Multi-point data flow overload processing method and system based on synchronous display

Abstract

The invention relates to the technical field of power equipment online monitoring and cloud computing technology crossing, in particular to a synchronous display-based multipoint data traffic overload processing method and system, and the method comprises the steps: receiving original monitoring data packets of a plurality of monitoring points; synchronously aligning to the same power frequency periodic coordinate system in real time so as to generate a global discharge event data set with time-space synchronization; performing parallel and batch calculation on the global discharge event data set to separate out partial discharge pulses, and determining synchronous phase information based on the partial discharge pulses to generate discharge diagnosis maps corresponding to the plurality of monitoring points; and constructing a composite index for the structured monitoring data, the feature data and the discharge diagnosis map according to a preset identifier, responding to a display request of a client, generating a synchronous display data packet, and displaying the synchronous display data packet based on the client. All monitoring data are uniformly stored in the cloud database, data fragmentation is avoided, and historical data comparison, trend analysis and deep data mining are facilitated.

Description

Technical Field

[0001] This application relates to the technical field of the intersection of online monitoring of power equipment and cloud computing technology, and in particular to a method and system for handling overloaded multi-point data traffic based on synchronous display. Background Technology

[0002] Power cables are an important component of the power grid, and their insulation condition directly affects the reliability of power supply. Partial discharge is one of the main signs of cable insulation deterioration, and the high-frequency pulse current method (HFCT) is a commonly used and effective method for detecting partial discharge.

[0003] Currently, many on-site partial discharge detection methods involve personnel at the detection point carrying laptops and running specialized software for data collection and preliminary analysis. This method has significant drawbacks:

[0004] Data fragmentation: Historical data is scattered across different laptops, making it difficult to manage in a unified manner, compare historical trends, and conduct comprehensive analysis, which is not conducive to managers having a full grasp of the cable insulation status.

[0005] Limited processing power: Laptops have limited computing and storage resources, making it difficult to handle real-time processing of multi-channel, long-term, high-sampling-rate data, let alone process data from multiple monitoring points simultaneously.

[0006] Poor scalability: When it is necessary to add monitoring points or perform more complex analyses (such as AI diagnostics), upgrading the hardware of laptops is difficult, costly, and cannot achieve flexible resource allocation.

[0007] Access is inconvenient: The analysis results are limited to laptops with specific software installed, making it impossible to access them conveniently anytime, anywhere, across multiple devices.

[0008] While some online monitoring systems exist, their data processing models often fail to fully leverage the centralized advantages of cloud computing architecture, or they suffer from shortcomings in system scalability, standardized data management, and convenient multi-user access. Therefore, there is an urgent need for a cable partial discharge monitoring solution that can fully utilize cloud resources to achieve centralized data processing, storage, and access. Summary of the Invention

[0009] In order to fully utilize cloud resources and realize a cable partial discharge monitoring scheme for centralized data processing, storage and access, this application provides a method and system for handling multi-point data traffic overload based on synchronous display.

[0010] Firstly, this application provides a method for handling multi-point data traffic overload based on synchronous display, employing the following technical solution:

[0011] A method for handling multi-point data traffic overload based on synchronous display includes the following steps:

[0012] Establish a unified cloud data receiving gateway to receive raw monitoring data packets from multiple monitoring points. The raw monitoring data packets carry power frequency phase stamps and timestamps.

[0013] Based on the phase stamp and timestamp, the original monitoring data packets are aligned in real time and synchronously to the same power frequency cycle coordinate system to generate a spatiotemporally synchronized global discharge event dataset.

[0014] The global discharge event dataset is processed in parallel and in batches to separate partial discharge pulses, and the synchronization phase information is determined based on the partial discharge pulses to generate discharge diagnostic maps corresponding to multiple monitoring points.

[0015] A composite index is constructed based on the structured monitoring data, feature data, and discharge diagnostic spectrum according to the preset identifier. The system responds to the client's display request and, based on the composite index, retrieves and assembles data and spectrums of multiple specified monitoring points within a selected time window and phase interval in real time to generate a synchronous display data package and display it on the client.

[0016] By adopting the above technical solution, all monitoring data is stored uniformly in the cloud database, avoiding data fragmentation and facilitating historical data comparison, trend analysis, and in-depth data mining. The cloud architecture enables it to simultaneously receive, process, and store data from a large number of widely distributed field monitoring terminals, achieving large-scale monitoring.

[0017] In one embodiment, the discharge diagnostic spectrum includes a PRPS spectrum. Generating discharge diagnostic spectra for multiple monitoring points based on synchronized phase information includes the following steps:

[0018] The continuous time axis is divided into equal-length segments based on the preset analysis granularity, and a PRPD map is generated for the global discharge events of each equal-length segment.

[0019] The PRPD maps corresponding to each equal-length time slice are arranged sequentially according to time sequence to form a three-dimensional discharge diagnostic map. The discharge diagnostic map characterizes the discharge occurrence at a specific phase, amplitude, and time.

[0020] By adopting the above technical solution and leveraging the high performance of cloud servers, multi-channel monitoring data can be easily processed, and complex calculations (such as PRPD / PRPS map generation and AI fault diagnosis) can be performed. Resources can be flexibly expanded as needed to support more monitoring points or more complex analysis algorithms.

[0021] In one embodiment, generating a PRPD map for global discharge events of each equal-length slice includes the following steps:

[0022] Based on the partial discharge pulse, the synchronous power frequency phase and the synchronous discharge pulse amplitude are confirmed and extracted;

[0023] All global discharge events are traversed, and a PRPD spectrum is plotted with the power frequency phase as the horizontal axis and the discharge pulse amplitude as the vertical axis. The PRPD spectrum represents the frequency of discharge occurrence at a specific phase and amplitude based on the density of coordinate points.

[0024] In one embodiment, after traversing all global discharge events and plotting a PRPD graph with the power frequency phase as the abscissa and the discharge pulse amplitude as the ordinate, the following steps are further included:

[0025] Obtain the defect types corresponding to historical PRPD maps in the historical database, and use the historical PRPD maps and defect types as training datasets;

[0026] A prediction model is trained based on the training dataset, and the PRPD map is input into the preset model to obtain the corresponding defect prediction type.

[0027] A preliminary assessment is determined based on the defect prediction type to generate a diagnostic report, which is then displayed using a visual interface.

[0028] In one embodiment, the original monitoring data packet is aligned synchronously to the same power frequency cycle coordinate system in real time based on the phase stamp and timestamp, wherein the synchronous alignment specifically includes the following steps:

[0029] The phase stamps of all monitoring points are calibrated and compensated using the power frequency phase of the first arriving monitoring terminal as a reference or through unified time synchronization via a cloud server.

[0030] In one embodiment, a composite index is constructed based on the structured monitoring data, feature data, and discharge diagnostic spectrum according to a preset identifier. The composite index adopts a multi-dimensional tag index structure of a time-series database and includes at least the monitoring point ID, data timestamp, power frequency phase value, and signal feature type tag. Based on the composite index, data and spectrum within a time window and phase interval are selected according to preset combination conditions.

[0031] In one embodiment, selecting data and spectra within a time window and phase interval based on the composite index according to preset combination conditions includes the following steps:

[0032] When a data query request is received, the query conditions in the request are first parsed. The conditions include at least the target device identifier, signal type, and time range.

[0033] Using the composite tag index, a set of storage point IDs that meet the conditions of device identifier and signal type can be quickly filtered out;

[0034] Using the primary index and the metadata records, physical files whose data timestamps and query time ranges intersect can be further located from the storage point ID set;

[0035] Only the located physical files are loaded for detailed searching and data retrieval, and the data and spectra within the selected time window and phase interval are returned.

[0036] In one embodiment, responding to a client's display request and retrieving and assembling data and maps from multiple specified monitoring points within a selected time window and phase interval based on the composite index in real time, to generate a synchronized display data package and display it on the client, includes the following steps:

[0037] The synchronous display data package covers the associated data and view configuration information of multiple monitoring points;

[0038] Based on the synchronous display data packet, multiple monitoring views are rendered, and the multiple monitoring views are globally linked for scaling and panning on the time axis and phase axis.

[0039] In one embodiment, after establishing a unified cloud data receiving gateway to receive raw monitoring data packets from multiple monitoring points, the following steps are also included:

[0040] Determine whether the number of monitoring terminals or the data receiving rate of the original monitoring data packet exceeds a preset threshold;

[0041] When the number of connected monitoring terminals or the data receiving rate exceeds a preset threshold, dynamic load balancing in the cloud is activated to automatically distribute data receiving and processing tasks to multiple cloud server instances for execution.

[0042] Secondly, this application provides a multi-point data traffic overload processing system based on synchronous display, which adopts the following technical solution:

[0043] A system for handling multi-point data traffic overload based on synchronous display, executing the multi-point data traffic overload handling method based on synchronous display as described in the first aspect, includes:

[0044] The data receiving module receives raw monitoring data packets from multiple monitoring points based on a unified cloud data receiving gateway. The raw monitoring data packets carry power frequency phase stamps and timestamps.

[0045] The data preprocessing module aligns the original monitoring data packets to the same power frequency cycle coordinate system in real time and synchronously based on the phase stamp and timestamp, so as to generate a spatiotemporally synchronized global discharge event dataset.

[0046] The core data processing module performs parallel and batch calculations on the global discharge event dataset to separate partial discharge pulses and determine synchronization phase information based on the partial discharge pulses to generate discharge diagnostic maps corresponding to multiple monitoring points.

[0047] The data storage and query module constructs a composite index based on the structured monitoring data, feature data, and discharge diagnostic maps according to preset identifiers. It responds to the client's display request and, based on the composite index, retrieves and assembles data and maps of multiple specified monitoring points within a selected time window and phase interval in real time to generate a synchronous display data package and display it on the client.

[0048] In summary, this application includes at least one of the following beneficial technical effects:

[0049] 1. All monitoring data is stored uniformly in a cloud database, avoiding data fragmentation and facilitating historical data comparison, trend analysis, and in-depth data mining;

[0050] 2. Powerful processing capabilities and scalability: Leveraging the high performance of cloud servers, it can easily process multi-channel monitoring data and perform complex calculations (such as PRPD / PRPS map generation and AI fault diagnosis). Resources can be elastically expanded as needed to support more monitoring points or more complex analysis algorithms;

[0051] 3. Its cloud architecture enables it to simultaneously receive, process, and store data from a large number of widely distributed field monitoring terminals, achieving large-scale monitoring.

[0052] 4. Convenient remote access and collaboration: Authorized users can view data and analysis results anytime, anywhere through any computer or mobile browser with an internet connection, improving operation and maintenance efficiency and collaborative work capabilities.

[0053] 5. High reliability and service continuity: Cloud servers typically have redundant backup and load balancing mechanisms, which ensure data security and service continuity, and overcome the risk of single point of failure (such as laptops). Attached Figure Description

[0054] Figure 1 This is a schematic diagram of the system architecture and data flow provided in the embodiments of this application;

[0055] Figure 2 This is a block diagram of a method for handling multi-point data traffic overload based on synchronous display, provided in an embodiment of this application. Detailed Implementation

[0056] To better understand the purpose, technical solutions, and advantages of this application, it has been described and illustrated below with reference to the accompanying drawings and embodiments. However, those skilled in the art should understand that this application can be implemented without these details. In some cases, to avoid obscuring various aspects of this application due to unnecessary description, well-known methods, processes, systems, components, and / or circuits already described at a higher level will not be elaborated upon. It will be apparent to those skilled in the art that various modifications can be made to the embodiments disclosed in this application, and the general principles defined in this application can be applied to other embodiments and application scenarios without departing from the principles and scope of this application. Therefore, this application is not limited to the illustrated embodiments, but conforms to the broadest scope consistent with the scope of protection claimed in this application.

[0057] refer to Figure 1 This application discloses a method for handling multi-point data traffic overload based on synchronous display, applied to a multi-point data traffic overload handling system based on synchronous display. The system includes a monitoring terminal deployed on-site and a cloud server. The monitoring terminal collects the raw pulse current signal of partial discharge in cables using a high-frequency pulse current sensor and uploads the data to the cloud server via a 4G module, network cable, or optical fiber. After receiving the data, the cloud server utilizes its high-performance computing resources for centralized processing (such as generating PRPD and PRPS maps), analysis, and visualization, and displays the results to users via a webpage. This system constructs a three-layer architecture of "on-site monitoring terminal + cloud server + remote access terminal." The on-site monitoring terminal is responsible for signal acquisition and uploading. The cloud server, as the "intelligent brain," undertakes heavy data storage, computational analysis, and visualization tasks. Users can remotely access the results via a webpage. This architecture fully leverages the high performance, high reliability, and high scalability advantages of the cloud.

[0058] like Figure 2 As shown, the method for handling overloaded multi-point data traffic based on synchronous display includes the following steps:

[0059] S100 establishes a unified cloud data receiving gateway to receive raw monitoring data packets from multiple monitoring points.

[0060] Combination Figure 1This embodiment demonstrates the solution using a cloud server. The cloud server establishes a unified cloud data receiving gateway, enabling it to receive raw monitoring data packets collected by monitoring terminals deployed at the cable site. The raw monitoring data packets include at least the pulse amplitude, a precise timestamp, and a phase stamp synchronized with the power frequency cycle. The cloud server centrally accesses, buffers, queues, and prioritizes data streams concurrently uploaded from multiple geographically dispersed cable partial discharge monitoring terminals to handle instantaneous traffic spikes caused by concurrent data uploads from multiple points.

[0061] The monitoring terminal collects partial discharge pulse current signals through a high-frequency pulse current sensor. Within the monitoring terminal, the signal is locally conditioned, converted from analog to digital, and compressed to obtain a raw monitoring data packet containing pulse amplitude, timestamp, and power frequency phase stamp.

[0062] It should be noted that the terminal is equipped with a high-frequency pulse current sensor (HFCT), whose bandwidth and sensitivity must meet the requirements for partial discharge signal detection. The pulse current signal on the grounding wire of the sensor coupling cable is conditioned (amplified, filtered) and sampled by an ADC before being uploaded to the cloud server by the communication module. The communication module can be configured with 4G / 5G (wireless, flexible), industrial Ethernet, or fiber optic (wired, high reliability) depending on the site conditions.

[0063] S200 aligns the original monitoring data packets to the same power frequency cycle coordinate system in real time and synchronously based on phase stamps and timestamps to generate a spatiotemporally synchronized global discharge event dataset.

[0064] The cloud server parses the original monitoring data packets and, based on the high-precision power frequency phase stamp and timestamp carried in the original monitoring data packets, aligns all data to the same virtual power frequency cycle coordinate system in real time and synchronously according to the unified clock reference and phase calibration algorithm in the cloud. Under the unified power frequency cycle coordinate system, the cloud calls upon powerful computing resources (such as wavelet transform and artificial intelligence classifiers) to filter out electromagnetic interference and noise pulses in batches, and accurately identify and extract the real global discharge event dataset.

[0065] A global discharge event dataset refers to a unified, structured collection of discharge events generated on a cloud server after the raw monitoring data packets uploaded from all deployed, geographically dispersed cable monitoring terminals have undergone a series of centralized and standardized processing steps. Each discharge event is an abstract description of real partial discharge activity, and each event is described by a set of clearly defined attribute fields, facilitating batch processing, statistics, and machine learning on the cloud server.

[0066] A discharge event typically includes a unique event ID, device identifier, high-precision timestamp, standardized phase stamp, pulse amplitude, pulse characteristics, and confidence label. The unique event ID is a globally unique identifier. The device identifier clearly indicates the specific cable joint, cable segment, or sensor from which the event originated (e.g., device ID: SUB-01-JT-202). The high-precision timestamp refers to the absolute time of the event, typically with microsecond-level accuracy, calibrated by a cloud-based unified time synchronization system. The standardized phase stamp refers to the phase angle (0°-360°) of the event within a unified virtual power frequency cycle in the cloud. This is the result after calibration through time-phase synchronization processing and is crucial for comparing multi-source data. The pulse amplitude is the quantified value of the discharge intensity, usually expressed in millivolts, picocoulombs, etc. Pulse characteristics may also include derived features such as pulse width, rise time, and equivalent frequency. The confidence label is a confidence score given after cloud-based denoising algorithms determine whether the event is a genuine partial discharge.

[0067] Aligning all data to the same virtual power frequency cycle coordinate system in real time and synchronously includes the following steps: First, establish an absolute time reference and virtual power frequency signal in the cloud. The cloud server connects to a high-precision time source, such as GPS / BeiDou satellite time synchronization or Network Time Protocol (NTP / PTP) server, to establish a unified absolute time reference T_cloud with millisecond or even microsecond precision for the entire system.

[0068] Based on this absolute time reference, the cloud-based system uses a highly stable digital oscillator to synthesize an ideal virtual power frequency signal with a rated frequency (e.g., 50.00Hz). The phase Φ_virtual(t) of this signal is a continuous, deterministic function of time t.

[0069] Φ_virtual(t) = 2π * f_nominal * (t - t_0) mod 360°. Here, f_nominal is the nominal frequency, and t_0 is a phase zero reference time. This Φ_virtual(t) is the unique phase coordinate system for the entire system.

[0070] Next, the terminal acquires data and marks the local phase. At the field monitoring terminal, hardware circuits (such as zero-crossing detection circuits or phase-locked loops) extract the local true power frequency zero-crossing point from the voltage transformer (PT) signal of the cable under test. When a discharge pulse is sampled by the ADC, the terminal records the local timestamp T_local and the local relative phase Φ_local. The local timestamp T_local is the absolute time of the pulse occurrence read from the terminal's local clock. The local relative phase Φ_local is calculated by converting the time interval between the pulse and the previous local power frequency zero-crossing point into a phase angle (0°-360°) relative to the local cycle. This identifies the pulse's position within the local true power frequency cycle.

[0071] Furthermore, regarding cloud reception and time delay compensation, the terminal sends pulse data packets {T_local, Φ_local, amplitude q, ...} to the cloud via the network. Upon receiving the data packet, the cloud immediately adds a cloud-based absolute timestamp T_arrival. Although network latency (Δ = T_arrival - T_local) is uncertain and variable, its impact can be estimated or partially eliminated. For regular, periodic heartbeat packets or synchronization messages, the cloud can dynamically estimate the network latency.

[0072] More importantly, the transmission delay does not affect the core logic of phase alignment, because the phase information Φ_local has already been determined locally on the terminal. The delay only affects the absolute time when the event is "seen" by the cloud, without changing its relative position within the power frequency cycle.

[0073] Next, the key alignment calculation—mapping to the virtual power frequency phase—requires the cloud to calculate the phase Φ_aligned corresponding to the pulse in the cloud's virtual power frequency Φ_virtual(t) coordinate system after receiving {T_local,Φ_local}.

[0074] Aligning the local timestamp with the cloud timeline: First, the terminal's local time T_local needs to be calibrated to the cloud's absolute time base. This is typically done using a clock offset estimation algorithm. The cloud monitors the relationship between each terminal's T_local and T_arrival over a long period, estimating the fixed offset (Offset) and possible drift rate of the terminal's clock relative to the cloud clock. The calibrated cloud-estimated time of occurrence is: T_estimated = T_local + Offset.

[0075] Calculate the absolute phase of the virtual power frequency: Substitute the calibrated time T_estimated into the virtual power frequency phase function: Φ_virtual_at_event=Φ_virtual(T_estimated)=2π*f_nominal*(T_estimated-t_0)mod360°.

[0076] At the moment the event occurs, the virtual power frequency signal in the cloud is in phase. However, Φ_local reflects the position of the event within the local real power frequency, and there may be slight differences in frequency and initial phase between the local real power frequency and the cloud virtual power frequency. Therefore, the final aligned phase Φ_aligned cannot directly use Φ_virtual_at_event, but should adopt the approach of relative phase shifting: within a short time window (such as several power frequency cycles), the instantaneous frequencies of the terminal's local real power frequency and the cloud virtual power frequency are approximately equal.

[0077] Alignment operation: Φ_aligned = Φ_local + ΔΦ.

[0078] Here, ΔΦ is a dynamic compensation quantity. It is calculated as follows: the cloud continuously tracks a series of pulse data from the same terminal, uses Φ_local and T_estimated to inversely deduce the instantaneous phase curve of the terminal's local power frequency, compares it with the Φ_virtual(t) curve in the cloud, and estimates the phase difference ΔΦ in real time, then performs smoothing filtering. Essentially, this is a software phase-locked loop process running in the cloud for each terminal.

[0079] Finally, after the above calculations, each original pulse data packet is assigned an aligned, globally uniform phase value Φ_aligned. This discharge event is written into the global discharge event dataset in a standardized format of {Device ID, T_estimated, Φ_aligned, Amplitude q, ...}.

[0080] The S300 performs parallel and batch calculations on the global discharge event dataset to separate partial discharge pulses and determine synchronization phase information based on the partial discharge pulses to generate discharge diagnostic maps corresponding to multiple monitoring points.

[0081] By utilizing the elastic computing resources of the cloud server, parallel batch computation is performed on the spatiotemporally synchronized global discharge event dataset. Signal deep denoising and feature extraction are performed uniformly, and standardized discharge diagnostic maps are generated in parallel for each monitoring point based on the synchronized phase information.

[0082] Based on the aligned and denoised data, the amplitude, number of pulses, and equivalent duration characteristics within each power frequency cycle are calculated, and phase-amplitude-number (PRPD) and / or phase-amplitude-time (PRPS) spectra are generated.

[0083] Specifically, the server first receives data packets from multiple monitoring points. Each data packet contains the maximum pulse value after edge compression, a precise timestamp, and a crucial power frequency phase stamp. Based on these phase stamps, the system accurately aligns all data points to a standard power frequency cycle (0-360 degrees), laying the foundation for subsequent PRPD map generation. Simultaneously, the data packets are decoded and sorted to ensure timing accuracy.

[0084] Raw pulse current signals often contain various types of noise. Cloud servers, with their powerful computing capabilities, can employ more complex algorithms for deep denoising than edge computing. Wavelet transform can decompose signals into different scales (frequency). Noise typically manifests as high-frequency components, while real partial discharge pulses have specific time-frequency characteristics. By analyzing wavelet coefficients at different scales, signals and noise can be distinguished.

[0085] The system accurately calculates fundamental parameters such as pulse amplitude (intensity), number of occurrences, equivalent duration, and equivalent bandwidth within each power frequency cycle. These are the basis for characterizing discharge activity. Specialized graphs are a crucial step in transforming the extracted features into intuitive, diagnostic graphics.

[0086] The S400 constructs a composite index based on the structured monitoring data, feature data, and discharge diagnostic spectrum according to the preset identifier. It responds to the client's display request and, based on the composite index, retrieves and assembles data and spectrums of multiple specified monitoring points within the selected time window and phase interval in real time to generate a synchronous display data package and display it on the client.

[0087] The processed monitoring data, feature data, and discharge diagnostic maps are used to construct a composite index. The index is centrally and persistently stored in a cloud-based distributed database or object storage with monitoring point identification and time-phase as key dimensions. A composite index is also established across monitoring points and time periods to support efficient correlation queries.

[0088] The cloud-driven synchronous display process, which targets multiple points, responds to the client's display request. The cloud visualization service engine retrieves and assembles data and maps of multiple specified monitoring points within a selected time window and phase interval from a unified storage in real time based on a composite index. It then generates a synchronous display data package and sends it to the client, driving it to perform multi-screen, multi-parameter, and interactive comparative displays.

[0089] It's worth noting that the human-computer interaction interface is developed based on web technology. Authorized users can log in via a browser to view real-time / historical data and graphs, and receive alarm information triggered by the system based on preset thresholds. The server can also integrate AI models to perform trend prediction and intelligent diagnosis of insulation status.

[0090] The core monitoring parameters and real-time status form the foundation of the display. It is necessary to provide real-time updates and overviews of key parameters such as device list, channel status, real-time discharge amplitude (such as maximum value and average value), discharge frequency, and pulse count.

[0091] In one embodiment, the discharge diagnostic spectrum includes a PRPS spectrum. Generating discharge diagnostic spectra for multiple monitoring points based on synchronized phase information includes the following steps:

[0092] S310 divides the continuous time axis into equal-length segments according to the preset analysis granularity, and generates PRPD maps for the global discharge events of each equal-length segment.

[0093] S320 arranges the PRPD maps corresponding to each equal-length time slice in chronological order to form a three-dimensional discharge diagnostic map. The discharge diagnostic map characterizes the discharge occurrence at a specific phase, amplitude, and time.

[0094] The cloud continuously receives a stream of discharge events occurring in chronological order from the monitoring points. The core information of each event packet is: [timestamp (t), power frequency phase (φ), pulse amplitude (q)]. The system divides the continuous time axis into a series of time slices of equal length according to the preset analysis granularity (e.g., every 1 second, every 1 power frequency cycle, or every 100 power frequency cycles as a time unit).

[0095] For each time slice, the system, similar to generating a single PRPD map, statistically analyzes all discharge events within that time period to form a two-dimensional "phase-amplitude" distribution map. This distribution records the intensity, phase mode, and frequency of discharges within that time slice.

[0096] Arrange the phase-amplitude distribution charts for each time slice in chronological order. The horizontal axis (X-axis) represents the power frequency phase (φ, 0-360°), the vertical axis (Y-axis) represents the discharge amplitude (q), and the newly added third axis (Z-axis) represents the time (t) or the sequence number of the time slice. This forms a three-dimensional data cube. The attribute value (such as color or brightness) of each point (φ, q, t) in the cube represents the density or intensity of the discharge at a specific phase, amplitude, and time (or time period).

[0097] The three-dimensional stereo view directly displays the three-dimensional coordinate space (φ, q, t), and the discharge activity is presented in the form of point cloud, isosurface or volume rendering. It can display the three-dimensional relationship most completely, but there are requirements for the viewing angle.

[0098] The three-dimensional cube is sliced ​​or projected, with time as the vertical axis (moving from top to bottom) and phase as the horizontal axis. The color or grayscale of each pixel represents the average or maximum amplitude of the discharge at that moment and phase. This view looks like a waterfall chart or fingerprint chart that scrolls over time, clearly showing the stability, burstiness, intermittency, or increasing trend of the discharge pattern over time.

[0099] Map generation and encoding: The cloud-based visualization engine encodes the rendered PRPS maps (usually high-resolution images or dynamic interactive graphic data) in preparation for output.

[0100] Building upon PRPD, a time (t) dimension was added, forming a three-dimensional map (φ-qt). This map can display the evolution of discharge activity over time, which is invaluable for observing the deterioration trend of defects and intermittent discharge phenomena. With high-quality features and maps, the ultimate goal of the analysis is to make an accurate judgment on the insulation status of the equipment and predict future risks.

[0101] In one embodiment, generating a PRPD map for global discharge events of each equal-length slice includes the following steps:

[0102] S311, based on partial discharge pulse, confirms and extracts the synchronous power frequency phase and the synchronous discharge pulse amplitude.

[0103] S312 iterates through all global discharge events and plots a PRPD spectrum with the power frequency phase as the horizontal axis and the discharge pulse amplitude as the vertical axis. The PRPD spectrum represents the frequency of discharge at a specific phase and amplitude based on the density of the coordinate points.

[0104] The PRPD spectrum correlates the amplitude (q) of each pulse with its phase (φ) occurring within the power frequency cycle and displays it as a scatter plot. Different types of insulation defects (such as internal air gaps and surface discharges) will produce PRPD spectra with unique distribution patterns (such as "double peaks"). Experts can directly determine the nature of the defect based on the spectrum morphology.

[0105] The system uses the power frequency phase (φ) as the horizontal axis (typically ranging from 0 to 360 degrees) and the discharge pulse amplitude (q) as the vertical axis. It iterates through all discharge events, treating each event as a data point and plotting it on the corresponding two-dimensional coordinate position according to its phase and amplitude.

[0106] Multiple discharges can occur at the same phase-amplitude location. PRPD maps visually represent the frequency (N) of discharges at a specific phase and amplitude by the density, color intensity, or size of the dots. Regions with higher density (darker color, more dense dots) represent more active discharge characteristics.

[0107] Finally, a standard two-dimensional scatter plot or density distribution map, namely the PRPD map, is generated in the cloud. This map clearly shows the dependence of discharge activity on the phase of the power frequency voltage, the distribution range of discharge amplitude, and the repetition frequency of the discharge.

[0108] In one embodiment, the original monitoring data packets are aligned synchronously to the same power frequency cycle coordinate system in real time based on phase stamps and timestamps. The synchronous alignment specifically includes the following steps:

[0109] S210 uses the power frequency phase of the first arriving monitoring terminal as a reference or synchronizes the time through a cloud server to calibrate and compensate the phase stamps of all monitoring points.

[0110] In one embodiment, a composite index is constructed based on the structured monitoring data, feature data, and discharge diagnostic spectrum according to a preset identifier. The composite index adopts a multi-dimensional label index structure of a time-series database and includes at least the monitoring point ID, data timestamp, power frequency phase value, and signal feature type label. Based on the composite index, data and spectrum within a time window and phase interval are selected according to preset combination conditions.

[0111] In one embodiment, selecting data and spectra within a time window and phase interval based on a composite index according to preset combination conditions includes the following steps:

[0112] S410, when a data query request is received, first parses the query conditions in the request. The conditions include at least the target device identifier, signal type and time range.

[0113] The S420 uses a composite tag index to quickly filter out a set of storage point IDs that meet the conditions of device identification and signal type.

[0114] S430 uses the primary index and metadata records to further locate physical files from the storage point ID set where the data timestamps and query time ranges intersect.

[0115] S440 loads only the located physical files for detailed searching and data retrieval, and returns the data and spectra within the selected time window and phase interval.

[0116] Each data point should be bound to a unique identifier of the device under test (such as device ID and test device information). Data binding aims to establish a relationship between different data fragments, while indexes are the key to enabling fast queries.

[0117] Implementation details: A unique ID is assigned to each telemetry point (such as the temperature or partial discharge amplitude of a specific sensor), and a hash linked list is built based on this ID. When writing data, the storage location in the linked list is quickly located based on the ID through hash calculation, and the new data is added to the tail of the linked list, ensuring that data from the same measurement point is stored continuously in time sequence, significantly improving write and query efficiency.

[0118] Partial discharge data is a typical time-series data. Using a time-series database automatically creates a timestamp-based index for the data. Composite indexes can also be created by combining tags such as device ID and signal type, supporting multi-dimensional queries by time range, device identifier, and other dimensions. File and record management: Data in the hash chain is saved as independent files according to a certain time window, and a record is generated for each file, containing metadata such as filename, data start and end times. During queries, the index is first used to locate the approximate target record group, and then a detailed search is performed within the group, avoiding a full file scan.

[0119] In one embodiment, responding to a client's display request and retrieving and assembling data and maps from multiple specified monitoring points within a selected time window and phase interval based on a composite index in real time, to generate a synchronized display data package and display it on the client, includes the following steps:

[0120] The S450 synchronously displays data packets covering associated data and view configuration information from multiple monitoring points.

[0121] The S460 renders multiple monitoring views based on synchronous display data packets and performs global linkage scaling and panning operations on the time axis and phase axis for the multiple monitoring views.

[0122] The client sends a request, for example: "Compare the discharge activity of sites A, B, and C in the vicinity of the voltage peak phase (80°-100°) over the past 24 hours."

[0123] Instead of directly transmitting raw, massive amounts of data, the cloud engine performs a "pre-computation": based on the device ID, time range, and phase interval in the request, it quickly retrieves relevant data from the global discharge event dataset and feature database using a composite index. This data is then aggregated and sliced ​​to generate formatted data suitable for consumption by front-end charting libraries (such as ECharts and D3.js).

[0124] Based on the number of monitoring points and the data type (PRPD or trend chart) for this comparison, an optimal view layout and configuration is automatically generated. For example, for three sites, a three-column PRPD chart configuration is generated, with a pre-defined uniform coordinate axis.

[0125] The associated data and view configuration information are packaged and sent to the client via WebSocket or HTTP interface. Upon receiving the data packet, the client (usually a web browser) uses a parser to break it down. The data portion is injected into the chart's data model, while the configuration portion is passed to the chart rendering engine.

[0126] The chart rendering engine instantiates multiple chart components on the same page based on the configuration information and populates them with corresponding data. Because a uniform coordinate axis is specified in the configuration, the rendered multiple views inherently have a consistent scale.

[0127] The client framework registers a unified event listener for the global timeline control, the global phase axis control, and the coordinate axes of each view.

[0128] When the user drags the timeline slider to pan, the event handler will:

[0129] a. Calculate the new unified time range.

[0130] b. Broadcast this new scope to all related views (such as PRPS charts, trend charts).

[0131] c. After receiving the instruction, each view synchronously updates the display range of its own X-axis (time axis) and redraws the data within the view.

[0132] Similarly, when a user zooms in on the phase axis of a PRPD image (e.g., focusing on the 30°-60° range), this operation can be automatically synchronized to the phase axes of all other displayed PRPD images through the configured "linkage relationship definition," enabling parallel viewing of multiple images in terms of phase details.

[0133] Synchronous display data packets are structured data sets dynamically assembled, packaged, and distributed by the cloud-based visualization service engine in response to a client's specific request for "multi-monitoring point comparison and viewing." It is not a simple image file, but a "data container" containing data, configuration, and instructions.

[0134] In one embodiment, after traversing all global discharge events and plotting a PRPD graph with the power frequency phase as the abscissa and the discharge pulse amplitude as the ordinate, the following steps are further included:

[0135] S500 retrieves the defect types corresponding to historical PRPD maps in the historical database and uses the historical PRPD maps and defect types as training datasets.

[0136] The S600 trains a prediction model based on a training dataset and inputs the PRPD map into a preset model to obtain the corresponding defect prediction type.

[0137] The S700 determines an initial assessment based on the defect prediction type to generate a diagnostic report, which is then displayed in a visual interface.

[0138] The cloud system can integrate pre-trained machine learning or deep learning models. These models have learned a large number of PRPD map features of known defect types, so they can automatically compare and classify new maps, output the judgment of the insulation defect type (e.g., "floating potential body discharge", "internal corona", etc.), and give a preliminary assessment of its severity.

[0139] The server stores long-term historical data, enabling it to plot trend graphs of key parameters (such as maximum discharge amplitude and pulse repetition rate) over time. By analyzing the slope and direction of these trends, the system can predict the rate of insulation degradation, issue early warnings before faults occur, provide a scientific basis for scheduling planned maintenance, and realize the transformation from "reactive emergency repair" to predictive maintenance.

[0140] All analysis results are presented to operations and maintenance personnel through intuitive visualization interfaces (such as web dashboards and mobile apps). These interfaces typically display a comprehensive list of devices, health scores, real-time PRPD graphs, trend curves, and alarm information. Simultaneously, the system can automatically generate detailed diagnostic reports to support decision-making.

[0141] In one embodiment, after establishing a unified cloud data receiving gateway to receive raw monitoring data packets from multiple monitoring points, the following steps are also included:

[0142] S800 determines whether the number of monitoring terminals or the data receiving rate of the original monitoring data packet exceeds a preset threshold.

[0143] When the number of connected monitoring terminals or the data receiving rate exceeds a preset threshold, the S900 initiates dynamic load balancing in the cloud, automatically distributing data receiving and processing tasks to multiple cloud server instances for execution.

[0144] The system diagnoses traffic pressure in real time using two key indicators: the number of online / active terminals and the number of terminals that have established connections and are continuously uploading data. If this number exceeds a preset threshold (e.g., 10,000 terminals), it means that the overhead of connection management and data distribution is approaching the limit for a single node.

[0145] Data reception rate refers to the number of data packets or the total amount of data (MB / s) received by the cloud per unit time (e.g., per second). A rate exceeding a preset threshold (e.g., 100,000pps or 1Gbps) indicates that the network throughput or processing pipeline of a single node has reached saturation. The preset threshold is pre-set based on performance stress testing and capacity planning for a single server instance and serves as the "water level" for triggering elastic scaling.

[0146] Deploy a lightweight monitoring agent at the cloud-based access gateway or API gateway level to continuously collect key metrics. The monitoring data flows through a rules engine and is compared with preset dynamic thresholds (which may not be fixed values ​​but are adjusted based on time and historical patterns). Once any metric consistently exceeds the threshold (to avoid false triggers due to momentary fluctuations), the system immediately generates a scaling event, indicating that the current single node or the existing cluster capacity is insufficient.

[0147] Dynamic load balancing is an automated, cloud platform API-driven process. Upon receiving a scaling event, the control center (such as a Kubernetes controller or a cloud provider's AutoScaling service) automatically invokes the cloud platform API to quickly start (hot-start) one or more new cloud server instances (virtual machines or containers) with identical configurations from a pre-built image template. After the new instance starts, it automatically pulls and runs standardized data receiving and preprocessing services. Once the new instance starts successfully, it automatically registers with the load balancer (such as Nginx, HAProxy, or the cloud provider's CLB / ALB service) and joins the backend server pool.

[0148] The load balancer updates its routing table in real time. Thereafter, connection requests from newly connected monitoring terminals, or new data streams, are intelligently distributed to all available server instances, including new instances, according to predetermined strategies (such as round-robin, least connections, hash (by terminal ID)).

[0149] Data receiving tasks are directly distributed to different instances, and each instance only processes the portion of terminal connections and data streams assigned to it.

[0150] The situation is more complex, involving state sharing. Systems are typically designed as follows:

[0151] a. Stateless processing: For pure computational tasks (such as denoising and feature extraction), any instance can process the data packet independently. After processing, the result (global discharge event) is written to a shared central storage (such as a cloud database or message queue).

[0152] b. Session persistence: For tasks that require maintaining context, the load balancer will use the "session persistence" strategy to ensure that all data packets from the same terminal are always sent to the same backend instance for processing during the session, so as to guarantee the continuity of data processing for that terminal (e.g., cumulative calculation of phase alignment).

[0153] This enables "scaling up when needed and scaling down when not in use," handling maximum business pressure with minimal resource cost and avoiding the need to maintain expensive high-configuration hardware year-round to cope with peak traffic. It also avoids single points of failure. Even if a processing instance crashes, the load balancer will automatically remove it from the service pool and route traffic to healthy instances, ensuring uninterrupted service.

[0154] In theory, as long as cloud resources are sufficient, the overall throughput and processing capacity of the system can be linearly improved by continuously adding instances, easily supporting a smooth evolution of the monitoring scale from hundreds of points to millions of points.

[0155] The entire expansion, scheduling, and service discovery process requires no manual intervention, greatly reducing the complexity of operation and maintenance, and forming the foundation for building an "unattended" intelligent operation and maintenance platform.

[0156] The key evolution of the system from a "static architecture" to a "dynamic organism" is not merely a functional description, but also embodies the modern cloud-native design philosophy—achieving elastic provisioning of computing resources and intelligent task scheduling through monitoring-driven and API collaboration. This, in turn, ensures the core infrastructure capabilities that enable the large-scale, centralized, real-time cable partial discharge monitoring platform in this embodiment to operate stably, reliably, and economically.

[0157] This application also discloses a multi-point data traffic overload processing system based on synchronous display.

[0158] like Figure 1 As shown, the multi-point data traffic overload processing system based on synchronous display includes a data receiving module, a data preprocessing module, a core data processing module, and a data storage and query module.

[0159] The data receiving module receives raw monitoring data packets from multiple monitoring points based on a unified cloud data receiving gateway. The raw monitoring data packets carry power frequency phase stamps and timestamps.

[0160] The data preprocessing module aligns the original monitoring data packets to the same power frequency cycle coordinate system in real time and synchronously based on phase stamps and timestamps to generate a spatiotemporally synchronized global discharge event dataset.

[0161] The core data processing module performs parallel and batch calculations on the global discharge event dataset to separate partial discharge pulses and determine synchronization phase information based on the partial discharge pulses in order to generate discharge diagnostic maps corresponding to multiple monitoring points.

[0162] The data storage and query module constructs a composite index based on the structured monitoring data, feature data, and discharge diagnostic maps according to the preset identifier. It responds to the client's display request and, based on the composite index, retrieves and assembles data and maps of multiple specified monitoring points within the selected time window and phase interval in real time to generate a synchronous display data package and display it on the client.

[0163] The system also includes a trend analysis module, a diagnostic alarm module, and an expansion module. The trend analysis module is used to display the change curve of specific parameters (such as maximum discharge) over time, which is used to assess the rate and trend of insulation degradation.

[0164] The diagnostic alarm module is used by the system to provide qualitative discharge type identification results, insulation status assessment (such as normal, warning, abnormal) and alarm information of different levels (such as amplitude alarm, frequency alarm) based on the spectrum and parameters and through built-in algorithms, and should provide specific handling suggestions.

[0165] The cloud can deploy sophisticated machine learning algorithms to train on massive amounts of partial discharge data from monitoring points across the entire network, forming a precise fault feature library. When new data is uploaded, the system can automatically compare it with the feature library, identifying not only the type of insulation defect (such as internal air gap discharge, surface discharge, etc.) but also assessing its severity and providing a diagnostic report. This significantly reduces reliance on on-site expert experience and improves the accuracy and consistency of diagnosis.

[0166] Based on long-term historical data stored in the cloud, the system can build a predictive model for equipment health status. By analyzing the changing trends of parameters such as partial discharge amplitude and frequency over time, the model can predict the rate of insulation degradation, issue early warnings before a fault occurs, and estimate the remaining lifespan, thus gaining valuable time for scheduled maintenance and transforming "reactive repair" into "proactive prevention." For example, the system may predict several days or weeks in advance that the insulation of a cable joint is about to break down.

[0167] As the monitoring scale expands, the expansion module meets the demand by adding cloud server instances (virtual machines), expanding storage capacity, or increasing network bandwidth. This elastic scalability allows the system to smoothly support upgrades from a few key points to large-scale, network-wide cable monitoring.

[0168] The implementation principle is as follows:

[0169] The server first receives data packets from multiple monitoring points. Each data packet contains the maximum pulse value after edge compression, a precise timestamp, and a crucial power frequency phase stamp. Based on these phase stamps, the system accurately aligns all data points to a standard power frequency cycle (0-360 degrees), laying the foundation for subsequent PRPD map generation. Simultaneously, the data packets are decoded and sorted to ensure timing accuracy.

[0170] Cloud servers, with their powerful computing capabilities, can employ more complex algorithms for deep denoising than edge computing. Wavelet transform can decompose signals into different scales (frequency). Noise typically manifests as high-frequency components, while real partial discharge pulses have specific time-frequency characteristics. By analyzing wavelet coefficients at different scales, signals and noise can be distinguished.

[0171] The system accurately calculates the pulse amplitude (intensity), number of occurrences, equivalent duration, and equivalent bandwidth within each power frequency cycle. It correlates the amplitude (q) of each pulse with its phase (φ) within the power frequency cycle and displays this correlation as a scatter plot. Different types of insulation defects (such as internal air gaps and surface discharges) produce PRPD patterns with unique distribution patterns (such as "double peaks"). Experts can directly determine the nature of the defect based on the pattern.

[0172] Building upon PRPD, a time (t) dimension was added, forming a three-dimensional map (φ-qt). This map can display the evolution of discharge activity over time, which is invaluable for observing the deterioration trend of defects and intermittent discharge phenomena. With high-quality features and maps, the ultimate goal of the analysis is to make an accurate judgment on the insulation status of the equipment and predict future risks.

[0173] The cloud system can integrate pre-trained machine learning or deep learning models. These models have learned a large number of PRPD map features of known defect types, so they can automatically compare and classify new maps, output the judgment of the insulation defect type (e.g., "floating potential body discharge", "internal corona", etc.), and give a preliminary assessment of its severity.

[0174] All analysis results are presented to operations and maintenance personnel through intuitive visualization interfaces (such as web dashboards and mobile apps). These interfaces typically display a comprehensive list of devices, health scores, real-time PRPD graphs, trend curves, and alarm information. Simultaneously, the system can automatically generate detailed diagnostic reports to support decision-making.

[0175] The other functions performed in the aforementioned data receiving module, data preprocessing module, core data processing module, and data storage query module, as well as the technical details of each function, are the same as or similar to the corresponding features in the multi-point data traffic overload processing method based on synchronous display described above, so they will not be repeated here.

[0176] It should be understood that although the steps in the flowcharts in the accompanying drawings are shown sequentially as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise expressly stated herein, there is no strict order in which these steps are performed, and they may be performed in other orders.

[0177] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for handling multi-point data traffic overload based on synchronous display, characterized in that, Includes the following steps: Establish a unified cloud data receiving gateway to receive raw monitoring data packets from multiple monitoring points. The raw monitoring data packets carry power frequency phase stamps and timestamps. Based on the phase stamp and timestamp, the original monitoring data packets are aligned in real time and synchronously to the same power frequency cycle coordinate system to generate a spatiotemporally synchronized global discharge event dataset. The global discharge event dataset is processed in parallel and in batches to separate partial discharge pulses, and the synchronization phase information is determined based on the partial discharge pulses to generate discharge diagnostic maps corresponding to multiple monitoring points. A composite index is constructed based on the structured monitoring data, feature data, and discharge diagnostic spectrum according to the preset identifier. The system responds to the client's display request and, based on the composite index, retrieves and assembles data and spectrums of multiple specified monitoring points within a selected time window and phase interval in real time to generate a synchronous display data package and display it on the client.

2. The method for handling multi-point data traffic overload based on synchronous display according to claim 1, characterized in that, The discharge diagnostic spectrum includes a PRPS spectrum, which is generated based on the synchronized phase information to correspond to multiple monitoring points. The process includes the following steps: The continuous time axis is divided into equal-length segments based on the preset analysis granularity, and a PRPD map is generated for the global discharge events of each equal-length segment. The PRPD maps corresponding to each equal-length time slice are arranged sequentially according to time sequence to form a three-dimensional discharge diagnostic map. The discharge diagnostic map characterizes the discharge occurrence at a specific phase, amplitude, and time.

3. The method for handling multi-point data traffic overload based on synchronous display according to claim 2, characterized in that, Generate a PRPD map for the global discharge events of each equal-length slice, including the following steps: Based on the partial discharge pulse, the synchronous power frequency phase and the synchronous discharge pulse amplitude are confirmed and extracted; All global discharge events are traversed, and a PRPD spectrum is plotted with the power frequency phase as the horizontal axis and the discharge pulse amplitude as the vertical axis. The PRPD spectrum represents the frequency of discharge occurrence at a specific phase and amplitude based on the density of coordinate points.

4. The method for handling multi-point data traffic overload based on synchronous display according to claim 3, characterized in that, After iterating through all global discharge events and plotting a PRPD graph with power frequency phase on the x-axis and discharge pulse amplitude on the y-axis, the following steps are also included: Obtain the defect types corresponding to historical PRPD maps in the historical database, and use the historical PRPD maps and defect types as training datasets; A prediction model is trained based on the training dataset, and the PRPD map is input into the preset model to obtain the corresponding defect prediction type. A preliminary assessment is determined based on the defect prediction type to generate a diagnostic report, which is then displayed using a visual interface.

5. The method for handling multi-point data traffic overload based on synchronous display according to claim 1, characterized in that, Based on the phase stamp and timestamp, the original monitoring data packet is aligned synchronously and in real time to the same power frequency cycle coordinate system, wherein the synchronous alignment specifically includes the following steps: The phase stamps of all monitoring points are calibrated and compensated using the power frequency phase of the first arriving monitoring terminal as a reference or through unified time synchronization via a cloud server.

6. The method for handling multi-point data traffic overload based on synchronous display according to claim 1, characterized in that, A composite index is constructed based on the structured monitoring data, feature data, and discharge diagnostic maps according to the preset identifier. The composite index adopts a multi-dimensional label index structure of time-series database and includes at least monitoring point ID, data timestamp, power frequency phase value, and signal feature type label. Based on the composite index, data and maps within the time window and phase interval are selected according to preset combination conditions.

7. The method for handling multi-point data traffic overload based on synchronous display according to claim 6, characterized in that, Based on the composite index, data and spectra within a time window and phase interval are selected according to preset combination conditions, including the following steps: When a data query request is received, the query conditions in the request are first parsed. The conditions include at least the target device identifier, signal type, and time range. Using the composite tag index, a set of storage point IDs that meet the conditions of device identifier and signal type can be quickly filtered out; Using the primary index and the metadata records, physical files whose data timestamps and query time ranges intersect can be further located from the storage point ID set; Only the located physical files are loaded for detailed searching and data retrieval, and the data and spectra within the selected time window and phase interval are returned.

8. The method for handling multi-point data traffic overload based on synchronous display according to claim 1, characterized in that, Responding to the client's display request and retrieving and assembling data and maps from multiple specified monitoring points within a selected time window and phase interval based on the composite index in real time, to generate a synchronized display data package and display it on the client, includes the following steps: The synchronous display data package covers associated data and view configuration information of multiple monitoring points; Based on the synchronous display data packet, multiple monitoring views are rendered, and the multiple monitoring views are globally linked for scaling and panning on the time axis and phase axis.

9. The method for handling multi-point data traffic overload based on synchronous display according to claim 1, characterized in that, After establishing a unified cloud data receiving gateway to receive raw monitoring data packets from multiple monitoring points, the following steps are also included: Determine whether the number of monitoring terminals or the data receiving rate of the original monitoring data packet exceeds a preset threshold; When the number of connected monitoring terminals or the data receiving rate exceeds a preset threshold, dynamic load balancing in the cloud is activated to automatically distribute data receiving and processing tasks to multiple cloud server instances for execution.

10. A multi-point data traffic overload processing system based on synchronous display, characterized in that, The method for handling multi-point data traffic overload based on synchronous display as described in any one of claims 1-9 includes: The data receiving module receives raw monitoring data packets from multiple monitoring points based on a unified cloud data receiving gateway. The raw monitoring data packets carry power frequency phase stamps and timestamps. The data preprocessing module aligns the original monitoring data packets to the same power frequency cycle coordinate system in real time and synchronously based on the phase stamp and timestamp, so as to generate a spatiotemporally synchronized global discharge event dataset. The core data processing module performs parallel and batch calculations on the global discharge event dataset to separate partial discharge pulses and determine synchronization phase information based on the partial discharge pulses to generate discharge diagnostic maps corresponding to multiple monitoring points. The data storage and query module constructs a composite index based on the structured monitoring data, feature data, and discharge diagnostic spectrum according to a preset identifier. It responds to the client's display request and, based on the composite index, retrieves and assembles data and spectra of multiple specified monitoring points within a selected time window and phase interval in real time to generate a synchronous display data package and display it on the client.

Images