Partial discharge monitoring strategy optimization method and system based on dynamic resource allocation

Abstract

The invention relates to the technical field of power system operation or management, in particular to a partial discharge monitoring strategy optimization method and system based on dynamic resource allocation, and the method comprises the steps: constructing a three-stage monitoring system comprising a sensor node, a sink node and a cloud processing center, firstly initializing monitoring parameters, and then obtaining system state information periodically or in a triggering manner, calculating the risk level of each monitoring point in combination with a dynamic risk evaluation model; constructing an efficiency-maximized resource allocation optimization model based on risk levels and resource constraints, solving an optimal scheme by adopting an improved multi-target particle swarm algorithm, and issuing the optimal scheme to each node to adjust monitoring behaviors to form closed-loop optimization; and meanwhile, model parameters are dynamically updated through an online learning mechanism. According to the method, dynamic matching of risks and resources is realized, the monitoring accuracy and the resource utilization rate are improved, the adaptability of the system to the equipment state and the environment change is enhanced, and the method is suitable for partial discharge monitoring scenes of various power equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system operation or management, in particular to a partial discharge monitoring strategy optimization method and system based on dynamic resource allocation. BACKGROUND

[0002] The current partial discharge monitoring technology faces many bottleneck problems that need to be solved in practical application, which is specifically embodied in the following aspects: In the operation and maintenance of power systems, partial discharge monitoring is a key means to evaluate the insulation state of equipment. It can detect the discharge signal caused by internal insulation defects of the equipment, and can detect the insulation deterioration trend in advance. It is of great significance to avoid equipment failure and ensure stable operation of the system. With the increase of types of power equipment and the complexity of operating conditions, the existing partial discharge monitoring technology gradually shows many limitations in practical application. At present, most monitoring strategies use a fixed parameter configuration mode, that is, fixed monitoring frequency, sampling accuracy and other parameters are preset for monitoring points, which remain unchanged throughout the process. This mode is difficult to adapt to the dynamic state of the equipment. When the equipment has abnormal discharge, insulation state changes, etc., the fixed parameters cannot be adjusted in time. If the initial parameter setting is too low, the discharge signal of high-risk equipment may not be captured in time, forming a monitoring blind area. If the parameters are generally too high, it will lead to excessive consumption of resources in low-risk equipment area, causing waste of computing resources, communication bandwidth and energy, and it is difficult to balance monitoring efficiency and resource cost. At the resource allocation level, the existing scheme lacks collaborative optimization logic based on the actual risk of the equipment, often uses "average allocation" or "experience allocation" method, and does not dynamically bind the real-time risk level of the equipment with the resource demand. This makes it possible for high-risk equipment to have insufficient monitoring accuracy due to insufficient resources, and low-risk equipment to occupy too many resources, resulting in low overall monitoring efficiency of the system. At the same time, the existing system has weak adaptability to dynamic scenarios. During equipment operation, the insulation state will continuously evolve with factors such as operating time, environmental interference (such as electromagnetic interference, temperature and humidity changes), etc. The traditional monitoring strategy lacks an effective closed-loop adjustment mechanism and cannot dynamically update the monitoring strategy according to the changes in equipment state, environmental interference intensity, etc. After long-term operation, parameter mismatch may occur, leading to risk misjudgment or omission. In addition, some schemes do not fully consider the differences in equipment types, historical operating characteristics and industry specification requirements during the initial parameter configuration stage, but rely solely on general parameters to start monitoring, which may lead to unreasonable parameters at the initial stage of system startup, affecting the basic monitoring effect and requiring frequent manual intervention for adjustment, increasing the operation and maintenance cost.

[0003] Therefore, a partial discharge monitoring strategy optimization method and system based on dynamic resource allocation are proposed to solve the above problems. SUMMARY

[0004] The application aims to provide a partial discharge monitoring strategy optimization method and system based on dynamic resource allocation to solve the problems in the background art.

[0005] To achieve the above-mentioned purpose, the application provides the following technical solutions. The partial discharge monitoring strategy optimization method based on dynamic resource allocation comprises the following steps: S1, a partial discharge monitoring system is constructed, which comprises sensor nodes deployed at multiple monitoring points, a convergence node and a cloud processing center; the sensor nodes are used to collect partial discharge signals; the convergence node is used to receive and pre-process the data of the sensor nodes; and the cloud processing center is used to execute an optimization algorithm of a monitoring strategy and issue control instructions; S2, the monitoring strategy is initialized, and initial monitoring frequencies and data acquisition accuracies are configured for the monitoring points; S3, the cloud processing center periodically or triggeredly acquires current system state information, which comprises real-time discharge amount data, historical discharge trend data, equipment working condition data, available resource states of the sensor nodes and a communication network of each monitoring point; S4, based on the system state information, a real-time discharge risk level of each monitoring point is calculated through a dynamic risk evaluation model; the dynamic risk evaluation model fuses real-time discharge amount, discharge trend change rate and equipment health state parameters, and outputs a quantitative risk evaluation value; S5, based on resource constraint conditions and the real-time discharge risk level of each monitoring point, a dynamic resource allocation optimization model is constructed, with the maximization of system overall monitoring efficiency as the target; the resource constraint conditions comprise total calculation resource constraint, total communication bandwidth constraint and total energy budget constraint; the monitoring efficiency is jointly defined by risk coverage rate, state sensing accuracy and resource consumption efficiency; S6, a multi-objective optimization algorithm is used to solve the dynamic resource allocation optimization model, and an optimal resource allocation scheme allocated to each monitoring point in the next period is obtained; the multi-objective optimization algorithm is an improved multi-objective particle swarm optimization algorithm based on chaotic mapping and adaptive crossover and mutation; the resource allocation scheme at least comprises calculation resource shares, communication bandwidth shares allocated to each monitoring point and monitoring task execution parameters determined therefrom, and the monitoring task execution parameters comprise monitoring frequencies and data acquisition accuracies; S7, the resource allocation scheme is issued to the corresponding convergence node and sensor node, and the monitoring behavior of each monitoring point is adjusted; S8, steps S3 to S7 are repeated to realize dynamic closed-loop optimization of the monitoring strategy.

[0006] As a preferred solution, the calculation process of the dynamic risk evaluation model in step S4 is as follows: the ratio of the real-time discharge amount to the discharge amount reference threshold is multiplied by the real-time discharge amount weight coefficient, and the absolute value of the discharge amount change rate is multiplied by the discharge amount change rate weight coefficient, and the output value of the mapping function of the influence of the equipment health state on the risk is multiplied by the equipment health state weight coefficient, and the sum of the three is the real-time discharge risk evaluation value; wherein the sum of the real-time discharge amount weight coefficient, the discharge amount change rate weight coefficient and the equipment health state weight coefficient is one.

[0007] As a preferred solution, the objective function of the dynamic resource allocation optimization model in step S5 is as follows: maximize the overall monitoring efficiency of the system, which is equal to the sum of the risk coverage efficiency and the perception accuracy efficiency of each monitoring point minus the total resource consumption cost of the system; wherein the risk coverage efficiency is the product of the risk evaluation value and the logarithm value of the monitoring frequency multiplied by the risk coverage weight factor, and the perception accuracy efficiency is the product of the data acquisition accuracy and the logarithm value multiplied by the perception accuracy weight factor. The resource constraints include: the total computational resource consumption of each monitoring point does not exceed the upper limit of the total system computational resource, the total bandwidth resource consumption of each monitoring point does not exceed the upper limit of the total system communication bandwidth, and the total energy resource consumption of each monitoring point does not exceed the upper limit of the total system energy budget; wherein the resource consumption of each monitoring point is the product of the single resource consumption coefficient of the monitoring point and the monitoring frequency and the data acquisition accuracy level.

[0008] As a preferred solution, the improved multi-objective particle swarm optimization algorithm based on chaotic mapping and adaptive crossover and mutation in step S6 has the following calculation process: Chaotic initialization: generate a chaotic sequence using the logistic mapping, and map the chaotic sequence to the value interval of the decision variable to generate an initial population; Adaptive inertia weight: the inertia weight decreases nonlinearly with the increase of the iteration number, and the initial inertia weight is greater than the final inertia weight; Particle velocity and position update: the new velocity is equal to the current inertia weight multiplied by the current velocity plus the individual learning factor multiplied by the random number multiplied by the difference between the individual historical optimal position and the current position plus the social learning factor multiplied by the random number multiplied by the difference between the global optimal position and the current position; the new position is equal to the current position plus the new velocity; Adaptive crossover and mutation: the crossover probability and the mutation probability are adaptively adjusted according to the particle fitness value, and the closer the fitness value is to the maximum fitness value, the smaller the crossover probability and the mutation probability are; External archive update and guide particle selection: maintain the external archive by using non-dominated sorting and congestion calculation, and select the guide particle from the non-dominated layer; The output of the improved multi-objective particle swarm optimization algorithm based on chaotic mapping and adaptive crossover and mutation is a resource allocation solution set that satisfies the Pareto optimality.

[0009] As a preferred solution, the final executed resource allocation scheme is selected from the Pareto optimal solution set according to the preset decision rule in step S6, specifically: calculating the comprehensive evaluation index of each scheme, which is equal to the normalized system overall monitoring effectiveness multiplied by the monitoring effectiveness decision weight coefficient plus a minus the difference between the normalized system total resource consumption cost multiplied by the resource cost decision weight coefficient; selecting the resource allocation scheme with the largest comprehensive evaluation index as the final execution scheme; wherein the sum of the monitoring effectiveness decision weight coefficient and the resource cost decision weight coefficient is one.

[0010] As a preferred solution, the adjustment of the monitoring behavior of each monitoring point in step S7 specifically includes: the aggregation node generates specific scheduling instructions according to the received monitoring task execution parameters and issues them to the corresponding sensor nodes; the sensor nodes adjust their signal collection frequency, sampling rate and signal preprocessing algorithm complexity according to the scheduling instructions.

[0011] As a preferred solution, the trigger condition for triggering the acquisition of the current system state information in step S3 includes: the real-time discharge amount of any monitoring point exceeds a preset threshold, a device operating state mutation signal is received, or the sensor node resource availability rate is lower than a safety threshold.

[0012] As a preferred solution, the method further comprises: S9, dynamically adjusting the risk evaluation weight coefficient in the dynamic risk evaluation model and the effectiveness balance weight factor in the objective function of the dynamic resource allocation optimization model according to historical monitoring data and resource allocation effects through an online learning mechanism.

[0013] The steps of the optimization system execution method of the partial discharge monitoring strategy optimization system based on dynamic resource allocation.

[0014] As can be seen from the above technical solutions provided by the present application, the partial discharge monitoring strategy optimization method and system based on dynamic resource allocation provided by the present application have the following beneficial effects: Improve the accuracy and risk sensitivity of partial discharge monitoring: by integrating multi-dimensional parameters such as real-time discharge amount, discharge change rate and device health status through the dynamic risk evaluation model, the risk level of each monitoring point is quantitatively calculated, which can more accurately capture the device insulation degradation trend (such as sudden discharge, gradual aging, etc.) compared to the traditional single threshold judgment method; at the same time, by dynamically adjusting the model weight coefficient and the mapping function parameter through the online learning mechanism, the device aging, environmental interference and other dynamic changes can be adapted, avoiding risk misjudgment or omission, and significantly improving the timeliness and accuracy of fault warning; Achieving efficient allocation of resources and cost control: the present application constructs a dynamic resource allocation optimization model with the goal of "maximizing the overall monitoring efficiency of the system", and solves the optimal scheme under the constraints of computing resources, communication bandwidth, energy, etc. by combining multi-objective optimization algorithm - by binding high-risk points with high resource demand, reducing resource investment moderately at low-risk points, and avoiding "one-size-fits-all" resource waste; in actual operation, the total resource consumption cost can be reduced while ensuring the monitoring coverage quality of high-risk areas; Enhancing the dynamic adaptability and long-term stability of the system: through the closed-loop iteration of "state perception-risk calculation-resource optimization-behavior adjustment", the system can periodically or triggeredly respond to changes in device working conditions and resource states (such as sudden increase in discharge amount, insufficient sensor power, etc.), and adjust the monitoring frequency and collection accuracy in real time to avoid the limitations of fixed strategies; on the other hand, the online learning mechanism continuously iterates the model parameters (such as risk evaluation weights, resource consumption coefficients, etc.) through historical data, enabling the system to adapt to scenarios such as device characteristic evolution (such as accelerated insulation aging) and environmental interference changes (such as enhanced electromagnetic interference) in the long term, without the need for frequent manual intervention and adjustment, ensuring the stability of long-term operation. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 The figure shows the step flowchart of the partial discharge monitoring strategy optimization method based on dynamic resource allocation of the present application. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical scheme and advantages of the present application clearer and more understandable, the following will further describe the present application in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0017] In order to better understand the above technical scheme, the above technical scheme will be described in detail in combination with the drawings and specific embodiments of the specification.

[0018] As Figure 1 shown, the present application provides a partial discharge monitoring strategy optimization method based on dynamic resource allocation, which includes the following steps: S1, constructing a partial discharge monitoring system, the partial discharge monitoring system includes sensor nodes deployed at multiple monitoring points, a gathering node and a cloud processing center; the sensor nodes are used to collect partial discharge signals; the gathering node is used to receive and preprocess the data of the sensor nodes; the cloud processing center is used to execute the optimization algorithm of the monitoring strategy and issue control instructions; S2, initializing the monitoring strategy, configuring the initial monitoring frequency and data collection accuracy for each monitoring point; S3, the cloud processing center periodically or triggeredly acquires current system state information, the system state information including real-time discharge amount data, historical discharge trend data, equipment working condition data, available resource state of the sensor nodes and the communication network of each monitoring point; S4, based on the system state information, calculating the real-time discharge risk level of each monitoring point through a dynamic risk evaluation model; the dynamic risk evaluation model fuses real-time discharge amount, discharge trend change rate and equipment health state parameters, and outputs a quantitative risk evaluation value; S5, based on the resource constraint conditions and the real-time discharge risk level of each monitoring point, a dynamic resource allocation optimization model is constructed with the goal of maximizing the overall monitoring efficiency of the system; the resource constraint conditions include total computing resource constraint, total communication bandwidth constraint and total energy budget constraint; the monitoring efficiency is jointly defined by risk coverage rate, state sensing accuracy and resource consumption efficiency; S6, a multi-objective optimization algorithm is used to solve the dynamic resource allocation optimization model, and the optimal resource allocation scheme allocated to each monitoring point in the next period is obtained; the multi-objective optimization algorithm is an improved multi-objective particle swarm optimization algorithm based on chaotic mapping and adaptive crossover and mutation; the resource allocation scheme at least includes the computing resource share, the communication bandwidth share allocated to each monitoring point and the monitoring task execution parameters determined therefrom, the monitoring task execution parameters including monitoring frequency and data acquisition accuracy; S7, the resource allocation scheme is issued to the corresponding sink nodes and sensor nodes, and the monitoring behavior of each monitoring point is adjusted; S8, repeating steps S3 to S7 realizes dynamic closed-loop optimization of the monitoring strategy; S9, through an online learning mechanism, dynamically adjusting the risk evaluation weight coefficients in the dynamic risk evaluation model and the efficiency balance weight factors in the objective function of the dynamic resource allocation optimization model according to historical monitoring data and resource allocation effects.

[0019] In this embodiment, the core of step S1 is to build a basic system architecture for partial discharge monitoring, which provides a hardware and operation carrier for subsequent monitoring strategy implementation and dynamic optimization, and the construction of a complete monitoring link is mainly realized through "node deployment - level configuration - system joint debugging": First, the selection and deployment of sensor nodes are needed, which need to be combined with the discharge characteristics of monitoring equipment (such as transformers, switch cabinets, etc.), select appropriate sensors (such as ultra-high frequency, ultrasonic sensors, etc.), and determine the monitoring point position according to the weak insulation parts of the equipment and the field intensity concentration area, and complete the node initial parameter (such as sampling rate, acquisition threshold) configuration and function verification to ensure the accuracy and coverage of signal acquisition; Secondly, the construction and configuration of the aggregation node, taking the edge computing terminal as the core, builds a hardware foundation containing processors, communication modules, and storage units, while implanting data preprocessing functions (such as noise filtering, feature compression, and integrity verification), and setting communication scheduling rules (such as time division multiple access mechanism) with sensors, to achieve efficient reception, processing, and transit of sensor data; Thirdly, the architecture of the cloud processing center is built, deploying server clusters and database systems, dividing the functions of application, database, and algorithm servers, while deploying core software modules such as state awareness, optimization algorithms, and instruction generation, developing a two-way communication interface with the aggregation node, forming an intelligent processing hub of "data reception-algorithm running-instruction issuance"; Finally, through system debugging, the coordination of each link is verified, the stability of the communication link between the sensor and the aggregation node, and the aggregation node and the cloud is tested, the integrity and timeliness of data flow are checked, and the running state (such as power consumption, temperature) of each node is confirmed to meet the design standards, finally building a three-level monitoring system of "sensor collection-aggregation node preprocessing-cloud intelligent optimization", providing hardware support and operating environment for subsequent steps.

[0020] In this embodiment, the core of step S2 is to set initial monitoring parameter benchmarks for each monitoring point, configure basic monitoring strategies by combining device characteristics and industry standards, provide initial operation basis for system startup, and lay a reference foundation for subsequent dynamic resource allocation, mainly including five parts: First, the determination of initial parameter benchmarks, which needs to combine device type differentiation analysis (such as the difference in insulation characteristics between high-voltage and low-voltage devices), historical data statistical modeling (based on discharge data of similar devices in the past three years to determine normal benchmarks), and industry standard adaptation adjustment (according to the power equipment monitoring guide to check parameter compliance), to form targeted benchmark values; Second, the classification configuration of monitoring frequency, setting basic frequency according to device importance level (high frequency monitoring for first-level devices and low frequency monitoring for third-level devices), while presetting frequency adjustment upper and lower limits (to avoid subsequent dynamic adjustment exceeding the limits), and adding trigger type acquisition thresholds (such as automatically starting continuous acquisition when single discharge exceeds the threshold); Third, the initial configuration of data acquisition accuracy, dividing the accuracy into three levels, corresponding to different sampling rates, sampling bits, and signal resolutions (high-risk devices adapt to high accuracy), while associating resource consumption verification (to ensure that the total resource consumption of initial configuration does not exceed 60% of the system rated capacity), and correcting the accuracy level for high interference environment; Fourth, parameter writing and effectiveness verification, the cloud generates standardized configuration instructions (including monitoring point ID, frequency, accuracy, etc.), which are issued to the aggregation node and sensor through an encrypted channel, the node stores the parameters and feeds back confirmation, the cloud triggers the first acquisition to verify the effectiveness of the parameters (such as sampling rate deviation and frequency execution error), and reconfigures the abnormal node; Five is to establish an initial policy benchmark database, cloud archive each monitoring point initial parameters (including configuration time, device type, resource consumption benchmark, etc.), evaluate the benchmark performance indicators (such as risk coverage, data integrity rate), and mark the policy version, form a traceable initial parameter archive, and provide a comparison benchmark for subsequent optimization.

[0021] In this embodiment, the trigger condition for triggering the acquisition of the current system state information in step S3 includes that the real-time discharge amount of any monitoring point exceeds the preset threshold, a device operating state mutation signal is received, or the sensor node resource availability rate is lower than the safety threshold; Further, the function of step S3 is to perceive the system operating state and the device discharge characteristics in real time, and to comprehensively acquire multi-dimensional state information by combining periodic collection with triggered response, so as to provide accurate and timely data input for dynamic risk assessment and resource allocation optimization, and to ensure the scientificity and timeliness of subsequent decision-making; the following are the detailed steps: Step S3-1: Periodic information collection mechanism is established: Collection cycle classification: According to the device type and importance, the collection cycle level is divided: Primary equipment (such as main transformer): the basic collection cycle is set to 5 minutes / time, and each collection contains 3 consecutive power frequency cycles (60ms) of discharge signal; Secondary equipment (such as GIS switch cabinet): the basic collection cycle is set to 15 minutes / time, and each collection contains 2 power frequency cycles (40ms); Tertiary equipment (such as cable branch box): the basic collection cycle is set to 30 minutes / time, and each collection contains 1 power frequency cycle (20ms); The cycle length can be dynamically adjusted by cloud instructions (adjustment step is 5 minutes, range is 1-60 minutes); Periodic data collection content: the core information to be acquired in each periodic collection includes: Discharge characteristic data: real-time discharge amount , discharge pulse frequency, discharge phase distribution (pulse distribution ratio in the range of 0-360°); Device operating data: device operating voltage (kV), load current (A), winding / housing temperature (℃), ambient temperature and humidity (℃ / %RH); Node resource state: sensor remaining power, storage space usage, communication module signal strength (dBm), CPU occupancy rate of aggregation node; Dynamic adaptation of collection frequency: When the equipment is in a special operating state (such as load rate > 90%, ambient humidity > 90%), the collection cycle is automatically shortened by 50% (for example, from 5 minutes to 2.5 minutes for level 1 equipment), and returns to the original cycle within 10 minutes after the state is restored, ensuring information density under high-risk conditions; Step S3-2: Setting trigger information collection conditions: Level 1 trigger condition (emergency response): High-density data collection is triggered immediately when any of the following conditions are met: Real-time discharge amount ( The discharge threshold of the equipment, such as transformer ); Discharge capacity change rate ( for Time monitoring point The discharge capacity change rate, which reflects the rate of change of the discharge capacity over time) (the discharge capacity suddenly increases by more than 500pC within 10 seconds); The sensor node battery level is ≤ 15% or the communication signal strength is ≤ -100dBm (resource critical state); After triggering, the continuous acquisition mode starts: the acquisition is performed once every 2 seconds for 1 minute, and then switches to collecting once every 10 seconds for 5 minutes, until the status returns to normal; Secondary trigger conditions (focus on response): When the following conditions are met, the enhanced acquisition is triggered: The real-time discharge amount is interval; The equipment temperature exceeds the rated value by 5°C (e.g. the oil temperature on the top layer of the transformer exceeds 85°C); Aggregation node cache usage rate ≥ 80% (data congestion warning); After the trigger, shorten the acquisition cycle to 1 / 2 of the original cycle (for example, from 15 minutes to 7.5 minutes for secondary equipment), and evaluate whether recovery has occurred after 3 cycles. Trigger signal priority processing: When multiple trigger conditions are met simultaneously, acquisition tasks are scheduled according to the priority of "first-level trigger > second-level trigger > periodic acquisition" to ensure priority allocation of resources in emergency situations (for example, periodic acquisition is suspended when a first-level trigger occurs, monopolizing the communication channel); Step S3-3: Multi-dimensional status information classification processing: Real-time discharge data processing: The original discharge signal collected by the sensor is pre-processed by the sink node to extract characteristic parameters: Use peak detection algorithm to identify the maximum discharge amount in each power frequency cycle , the error is controlled within ±3%; Calculate the discharge phase distribution entropy through phase window statistics (each 10° is a window) (Reflecting the randomness of discharge, the higher the entropy value, the closer to the fault state) Smooth the discharge amount collected for 3 times in a row to get the smooth value , filter out transient interference (such as false peak value caused by electromagnetic pulse); Equipment working condition data standardization: normalize various working condition parameters to the [0, 1] interval to facilitate subsequent model calculation: Temperature normalization: ( is the normalized temperature value; is the measured temperature; is the lowest / highest temperature allowed by the device); Load rate calculation: ( is the device load rate; is the real-time current, is the rated current); Environmental temperature correction: when the humidity is > 85%, multiply the discharge amount data by the correction coefficient 1.2 (because high humidity can aggravate discharge); Resource state data quantification: convert node resource state to calculable quantitative indicators: Resource availability: calculate the ratio of resource remaining to total (such as power availability , where is the power availability; is the remaining power; is the total power); Communication quality score: based on the packet loss rate of the last 10 transmissions , (Where, is the communication quality score; is the packet loss rate of the last 10 transmissions) (full score is 100 points, packet loss rate > 2096 is scored as 0); Calculate the load index: (Where, is the calculation load index; is the CPU usage rate; is the memory usage rate) (the average of CPU and memory usage rate, reflecting the calculation pressure); Step S3-4: Data transmission and integrity check: Hierarchical transmission protocol adaptation: select the communication protocol according to the data type: Real-time discharge data and trigger type collection data: use MQTT protocol (QoS=2, ensure that the message is delivered exactly once), transmit through 5G network, transmission delay control ≤100ms; Periodic working condition data and resource status data: LoRaWAN protocol (Class A mode) is adopted, taking into account low power consumption and coverage, transmission period and collection period are synchronized; Large file historical data (such as continuous 1-hour discharge waveform): HTTP protocol is used for fragmented upload, each piece is ≤1MB in size, and supports breakpoint resume; Data encryption and identity authentication: all transmission data uses "end-to-end encryption": Sensor to sink node: AES-128 symmetric encryption is used, the key is dynamically generated by the sink node and updated regularly (every 24 hours); Sink node to cloud: TLS1.3 protocol is used for encrypted transmission, combined with device unique certificate for two-way identity authentication, to prevent data tampering and masquerading attacks; Integrity check and retransmission mechanism: Each data packet is attached with a 32-bit CRC check code, which is verified by the cloud after receiving. If it is inconsistent, it is marked as "damaged packet" and retransmission is requested; Set the retransmission threshold: if the same data packet fails to retransmit 3 times, start the backup communication channel (such as switching from LoRa to 4G), and record the channel quality degradation alarm; For data with high timing requirements (such as discharge phase data), if the transmission delay exceeds 500ms, it is marked as "timeout data" and only used for historical analysis, not for real-time risk calculation; Step S3-5: State information fusion storage and visualization: Multi-source data fusion storage: the cloud database uses a "time series + relationship" hybrid storage architecture: Time series database (InfluxDB): stores real-time discharge, temperature, current and other parameters that change over time, indexed by "monitoring point ID + timestamp", supports millisecond-level queries; Relational database (MySQL): stores device basic information (model, commissioning time), resource configuration parameters (initial frequency, precision level), trigger event records (trigger time, reason, processing result), and associates time series data through foreign keys; Data life cycle management: real-time data is retained for 7 days (high precision), historical data is aggregated at "hour level" and retained for 1 year, and data beyond the retention period is automatically archived to cold storage (queryable but not used for real-time calculation); State information visualization: the cloud platform builds multi-dimensional monitoring interfaces: Real-time status board: displays the location of each monitoring point in a topology graph, uses color to mark the status (green: normal; yellow: attention; red: emergency), and displays real-time discharge, temperature, resource availability and other core parameters when the mouse hovers over; Trend curve analysis: draw the discharge capacity change curve and temperature curve for the past 24 hours, and automatically mark the trigger event point (such as the first level trigger occurs at 10:23); Resource status dashboard: Statistics on the power consumption, communication quality, and computing load distribution of each node, and uses heat maps to display resource-constrained areas (red indicates resource utilization >80%). Instant push of abnormal information: When data abnormalities are detected (such as three consecutive data collection failures or discharge volume exceeding the threshold), the system automatically generates an alarm message and pushes it to the operation and maintenance personnel via SMS and APP. The content includes: monitoring point ID, abnormality type, occurrence time, and current status parameters to ensure timely operation and maintenance response (target response time ≤ 15 minutes).

[0022] In this embodiment, the calculation process of the dynamic risk assessment model in step S4 is as follows: the ratio of the real-time discharge amount to the discharge amount reference threshold is multiplied by the real-time discharge amount weight coefficient, the absolute value of the discharge amount change rate is multiplied by the discharge change rate weight coefficient, and the output value of the mapping function of the device health status on the risk is multiplied by the device health status weight coefficient. The sum of the three is the real-time discharge risk assessment value; wherein, the sum of the real-time discharge amount weight coefficient, the discharge change rate weight coefficient, and the device health status weight coefficient is one; Furthermore, step S4 converts the system status information into a quantitative risk indicator. By integrating multi-dimensional status parameters through a dynamic risk assessment model, the real-time discharge risk level of each monitoring point is accurately calculated, providing a targeted decision-making basis for subsequent resource allocation optimization. The following are the detailed steps: Step S4-1: Preprocessing of model input parameters: Real-time discharge amount smoothing: Real-time discharge amount obtained by S3 , use the sliding average method to eliminate instantaneous interference, take three consecutive acquisitions Calculate the mean , the formula is (in, for Time monitoring point The amount of discharge after smoothing; 、 They are 、 Time monitoring point If there is missing data, it will be supplemented with the most recent valid collected value; Calculation of discharge capacity change rate: Based on the smoothed discharge capacity, the discharge capacity change rate is calculated using the first-order difference method. , the formula is (Wherein, T is the acquisition period; for Time monitoring point The smoothed discharge amount) is taken as the absolute value and used as the model input to reflect the change trend of the discharge amount; Normalization of equipment health status parameters: The equipment operating condition data processed by S3 (such as temperature normalization value, load rate, etc.) are integrated into equipment health status parameters , calculated using the weighted summation method: (in, 、 、 are the normalized temperature, load rate, and operating condition parameters respectively; 、 、 is the fusion weight, the sum is 1), ensuring The value range is [0,1]; Step S4-2: Dynamic risk assessment model parameter configuration Initial setting of weight coefficient: According to the type and importance of the equipment corresponding to the monitoring point, set the risk assessment weight coefficients α, β, and γ: For equipment sensitive to insulation aging (such as cable connectors), set α=0.4, β=0.3, and γ=0.3, focusing on real-time discharge and equipment health status; For equipment sensitive to sudden discharge (such as switchgear), set α=0.3, β=0.4, and γ=0.3, focusing on the rate of change of discharge amount; The weight coefficient can be dynamically adjusted later through S9’s online learning mechanism; Mapping Function Selection: According to the characteristics of the impact of the health status of the equipment on the discharge risk, select the appropriate mapping function: If the impact of the health status deterioration on the risk increases linearly, use the linear function ( is the proportional coefficient, with a value of 1.2-1.5); if it shows nonlinear saturation growth (such as the risk growth rate slows down when the health status is extremely poor), the Sigmoid function is used. (m and n are shape parameters, m is 5-8, n is 0.5), ensure The value range matches other risk components; Discharge reference threshold Determination: Based on the historical fault data of the equipment, the minimum discharge amount when the equipment has obvious insulation degradation is used as If there is no historical data, refer to the standard value of similar equipment (such as first-class equipment Take 1.2 times the typical value, and 1.5 times for secondary equipment); Step S4-3: Calculation of real-time discharge risk assessment value: Substitute the model formula into the calculation: Substitute the preprocessed parameters into the dynamic risk assessment model, and the calculation formula is: ,in, for Time monitoring point The real-time discharge risk assessment value; α, β, γ are risk assessment weight coefficients, the sum of which is 1; for Time monitoring point The amount of discharge after smoothing; is the discharge reference threshold; for Time monitoring point The rate of change of discharge capacity; is the mapping function of the equipment health status to the risk impact; for Time monitoring point Equipment health status parameters); Calculation result range check: If If the calculation result exceeds the range of [0,1], it will be truncated (0 when ≤0, 1 when ≥1); if the result is abnormal due to abnormal parameters (such as If the value is 0 but a high-risk value appears), it is marked as "invalid value" and a parameter recheck is triggered; Step S4-4: Risk level classification and output: Level threshold setting: According to The distribution characteristics of the risk are combined with historical risk-fault correlation data to divide the risk into three levels: Low risk: , ,corresponding to the normal operating state of the equipment, the discharge risk is controllable; Medium risk: , the corresponding equipment has potential insulation problems and needs to be monitored more closely; High risk: ,The corresponding equipment may have obvious discharge faults, and resources need to be allocated first;,the threshold can be dynamically calibrated based on the actual fault situation through S9’s,online learning mechanism; Risk level association output: Generate a binary result of “risk assessment value + level” for each monitoring point (such as Corresponding to "medium risk"), the monitoring point ID and calculation timestamp are associated and stored in the cloud database, and pushed to the input interface of the dynamic resource allocation optimization model as the core input parameters of S5; Step S4-5: Check and correct abnormal results: Cross-time consistency check: Compare 3 consecutive calculations If the difference between two adjacent results is ≥0.4 and there is no obvious change in working conditions (such as equipment load and environment), it is judged as "abnormal fluctuation" and the discharge capacity change rate and , if necessary, recalculate using the weight coefficient of the previous period; Collaborative verification of multiple monitoring points on the same equipment: For multiple monitoring points on the same equipment (such as three monitoring points on a transformer), if the risk level of a monitoring point is 2 levels or higher than that of the other two points (for example, a single point is high risk, and the other two points are low risk), the resource status of the sensor at that point (such as signal strength and power) will be combined to check whether it is caused by a sensor abnormality. If an abnormality is confirmed, the average risk level of other monitoring points on the same equipment will be used as a replacement.

[0023] In this embodiment, the objective function of the dynamic resource allocation optimization model in step S5 is to maximize the overall monitoring efficiency of the system, which is equal to the sum of the risk coverage efficiency and the perception accuracy efficiency of each monitoring point minus the total resource consumption cost of the system; wherein, the risk coverage efficiency is the product of the risk assessment value and the logarithm of the monitoring frequency multiplied by the risk coverage weight factor, and the perception accuracy efficiency is the logarithm of the data collection accuracy multiplied by the perception accuracy weight factor; Resource constraints include: the total computing resource consumption of each monitoring point does not exceed the upper limit of the system's total computing resources; the total bandwidth resource consumption of each monitoring point does not exceed the upper limit of the system's total communication bandwidth; and the total energy resource consumption of each monitoring point does not exceed the upper limit of the system's total energy budget. The resource consumption of each monitoring point is the product of the single resource consumption coefficient of the monitoring point, the monitoring frequency, and the data acquisition accuracy level. Furthermore, step S5 is used to build a "risk-resource-efficiency" linkage optimization framework. Based on the real-time discharge risk level output by S4 and the resource status information obtained by S3, the resource constraint boundaries and system efficiency goals are clarified, and the monitoring strategy optimization problem is converted into a quantifiable mathematical model. This provides a clear goal orientation and constraint basis for the algorithm solution of S6. The detailed steps are as follows: Step S5-1: Quantitative definition of the overall monitoring performance of the system: Definition of risk coverage ratio component: Quantify the demand for “prioritized monitoring of high-risk points” into risk coverage items, the formula is: (in, is the total number of monitoring points; is the risk coverage weight factor; For monitoring points Real-time discharge risk assessment value; The improvement factor of monitoring frequency on performance; For monitoring points The logarithmic function reflects that “the higher the risk, the greater the marginal benefit of frequency increase” (e.g., high risk points hour, The performance gain from 2 to 4 is a low-risk point 4 times of the time); Definition of state perception accuracy component: The contribution of data acquisition accuracy to state cognition is quantified as the perception accuracy term, and the formula is: (in, is the perception accuracy weight factor; is the performance improvement coefficient of collection accuracy; is the data collection accuracy level of the monitoring point ; similarly, a logarithmic function is used to avoid resource waste caused by excessive accuracy improvement (e.g., the performance gain of increasing the accuracy level from 3 to 4 is lower than that of increasing from 1 to 2); Resource consumption efficiency component definition: quantify the loss of resource cost to overall performance as a cost item, with the total resource consumption cost of the system as the characterization, the formula is (wherein, is the resource cost weight factor); Through the integration of computing resources, communication bandwidth, and energy consumption: (wherein, , , are the total computing resources, bandwidth, and energy consumption, respectively; , , are resource type weights, and the sum is 1); Overall performance function integration: integrate the above three parts into the system overall monitoring performance objective function: , to ensure that the objective function reflects the dual demands of "improving risk coverage and perception accuracy" and "reducing resource consumption"; Step S5-2: Definition and quantification of resource constraints: Total computing resource constraint setting: based on the upper limit of the computing power of the cloud and the aggregation node, determine the computing resource constraint formula: (wherein, is the computing resource consumption coefficient of the monitoring point for single collection and processing, related to the sensor type; , are the monitoring frequency and collection accuracy level, respectively; is the upper limit of the total computing resources of the system); Determined by real-time monitoring of server CPU / memory usage (e.g., taking 80% of the total computing power as the upper limit, reserving 20% for sudden computing); Total communication bandwidth constraint setting: according to the actual bearing capacity of the communication network, determine the bandwidth constraint formula: (wherein, is the bandwidth consumption coefficient of the monitoring point for single data transmission, related to the data compression rate; is the upper limit of the total communication bandwidth of the system); Refer to the average network throughput of the past hour for settings (for example, take 90% of the average throughput to avoid network congestion); Total energy budget constraint setting: Combined with the power supply capabilities of the sensor and the sink node (such as battery capacity and mains power stability), determine the energy constraint formula: (in, For monitoring points The energy consumption coefficient of a single acquisition transmission is related to the transmission power; is the upper limit of the total energy budget of the system); For battery-powered nodes, the calculation is based on "average daily energy consumption × remaining flight time" (e.g., if the remaining flight time is 7 days, take 7 times the average daily energy consumption); Dynamic adaptation of constraint parameters: Set a dynamic adjustment mechanism for each constraint condition: when the utilization rate of a certain type of resource is less than 50% for three consecutive cycles (for example, the computing resource utilization rate is only 30%), the corresponding constraint upper limit will be increased by 10% (for example, If the utilization rate exceeds 90% for three consecutive cycles, it is reduced by 10% to ensure that the constraints match the actual resource status; Step S5-3: Determination of performance balance weight and consumption coefficient: Performance balance weight factor 、 、 Setting: Differentiated setting according to system operation objectives: Fault warning priority scenario (such as equipment maintenance period): =0.4, =0.3, =0.3, focusing on risk coverage and perception accuracy; Resource energy saving priority scenario (such as when the battery node is low): Set =0.2, =0.2, =0.6, focusing on controlling resource consumption; The weight factor can be dynamically adjusted based on the historical performance improvement rate through S9's online learning mechanism (if the performance improvement is significant under a certain weight, it will be appropriately increased); Resource consumption coefficient 、 、 Calibration: Determine the base value through experiments and historical data statistics: For each monitoring point, 10 different 、 The resource consumption data under the condition (such as computing resource consumption, bandwidth occupancy, energy consumption) is obtained by linear regression fitting to obtain the consumption coefficient (such as , solved by the least squares method 、 ; The monitoring points with large environmental interference (such as high electromagnetic interference area) will be multiplied by a correction factor of 1.1-1.3 to compensate for the additional resource loss; Monitoring efficiency coefficient 、 Set: Reflect the efficiency of monitoring frequency on risk coverage, high-risk points in the corresponding area are set (Frequency promotion is more effective), low-risk points are set ; Reflect the efficiency of collection accuracy on perception, set (accuracy promotion is more necessary), the state is clear ; Step S5-4: Integration and verification of dynamic resource allocation optimization model: Model variables and boundaries are clear: the decision variables of the model are the 、 (need to meet 、 、 、 、 、 preset adjustment interval for S2); the constraint conditions are the three resource constraints of step S5-2; the objective function is the overall performance function of step S5-1 , forming a complete "variable-constraint-target" model framework; Model feasibility verification: select historical typical scenarios (such as 3 high-risk points + 5 medium-risk points + 10 low-risk points), and substitute the actual resource state into the model to check whether there is a feasible solution: if (that is, it still exceeds the upper limit of resources under the minimum configuration), trigger the resource emergency scheduling (such as temporarily closing unnecessary collection of 2 low-risk points), and mark it as "resource overload" state; if there is a feasible solution, record the constraint tightness of the current model (such as the usage rate of resource constraints is 75%); Model and algorithm adaptability check: confirm that the objective function and constraint condition of the model meet the solution requirements of the improved multi-objective particle swarm optimization algorithm used in S6: whether the objective function is derivable (the logarithmic function is derivable, meeting the algorithm gradient calculation demand), whether the constraint is linear (the resource constraint is all linear inequality, which is convenient for algorithm processing); if there is a nonlinear constraint (such as the communication delay constraint introduced later), linearization processing (such as approximation with piecewise linear function) needs to be done in advance to ensure that the algorithm can be efficiently solved; Step S5-5: Real-time updating mechanism of model parameters: Parameter updating trigger condition: set the trigger scenario of model parameter updating: Periodic update: update the upper limit of resource constraints ( , , ) and consumption coefficients ( , , ) once every 3 optimization periods (e.g., 1 hour per period); event-triggered update: when a monitoring point sensor is replaced (e.g., replacing a UHF sensor), the communication network is upgraded (e.g., switching from 4G to 5G), or the device working condition changes drastically (e.g., the load rate increases from 60% to 90%), the corresponding parameters are updated immediately (e.g., recalibration or adjustment ; Parameter update process: after receiving the parameter update trigger signal, the cloud processing center automatically retrieves the latest resource state data (e.g., server monitoring data, sensor energy consumption data), recalculates and replaces the corresponding parameters in the model, and at the same time, preserves the parameter update record (including update time, reason, old value, new value) to facilitate model tracing and rollback (e.g., when the updated model has no feasible solution, it can be rolled back to the previous version of parameters).

[0024] In this embodiment, the improved multi-objective particle swarm optimization algorithm based on chaotic mapping and adaptive crossover and mutation in step S6 includes the following calculation processes: Chaotic initialization: generate a chaotic sequence using the logistic map, and map the chaotic sequence to the value interval of the decision variable to generate an initial population; Adaptive inertia weight: the inertia weight decreases nonlinearly with the increase of iteration number, and the initial inertia weight is greater than the final inertia weight; Particle velocity and position update: new velocity equals current inertia weight multiplied by current velocity plus individual learning factor multiplied by random number multiplied by the difference between individual historical optimal position and current position plus social learning factor multiplied by random number multiplied by the difference between global optimal position and current position; new position equals current position plus new velocity; Adaptive crossover and mutation: the crossover probability and mutation probability are adjusted adaptively according to the particle fitness value, and the closer the fitness value is to the maximum fitness value, the smaller the crossover probability and mutation probability are; External archive update and guide particle selection: maintain the external archive by using non-dominated sorting and congestion calculation, and select the guide particle from the non-dominated layer; The output of the improved multi-objective particle swarm optimization algorithm based on chaotic mapping and adaptive crossover and mutation is a set of resource allocation scheme solutions that meet the Pareto optimality; In step S6, the resource allocation plan to be finally executed is selected from the Pareto optimal solution set according to the preset decision rule. Specifically, the comprehensive evaluation index of each plan is calculated, which is equal to the normalized system overall monitoring efficiency multiplied by the monitoring efficiency decision weight coefficient plus the difference between one and the normalized system total resource consumption cost multiplied by the resource cost decision weight coefficient; the resource allocation plan with the largest comprehensive evaluation index is selected as the final execution plan; wherein the sum of the monitoring efficiency decision weight coefficient and the resource cost decision weight coefficient is one; Furthermore, step S6 solves the dynamic resource allocation optimization model through an improved multi-objective optimization algorithm. Under the premise of satisfying resource constraints, a Pareto optimal solution set that takes into account both system monitoring efficiency and resource consumption is generated. The final execution plan is selected based on the decision rule, providing a quantitative basis for monitoring strategy adjustment. The detailed steps are as follows: Step S6-1: Initialization of parameters of improved multi-objective particle swarm optimization algorithm: Particle swarm size and dimension setting: according to the number of monitoring points Determine the particle size as (Each monitoring point corresponds to two decision variables: monitoring frequency and acquisition accuracy level ); the particle swarm size is set to (Ensure population diversity, e.g. When the scale is 60), each particle represents a resource allocation plan ; Decision variable range definition: Set boundaries for each decision variable: ( 、 The upper and lower limits of the monitoring frequency preset for S2) ( 、 The upper and lower limits of the accuracy level preset for S2) are used to prevent particles from exceeding the practical feasible range; Initial configuration of algorithm core parameters: setting the maximum number of iterations (Dynamically adjusted with the number of monitoring points); initial inertia weight , the final inertia weight ; Learning factor ; Initial crossover probability , initial mutation probability ;The external archive capacity is set to 100 (the maximum number of non-dominated solutions); Step S6-2: Initial population generation based on chaotic mapping: Logistic chaotic sequence generation: Logistic mapping is used to generate chaotic sequences. The formula is: (in, For the Chaos value of iteration; is the chaos control parameter; For the The chaos value of the iteration, the initial value Randomly select a value in (0,1) that is not equal to 0.25, 0.5, or 0.75) to generate a chaotic sequence with the same length as the particle dimension to ensure sequence ergodicity and randomness; Mapping of chaotic sequences to decision variables: transforming chaotic values Mapped to the value interval of the decision variable, the formula is , (in, 、 is the monitoring frequency and accuracy level after mapping; 、 、 、 is the decision variable boundary), and the mapping results are rounded (the precision level is an integer) to form the initial particles; Feasibility check of the initial population: For each generated particle, substitute the resource constraints of S5 (computing resources, bandwidth, energy constraints) for verification. If the constraints are not met, randomly adjust the decision variables outside the high-risk points (reduce the low-risk points) or ) to make it feasible and ensure that the proportion of feasible solutions in the initial population ; Step S6-3: Adaptive update of particle velocity and position: Adaptive inertia weight calculation: Calculate the inertia weight according to the current iteration number k. The formula is: (in, is the current inertia weight; 、 are the initial and final inertia weights; is the current iteration number; is the maximum number of iterations), achieving nonlinear decrease of weight with iteration (large weight in the early stage enhances global search, small weight in the later stage enhances local optimization); Particle velocity update: according to the formula Update speed (where For particles No. Vidi The speed of iterations; For the The speed of iterations; 、 is the learning factor; 、 is a random number (0,1); For particles No. The optimal position of an individual in the dimension; For particles No. Vidi The position of the iteration; For the global guide particle dimensional position) and impose boundary limits on the velocity , is 20% of the interval of the decision variable); Particle position update: according to the formula Update location (where For particles No. Vidi The position of the iteration is cut off if it exceeds the boundary (such as Time ), rounding off integer variables such as precision level; Step S6-4: Adaptive crossover mutation and external archive update: Adaptive crossover operation: Cross the updated particles with randomly selected elite particles from the external archive, and the crossover probability is Calculate (where, is the crossover probability; is the initial crossover probability; 、 is the maximum and minimum fitness value of the current population; is the fitness value of the particle to be crossed; is a very small constant), particles with low fitness (poor performance) use high crossover probability to promote population evolution; the crossover method uses arithmetic crossover: ( A random number (0,1) For elite particles dimensional position); Adaptive mutation operation: mutate the particles after crossover, and the mutation probability is Calculate (where, is the mutation probability; is the initial mutation probability; is the fitness value of the current particle), particles with low fitness adopt high mutation probability to increase diversity; the mutation method adopts Gaussian mutation: ( With mean 0 and variance Gaussian random numbers, It decreases linearly from 0.2 to 0.05 with iterations); External archive maintenance: Merge the cross-mutated particles with the current external archive, filter non-dominated solutions through non-dominated sorting (i.e., no other solution is better than all other solutions in all objectives), calculate the congestion of non-dominated solutions (the average distance between the solution and the neighboring solution in the same layer; the greater the distance, the lower the congestion), and retain solutions with high congestion until the archive capacity is filled, ensuring that the solutions in the archive are evenly distributed and cover the entire Pareto front. Step S6-5: Guide particle selection and iteration termination judgment: Risk-weighted guide particle selection: selecting guide particles from external archives When the risk level weighting factor is introduced: for each non-dominated solution, calculate its risk weighted value ( For monitoring points Risk assessment value), priority A larger solution (focusing more on resource allocation at high-risk points) If they are the same, the solution with greater congestion is selected to balance risk sensitivity and solution diversity; The iteration is terminated if any of the following conditions are met: Reached the maximum number of iterations ; System performance of the optimal solution in external archiving Change rate for 20 consecutive iterations (convergence and stability); The calculation time exceeds the preset threshold (e.g. 5 seconds, to ensure real-time performance); Pareto optimal solution set output: The non-dominated solutions in the final external archive are used as the Pareto optimal solution set, each solution contains the 、 and corresponding system performance , total resource consumption , providing alternatives for subsequent decision-making; Step S6-6: Decision-making on the final resource allocation plan: Calculation of comprehensive evaluation index: Calculate the comprehensive evaluation index for each solution in the Pareto optimal solution set (in, is a comprehensive evaluation indicator; 、 is the decision weight coefficient and ; is the normalized system performance; is the normalized total resource consumption); 、 Adjustment according to the operating mode: Fault warning mode 、 Energy saving mode ; Optimal scheme selection and verification: select The largest scheme as the final execution scheme, and the resource constraint condition is substituted for secondary verification (to ensure actual feasibility). If there is a constraint violation, select a suboptimal scheme until a feasible scheme is found; output the scheme including the calculation resource share , communication bandwidth share and corresponding , as the adjustment basis of S7.

[0025] In this embodiment, the adjustment of the monitoring behavior of each monitoring point in step S7 specifically includes: the aggregation node generates specific scheduling instructions according to the received monitoring task execution parameters and sends them to the corresponding sensor nodes; the sensor node adjusts its signal acquisition frequency, sampling rate and signal preprocessing algorithm complexity according to the scheduling instructions; Further, the role of step S7 is to convert the optimal resource allocation scheme generated in S6 into actual monitoring behavior adjustment instructions, and through the collaborative execution of the cloud, the aggregation node and the sensor node, to realize the dynamic adaptation of the monitoring strategy of each monitoring point and ensure that the resource allocation optimization result takes effect; the following are the detailed steps: Step S7-1: Resource allocation scheme instruction packaging and delivery: Scheme instruction standardized packaging: the cloud processing center converts the final resource allocation scheme (including the , , calculation resource share, bandwidth share, etc. of each monitoring point) into standardized instructions, which are stored in a JSON format structure, and the fields include: instruction ID (unique identification), target monitoring point ID list, each monitoring point's (monitoring frequency), (acquisition accuracy level), execution start time (accurate to seconds), instruction validity period (default 30 minutes); a digital signature (based on device unique key generation) is added to the instruction to prevent tampering; Hierarchical delivery path planning: according to the hierarchy of “cloud→aggregation node→sensor node”: From the cloud to the aggregation node: use encrypted HTTP protocol (TLS1.3) transmission, preferentially select 5G communication channel (time delay ≤100ms), and switch to 4G if the 5G signal is weak; when delivering, send by range according to the jurisdiction of the aggregation node (such as one aggregation node corresponds to 10 monitoring points, then pack separately) to avoid instruction congestion; From the aggregation node to the sensor node: use LoRaWAN protocol (Class C mode, supports instant downlink), allocate independent time slots (time slot length 20ms) for each sensor node, and deliver according to the risk level of the monitoring point (high-risk point instructions are sent first); Delivery progress tracking: The cloud records the status of command delivery in real time ("pending", "sending", "delivered", "executed"). After receiving the command, the aggregation node and sensor node must return a "receipt confirmation" (including the command ID and local timestamp) within 1 second. If no confirmation is received within 3 seconds, the cloud automatically retransmits (up to 3 times). If the retransmission still fails, it is marked as "delivery abnormality" and triggers manual investigation. Step S7-2: Instruction parsing and scheduling at the aggregation node: Instruction integrity check: After receiving an instruction, the sink node first verifies the digital signature (comparing it with the locally stored device public key) and the instruction format (checking whether the required fields are complete). If the verification fails, the execution is rejected and a "verification error" is returned. After the verification passes, the parameters of each monitoring point in the instruction are parsed and matched with the locally stored monitoring point ID list (invalid IDs are eliminated) to generate a "monitoring point-parameter" mapping table. Local resource adaptation and fine-tuning: The aggregation node fine-tunes the instruction parameters locally based on its own real-time resource status (such as current CPU usage and remaining bandwidth): If the computing resource share corresponding to a monitoring point exceeds the current remaining computing capacity of the aggregation node (e.g. the instruction requires a 20% share, but only 15% is left), the computing resource share of the monitoring point will be reduced proportionally. (e.g. from 5 times / minute to 4 times / minute), while maintaining Unchanged (prioritizing accuracy); If the communication bandwidth is insufficient (e.g. the total bandwidth requirement of the instruction exceeds 10% of the current available bandwidth), the low risk point ( )of Temporarily reduce it by 20% and record the fine-tuning amount (which will be fed back to the cloud for model optimization later); Scheduling instruction generation and distribution: For each sensor node, the sink node generates detailed scheduling instructions, including: local sampling rate (according to OK, if Corresponding to 50MHz), acquisition cycle (1 / ), data compression algorithms (such as Use lossless compression when high and lossy compression when low), pre-processing task type (such as whether to enable phase distribution entropy calculation); distribute through the sensor node's dedicated communication port (such as UART interface) to ensure the command transmission bit error rate ; Step S7-3: Adjusting the monitoring behavior of the sensor node: Parameter configuration update: After receiving the scheduling instruction, the sensor node writes the new parameters to the local configuration file (overwriting the original initial parameters) and simultaneously updates the hardware registers: Adjust the sampling frequency: Reset the sampling trigger interval through the timer (such as times / minute (set 12-second interval) to ensure trigger error ms; Adjust the acquisition accuracy: Press Switching sampling circuit gain (such as When the gain is set to 20dB, 10dB) and update the ADC sampling bit (16 bits / 14 bits / 12 bits); Adjust the preprocessing algorithm: If Improve (high precision), enable high-order filtering algorithms (such as wavelet denoising); if Reduce, switch to basic filtering (such as mean filtering) to reduce energy consumption; Self-test of the adjusted state: After the sensor node completes the parameter update, it performs a test acquisition: it collects the signal of one power frequency cycle, checks whether the actual sampling rate and accuracy are consistent with the instructions (if the sampling rate deviation is ≤5%, it is qualified), and measures the current energy consumption (compared with the state before adjustment, the deviation should be ≤10%). If the self-test passes, it records "adjustment completed"; if it fails, it rolls back to the parameters before adjustment and returns "adjustment failed" (the aggregation node reissues the instruction); Historical parameter archive: The sensor node will adjust the parameters before and after ( 、 , sampling rate, etc.) and the adjustment timestamp are stored in non-volatile memory (such as Flash), and the latest 10 adjustment records are retained to facilitate subsequent fault tracing (for example, in the event of abnormal discharge, it is possible to verify whether the parameters are correctly applied); Step S7-4: Real-time verification of adjustment effect: Verification of the first data collection: After the adjustment is completed, the sensor node performs the first data collection according to the new parameters, and the sink node verifies the data after receiving it: Integrity check: Check the number of data frames (with Match, such as times / minute, 5 frames should be received within 1 minute); Accuracy verification: Extract discharge characteristics from the data (such as peak ), compared with the characteristics under the same working conditions before adjustment (high precision The feature details should be richer, such as a more complete phase distribution); Resource consumption verification: Statistics on the actual bandwidth usage of the monitoring point after adjustment (such as the size of each frame data ) and energy consumption (measured by current sensor) to confirm that they do not exceed the instruction limit; State feedback chain construction: The aggregation node summarizes the verification results (including "pass / fail" identification, actual parameter values, and resource consumption data) and uploads them to the cloud through the 5G / 4G channel; the cloud compares the feedback results with the original allocation plan and calculates the "parameter execution deviation rate" (such as ), if the deviation rate is less than or equal to 10%, the label "adjustment effective" is marked, otherwise the sample is included in the sample library of S9 online learning (used to optimize the subsequent allocation model); Step S7-5: emergency treatment of abnormal scenarios: Sensor node unresponsive: if the sink node does not receive the collected data of a certain sensor node for 3 consecutive times (and confirms that the instruction has been sent), it is determined that the "node is unresponsive", and the backup strategy is immediately started: If the monitoring point is a high-risk point ( ), the parameter improvement scheme of other monitoring points of the same device (if any) is enabled (such as increasing the adjacent monitoring points by 50%), which compensates for the monitoring blind area; If it is a low-risk point, the collection task of the node is suspended, its resource share is allocated to other nodes, and a "node failure" alarm is pushed to the operation and maintenance terminal; Parameter adjustment conflict processing: if multiple sensor nodes appear resource competition after adjustment (such as simultaneously occupying the communication channel), the sink node re-allocates the communication time slot through dynamic time division multiplexing (high-risk point time slot priority > low-risk point), and reduces the of the conflict nodes (temporarily reduced by 10%) until the conflict is resolved; the conflict processing process needs to be completed within 2 seconds to avoid affecting the monitoring continuity; Emergency scenario forced adjustment: if the S3 first trigger condition is triggered during the adjustment process (such as a sudden increase in the discharge amount of a certain monitoring point), the sink node can temporarily interrupt the current adjustment process and execute the emergency collection instruction (such as increasing the point to the maximum allowed value), and then restore the original adjustment plan after the emergency state is resolved, to ensure the monitoring reliability in abnormal situations.

[0026] In this embodiment, step S9 realizes continuous optimization of model parameters through the construction of a dynamic learning mechanism, so that the system can adapt to the evolution of device state and environmental changes, and the core includes four parts: First, the dynamic sample library is constructed, complete and representative samples are selected from historical data (covering different risk levels, device states and environmental conditions), and derivative features such as risk trend slope and resource efficiency are expanded, and old data is eliminated according to timeliness (preferably within 3 months, and 10% of historical data is eliminated every month), to provide high-quality input for learning; Second, the model parameter iteration, for the risk evaluation model, the weight coefficients a, b, and g are optimized by stochastic gradient descent, and the mapping function parameters are adjusted by maximum likelihood estimation; for the resource allocation model, the performance balance weight is optimized by reinforcement learning (Q-Learning), and the resource consumption coefficient is calibrated by sliding window least squares method, and the particle swarm algorithm control parameters are adaptively adjusted according to the convergence speed of the algorithm; Third, the effectiveness of the verification, first by offline validation set test (risk prediction accuracy requirement to improve ≥3% or resource efficiency to improve ≥2%), and then select typical monitoring points for 24 hours online pilot, verify the risk coverage, resource efficiency and fault warning effect, up to the full system promotion; Fourth, the whole system synchronization, according to the priority (core parameter real-time synchronization, general parameters daily / weekly synchronization) using RSA + AES encryption transmission, through the chain synchronization mechanism to push to each node, while establishing version management and rollback mechanism, if the performance decreases after the update to restore to stable version, ensure the long-term reliable operation of the system.

[0027] The partial discharge monitoring strategy optimization system based on dynamic resource allocation, and the steps of the optimization system execution method.

[0028] Although embodiments of the present application have been shown and described, it will be understood by those having ordinary skill in the art that various changes, modifications, alternatives, and variations can be made thereto without departing from the principles and spirit of the application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A partial discharge monitoring strategy optimization method based on dynamic resource allocation, characterized by: The following steps are involved: S1. Build a partial discharge monitoring system. The partial discharge monitoring system includes sensor nodes deployed at multiple monitoring points, aggregation nodes, and a cloud processing center. The sensor nodes are used to collect partial discharge signals. The aggregation nodes are used to receive and pre-process data from the sensor nodes. The cloud processing center is used to execute the optimization algorithm of the monitoring strategy and issue control instructions. S2. Initialize the monitoring strategy and configure the initial monitoring frequency and data collection accuracy for each monitoring point; S3. The cloud processing center periodically or triggers the acquisition of current system status information, which includes real-time discharge data of each monitoring point, historical discharge trend data, equipment operating condition data, and available resource status of sensor nodes and communication networks. S4. Based on the system status information, the real-time discharge risk level of each monitoring point is calculated through a dynamic risk assessment model; The dynamic risk assessment model integrates real-time discharge amount, discharge trend change rate and equipment health status parameters to output a quantitative risk assessment value; S5. Based on resource constraints and the real-time discharge risk level of each monitoring point, a dynamic resource allocation optimization model is constructed with the goal of maximizing the overall monitoring efficiency of the system. Resource constraints include total computing resource constraints, total communication bandwidth constraints, and total energy budget constraints. Monitoring efficiency is defined by risk coverage, state perception accuracy, and resource consumption efficiency. S6. Solve the dynamic resource allocation optimization model using a multi-objective optimization algorithm to obtain the optimal resource allocation plan for each monitoring point in the next cycle; the multi-objective optimization algorithm is an improved multi-objective particle swarm optimization algorithm based on chaotic mapping and adaptive crossover mutation; the resource allocation plan includes at least the computing resource share and communication bandwidth share allocated to each monitoring point, and the monitoring task execution parameters determined accordingly, wherein the monitoring task execution parameters include the monitoring frequency and data acquisition accuracy; S7, sending the resource allocation plan to the corresponding sink node and sensor node, and adjusting the monitoring behavior of each monitoring point; S8. Repeat steps S3 to S7 to achieve dynamic closed-loop optimization of the monitoring strategy.

2. The method for optimizing partial discharge monitoring strategy based on dynamic resource allocation according to claim 1, characterized in that: The calculation process of the dynamic risk assessment model in step S4 is as follows: multiply the ratio of the real-time discharge amount to the discharge amount reference threshold by the real-time discharge amount weight coefficient, add the absolute value of the discharge amount change rate multiplied by the discharge change rate weight coefficient, and add the output value of the mapping function of the impact of the equipment health status on the risk multiplied by the equipment health status weight coefficient. The sum of the three is the real-time discharge risk assessment value; wherein, the sum of the real-time discharge amount weight coefficient, the discharge change rate weight coefficient, and the equipment health status weight coefficient is one.

3. The method for optimizing partial discharge monitoring strategy based on dynamic resource allocation according to claim 2, characterized in that: The objective function of the dynamic resource allocation optimization model in step S5 is to maximize the overall monitoring efficiency of the system, which is equal to the sum of the risk coverage efficiency and the perception accuracy efficiency of each monitoring point minus the total resource consumption cost of the system; wherein the risk coverage efficiency is the product of the risk assessment value and the logarithm of the monitoring frequency multiplied by the risk coverage weight factor, and the perception accuracy efficiency is the logarithm of the data collection accuracy multiplied by the perception accuracy weight factor; The resource constraints include: the total computing resource consumption of each monitoring point does not exceed the upper limit of the system's total computing resources, the total bandwidth resource consumption of each monitoring point does not exceed the upper limit of the system's total communication bandwidth, and the total energy resource consumption of each monitoring point does not exceed the upper limit of the system's total energy budget; wherein the resource consumption of each monitoring point is the product of the single resource consumption coefficient of the monitoring point, the monitoring frequency, and the data acquisition accuracy level.

4. The method for optimizing partial discharge monitoring strategy based on dynamic resource allocation according to claim 3, characterized in that: The calculation process of the improved multi-objective particle swarm optimization algorithm based on chaotic mapping and adaptive crossover mutation in step S6 includes: Chaos initialization: Use logistic mapping to generate a chaotic sequence, and then map the chaotic sequence to the value interval of the decision variable to generate the initial population; Adaptive inertia weight: The inertia weight decreases nonlinearly with the number of iterations, and the initial inertia weight is greater than the final inertia weight; Particle speed and position update: The new speed is equal to the current inertia weight multiplied by the current speed plus the individual learning factor multiplied by the random number multiplied by the difference between the individual's historical optimal position and the current position plus the social learning factor multiplied by the random number multiplied by the difference between the global optimal position and the current position; the new position is equal to the current position plus the new speed; Adaptive crossover and mutation: The crossover probability and mutation probability are adaptively adjusted according to the particle fitness value. The closer the fitness value of the particle is to the maximum fitness value, the smaller the crossover probability and mutation probability; External archive update and guide particle selection: Use non-dominated sorting and crowding calculation to maintain the external archive and select guide particles from the non-dominated layer; The output of the improved multi-objective particle swarm optimization algorithm based on chaotic mapping and adaptive crossover mutation is a resource allocation solution set that satisfies Pareto optimality.

5. The method for optimizing partial discharge monitoring strategy based on dynamic resource allocation according to claim 4, characterized in that: In step S6, the resource allocation plan to be finally executed is selected from the Pareto optimal solution set according to the preset decision rules, specifically: calculating the comprehensive evaluation index of each plan, which is equal to the normalized system overall monitoring efficiency multiplied by the monitoring efficiency decision weight coefficient plus the difference between one and the normalized system total resource consumption cost multiplied by the resource cost decision weight coefficient; selecting the resource allocation plan with the largest comprehensive evaluation index as the final execution plan; wherein, the sum of the monitoring efficiency decision weight coefficient and the resource cost decision weight coefficient is one.

6. The method for optimizing partial discharge monitoring strategy based on dynamic resource allocation according to claim 1, characterized in that: Adjusting the monitoring behavior of each monitoring point in step S7 specifically includes: the aggregation node generates a specific scheduling instruction based on the received monitoring task execution parameters and sends it to the corresponding sensor node; the sensor node adjusts its own signal acquisition frequency, sampling rate and signal preprocessing algorithm complexity according to the scheduling instruction.

7. The method for optimizing partial discharge monitoring strategy based on dynamic resource allocation according to claim 1, characterized in that: The trigger conditions for triggering the acquisition of the current system status information in step S3 include: the real-time discharge amount of any monitoring point exceeds a preset threshold, a sudden change signal of the equipment operation status is received, or the sensor node resource availability is lower than a safety threshold.

8. The method for optimizing partial discharge monitoring strategy based on dynamic resource allocation according to claim 3, characterized in that: The method further comprises: S9. Through an online learning mechanism, based on historical monitoring data and resource allocation effects, dynamically adjust the risk assessment weight coefficient in the dynamic risk assessment model and the efficiency balance weight factor in the objective function of the dynamic resource allocation optimization model.

9. Partial discharge monitoring strategy optimization system based on dynamic resource allocation, characterized by: The optimization system executes the steps of the method according to any one of claims 1 to 8.

Images