands other than the electrical characteristic time-frequency spectrum. From the preset time-frequency spectrum template, extract the electrical template spectrum and the mechanical template spectrum corresponding to the electrical feature time-frequency spectrum and the mechanical feature time-frequency spectrum; Calculate the first similarity score between the electrical feature time-frequency spectrum and the electrical template spectrum, and the second similarity score between the mechanical feature time-frequency spectrum and the mechanical template spectrum, respectively. The first similarity score and the second similarity score are converted into an electrical degradation index and a mechanical degradation index, respectively, and the degradation index is obtained by weighted summation based on the electrical degradation index and the mechanical degradation index.

3. The method of claim 2, wherein, The method further includes: Based on the reflected waveform data, the round-trip time of the preset square wave pulse in the branch cable corresponding to the target load is calculated to determine the defect location information of the branch cable corresponding to the target load. Based on the defect location information and the preset cable network topology data of the auxiliary power system, a target associated load that is in the same branch cable as the target load and is far away from the main bus of the auxiliary power system relative to the target load is determined from the cable network topology data. By calculating the absolute value of the difference between the degradation index of the target load and the degradation index of the target associated load, and comparing the absolute value of the difference with a preset threshold, range definition information is generated.

4. The method of claim 1, wherein, The step of locating the start / stop event window of each load in the high-frequency waveform data based on the load start / stop event markers of the remote signaling data, and performing multi-scale time-frequency decomposition on the high-frequency waveform data within the start / stop event window to generate a two-dimensional time-frequency spectrum includes: The start / stop event time points marked by the load start / stop event for each load are determined in the operational remote signaling data; Based on the start and stop event time points, the start and end boundaries of the start and stop event windows for each load are determined according to a preset time offset. In the high-frequency waveform data, waveform data segments corresponding to the time period are extracted according to the start boundary and the end boundary; The waveform data segment is divided into multiple time scales, and the frequency distribution is calculated at each time scale, which are then combined to form a two-dimensional time-frequency spectrum.

5. The method of claim 1, wherein, The process of controlling the auxiliary power system to inject a preset square wave pulse into the branch cable corresponding to the target load and acquiring the reflected waveform data includes: Send a status query command to the auxiliary power system to obtain the operating status of the target load; When the target load is in standby or light load condition, a synchronization trigger signal is sent to the signal generation unit and data acquisition unit in the auxiliary power system. The synchronization trigger signal is used to drive the signal generation unit to generate an injection signal based on the preset square wave pulse and inject it into the branch cable corresponding to the target load. At the same time, it drives the data acquisition unit to record the mixed waveform including the injection signal and the reflected signal. The reflected waveform data is separated from the mixed waveform by a bandpass filter, and the passband range of the bandpass filter covers the fundamental frequency and third harmonic frequency band of the preset square wave pulse.

6. The method of claim 1, wherein, The step of determining the defect nature information of the branch cable corresponding to the target load by determining the polarity information of the reflected waveform data includes: Zero-point level correction is performed on the reflected waveform data, and the first main reflected pulse is located in the corrected reflected waveform; The peak direction of the main reflected pulse is determined by calculating the maximum deviation direction of the waveform amplitude of the main reflected pulse relative to the zero-point level. The defect nature information is obtained by comparing the peak direction with the reference polarity of the preset square wave pulse. The defect nature information includes high-impedance defects and low-impedance defects.

7. The method of claim 1, wherein, The step of determining the fault diagnosis information of the target load in a preset diagnostic rule knowledge base based on the degradation index and the defect nature information includes: Based on the defect nature information, a subset of candidate fault modes is selected from the preset diagnostic rule knowledge base; The target fault mode is determined by comparing the degradation index with the degradation index threshold range corresponding to each candidate fault mode to obtain the fault diagnosis information.

8. An intelligent interactive diagnostic system based on the fusion of multi-dimensional information and cloud computing power from substations, applied to a cloud computing platform, characterized in that... The system includes: The receiving module is used to receive high-frequency waveform data of the main bus of the auxiliary power system sent by the auxiliary power system of the substation, as well as the remote signaling data of the operation of multiple loads in the auxiliary power system. The generation module is used to locate the start-stop event window of each load in the high-frequency waveform data according to the load start-stop event marker of the running remote signaling data, and to perform multi-scale time-frequency decomposition on the high-frequency waveform data within the start-stop event window to generate a two-dimensional time-frequency spectrum. The generation module is also used to perform structural similarity comparison between the two-dimensional time-frequency spectrum and the preset time-frequency spectrum template, generate the degradation index of each load, and display the degradation index of each load on the interactive interface. The receiving module is also used to receive a pulse detection command input by the user to detect the target load selected by the user, send the pulse detection command to the auxiliary power system, and receive the reflected waveform data of the target load sent by the auxiliary power system. The pulse detection command is used to control the auxiliary power system to inject a preset square wave pulse into the branch cable corresponding to the target load and collect the reflected waveform data. The determination module is used to determine the defect nature information of the branch cable corresponding to the target load by determining the polarity information of the reflected waveform data; The display module is used to determine the fault diagnosis information of the target load in a preset diagnostic rule knowledge base based on the degradation index and the defect nature information, and to display the fault diagnosis information of the target load on the interactive interface.

9. An electronic device, characterized in that, The device includes: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, it implements the intelligent interactive diagnostic method based on the fusion of multi-dimensional information cloud computing power of substations as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions, which, when executed by a processor, implement the intelligent interactive diagnostic method based on the fusion of multi-dimensional information cloud computing power of substations as described in any one of claims 1-7.

Images