Grounding grid grounding resistance value change on-line monitoring system

Abstract

The invention discloses a grounding grid grounding resistance value change on-line monitoring system, and relates to the field of electrical safety monitoring. The system comprises a field monitoring device, a communication network and a cloud service platform. The on-site monitoring device integrates a dynamic uncertainty evaluation and fusion module and a communication-energy joint optimization scheduling module, and full-scene high-precision measurement is realized through real-time quantification of measurement errors of a potential drop method and an electromagnetic induction method and adaptive weighted fusion. In a Beidou short message mode, transmission content is optimized according to data value density, and reliable return of key data in an extreme environment is guaranteed. The cloud service platform is provided with a multi-stage risk early warning engine based on failure physics, and hierarchical risk early warning is realized by performing hybrid model decomposition on resistance time sequence data. The industrial problems that the precision is limited by scenes, data transmission in an extreme environment is unreliable and risk early warning is lagged and high in false alarm in grounding resistance monitoring are systematically solved.

Description

Technical Field

[0001] This invention belongs to the field of online monitoring technology for electrical equipment status, specifically an online monitoring system for changes in the grounding resistance value of a grounding grid. Background Technology

[0002] As a core infrastructure ensuring the safe and stable operation of the power system, the grounding grid's grounding resistance value is the only quantitative indicator for measuring fault current discharge capacity and assessing the system's safety status. Abnormal changes in grounding resistance are directly related to potential faults such as grounding electrode corrosion, loose connection points, or external damage. Accurate, reliable, and continuous online monitoring of grounding resistance is a technical prerequisite for achieving predictive maintenance and avoiding major safety accidents.

[0003] Existing online monitoring technologies face three fundamental bottlenecks caused by a combination of physical principles and engineering constraints: First, there is a contradiction between the inherent limitations of measurement principles and complex field environments. Among mainstream measurement methods, the potential drop method (three-electrode method) requires the voltage and current electrodes to be placed at "infinity," a condition that cannot be met in space-constrained environments such as mountainous areas and urban areas. Lead mutual inductance and layout deviations will introduce non-negligible systematic errors. While the electromagnetic induction method (clamp meter method) is easy to implement, its measurement effectiveness strictly depends on the formation of a closed loop and a low-interference environment. In areas with complex grounding grid structures or strong electromagnetic interference, the measurement signal-to-noise ratio deteriorates sharply. Existing attempts using dual principles are mostly limited to simple mode switching or fixed-weight fusion, failing to establish a dynamic quantitative relationship between measurement uncertainty and real-time environmental parameters, and thus unable to achieve adaptive optimal estimation across all scenarios in principle.

[0004] Second, there is a contradiction between the requirement for continuous data transmission and the extremely limited communication resources. For monitoring points deployed in remote mountainous areas, deserts, and other areas without public network coverage, BeiDou short message service becomes the only means of communication. However, it suffers from physical constraints such as stringent channel capacity, limited daily communication frequency, and high transmission power consumption. Traditional data caching and polling retransmission mechanisms essentially treat communication as a transparent channel, ignoring the core optimization question of "which data to transmit to maximize efficiency" under resource boundary conditions. This leads to systemic risks such as the loss of critical abnormal data and premature energy depletion.

[0005] Third, there is a contradiction between the need for predictive risk warnings and the interference of mixed signals. The time-varying characteristics of grounding resistance are a complex process driven by multiple factors, including soil physicochemical properties, electrical connection status, and environmental temperature and humidity. Currently widely used static thresholds or simple differential alarms cannot separate the long-term trend components representing structural degradation and the periodic components reflecting environmental disturbances from the time-series signals. Therefore, they cannot distinguish between normal fluctuations and early fault symptoms, resulting in frequent false alarms or delayed warnings, and losing the decision-making window for predictive maintenance.

[0006] Therefore, there is an urgent need for a monitoring system that starts from the aforementioned contradictory nature and innovates its systematic principles at the three levels of perception, transmission, and analysis, in order to break through the existing technological bottlenecks. Summary of the Invention

[0007] The purpose of this invention is to provide an online monitoring system for changes in the grounding resistance value of a grounding grid, in order to solve the problems in the prior art mentioned in the background, such as measurement accuracy being limited by the environment, unreliable data transmission under extreme conditions, and lack of predictability in risk warning.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: An online monitoring system for changes in grounding grid resistance includes a field monitoring device, a communication network, and a cloud service platform; The field monitoring device is used to collect and process grounding resistance data. It integrates a dynamic uncertainty assessment and fusion module as well as a communication and energy joint optimization scheduling module. A communication network is used to provide a data transmission channel between field monitoring devices and cloud service platforms; The cloud service platform is used to receive and store data and perform risk analysis and early warning. It has a multi-stage risk warning engine based on failure physics deployed inside. in, The dynamic uncertainty assessment and fusion module is configured to simultaneously perform measurements using the potential drop method and the electromagnetic induction method. Based on the measurement uncertainties of the two methods assessed in real time, it uses an inverse variance weighted model to perform dynamic data fusion and outputs the optimal estimated resistance value. The communication and energy joint optimization scheduling module is configured to, in the BeiDou short message communication mode, decide the content of the data to be sent by solving an optimization problem based on the constructed data value density model and under the constraints of message length, number of transmissions and remaining energy. The multi-stage risk warning engine based on failure physics is configured to perform hybrid model decomposition on grounding resistance time series data, extract trend acceleration, deperiodic residual cumulative sum and mutation detection statistics, and trigger graded warnings based on the logical combination of the above feature parameters.

[0009] According to the above technical solution, the dynamic uncertainty assessment and fusion module includes: The uncertainty assessment submodule is configured as follows: Calculate the spacing distortion factor using the potential drop method: in The actual wiring length obtained through the environmental sensing unit. To satisfy the theoretical minimum length of the semi-infinite field assumption, and based on the pre-defined mapping relationship fv Obtain the estimated standard deviation of the measurement error using the potential drop method. ; Calculate the signal-to-interference-plus-noise ratio (SIR) of the electromagnetic induction method: in The signal current of the circuit under test, Interference current from adjacent circuits, The background noise equivalent current is determined based on a preset mapping relationship. f c Obtain the estimated standard deviation of the measurement error using the electromagnetic induction method. ; The data fusion submodule is configured to calculate dynamic weights and perform fusion using the following formula: in, R v and R c These are the measured values ​​using the potential drop method and the electromagnetic induction method, respectively. R fused This represents the optimal estimated resistance value after fusion.

[0010] According to the above technical solution, the on-site monitoring device also includes an environmental sensing unit, which includes at least a tilt sensor and an odometer for real-time estimation of the actual wiring length. d aCtual Electromagnetic induction measurement employs a dual-coil differential clamp structure.

[0011] According to the above technical solution, the communication and energy joint optimization scheduling module includes: The value calculation submodule is configured to calculate the information value density Vi for each piece of data i to be transmitted in the cache queue, using the following formula: in, For data delay time, This is the resistance measurement value. This is the historical baseline value. The historical correlation coefficient with the data of the j-th geographically or electrically adjacent monitoring point. These are configurable weighting coefficients. It is the attenuation constant; The optimization decision submodule is configured to solve the following 0-1 knapsack problem to determine the sending set each time a wake-up call is sent: Objective function: ; Constraints: C1: ; C2: e≤E; C3: N>0; C4:x i ∈{0,1}; in, Let i be the number of bytes after compression. Here, e represents the maximum capacity of the BeiDou short message service, e represents the energy consumption of the BeiDou module per transmission, E represents the current remaining energy of the device, N represents the remaining number of transmissions available for the day, and x represents the maximum capacity of the BeiDou short message service. i These are decision variables.

[0012] According to the above technical solution, the communication and energy joint optimization scheduling module also includes a communication mode switching submodule, configured to: monitor the reference signal received power (RSRP) of the 4G network in real time; when RSRP remains below a first threshold T low When the preset duration t1 is reached, the control field monitoring device switches to BeiDou short message communication mode and activates the optimization decision submodule; when RSRP recovers to a level higher than the second threshold T... high When the preset duration t2 is reached, the control field monitoring device switches back to 4G communication mode, where T... high >T low .

[0013] According to the above technical solution, the multi-stage risk warning engine based on failure physics includes: The sequence decomposition submodule is configured to decompose the resistance time series data R(t) into a trend term T(t), a periodic term S(t), and a residual term. Satisfying: R(t) = T(t) + S(t) + The trend term T(t) is used to characterize the structural deterioration of the grounding electrode, and the periodic term S(t) is used to characterize the periodic influence of environmental factors. The feature extraction submodule is configured as follows: Calculate the second discrete derivative of the trend term T(t) as the trend acceleration a. t ; Calculate the cumulative sum CUSUM statistic C for the deperiodic residual sequence R(t)−S(t). t ; The sequential probability ratio test (SPRT) was performed on the original resistance time series data R(t) to obtain the mutation detection statistic J. t .

[0014] According to the above technical solution, the multi-stage risk warning engine based on failure physics also includes a warning triggering submodule, which is configured to execute hierarchical warning logic: When the condition is met: |a t ∣>Th a1 Or Ct >Th c1 At that time, a Level 1 warning was triggered; When the condition is met: |a t ∣>Th a2 And C t When the growth continues for more than a preset duration, a level-two warning is triggered, in which Th a2 >Th a1 ; When condition: J t Exceeding the preset mutation threshold or R(t)>Th abs At that time, a level 3 alarm was triggered.

[0015] According to the above technical solution, the field monitoring device has a built-in power supply unit and a main control unit. The main control unit supports low-power sleep mode and timed wake-up mode, and works in conjunction with the communication and energy joint optimization scheduling module to ensure communication optimization scheduling under energy constraints.

[0016] Based on the above technical solution, the cloud service platform adopts a microservice architecture, which includes at least: The device access gateway group is used to concurrently receive and parse data from different communication protocols, and to standardize the data based on a predefined object model; Message middleware is used to asynchronously decouple data streams and distribute standardized data to downstream services. The computing and alerting service, as a containerized microservice, runs a multi-stage risk warning engine based on failure physics. The hierarchical data storage cluster includes Redis for caching real-time data, MySQL for storing business relationship data, and TDengine for storing historical time-series data.

[0017] According to the above technical solution, the cloud service platform specifically includes: The device access gateway group is used to concurrently receive and parse data from different communication protocols, and to standardize the data based on a predefined object model; The message middleware connects to the device access gateway group and is used for asynchronous distribution of standardized data; The computing and alerting service, connected to the message middleware, is a containerized microservice that runs a multi-stage risk warning engine based on failure physics. A tiered data storage cluster, connected to computing and alarm services, includes a cache database for storing real-time data, a relational database for storing business relationship data, and a time-series database for storing historical time-series data.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This invention achieves a full-link performance improvement from precise on-site perception and reliable data transmission to intelligent cloud-based early warning through the closed-loop collaboration and system-level optimization of the three core modules mentioned above. Specifically: the fusion perception module based on dynamic uncertainty assessment effectively overcomes the limitations of single measurement methods in complex terrain and electromagnetic environments through an adaptive weighted fusion strategy, significantly reducing the overall measurement error across all scenarios; the communication and energy joint optimization scheduling module ensures highly reliable uploading of critical status information by intelligently selecting and compressing key data under extreme communication conditions, while significantly extending the continuous operating time of the device in passive scenarios through an energy-efficient scheduling strategy; the multi-stage risk early warning engine based on failure physics achieves a fundamental shift from passive threshold alarms to early trend prediction by deeply decomposing time-series signals and extracting forward-looking features, greatly improving early warning accuracy and effectively identifying latent faults, providing a critical time window for preventative maintenance. The overall system provides an efficient and reliable systematic technical solution for solving the measurement, communication, and early warning challenges in grounding grid status monitoring. Attached Figure Description

[0019] Figure 1 This is a structural block diagram of the online monitoring system of the present invention; Figure 2 This is a structural block diagram of the uncertainty assessment and fusion module of the present invention; Figure 3 This is a flowchart of the communication and energy joint optimization scheduling module of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1 like Figure 1 As shown, an online monitoring system for changes in grounding grid resistance includes a field monitoring device, a communication network, and a cloud service platform. The field monitoring device is responsible for collecting, processing, and uploading grounding resistance data locally; the communication network provides a primary / backup heterogeneous data transmission channel between the field monitoring device and the cloud service platform; and the cloud service platform is responsible for data aggregation, storage, in-depth analysis, and visual early warning.

[0022] This embodiment details the composition and workflow of the dynamic uncertainty assessment and fusion module in the field monitoring device. Its core lies in establishing and applying the dynamic mapping relationship between measurement uncertainty and environmental parameters. The specific implementation of the dynamic uncertainty assessment and fusion module is as follows: like Figure 2 As shown, the core of this module lies in establishing and applying the dynamic mapping relationship between measurement uncertainty and environmental parameters, which consists of a hardware measurement unit, an environmental sensing unit, an uncertainty mapping unit, and a fusion calculation unit forming a closed loop.

[0023] Before describing the various embodiments in detail, it is necessary to first clarify the structural relationship between the modules and sub-modules described in the claims in specific technical implementations. Those skilled in the art should understand that the modules named by function in this application (such as the dynamic uncertainty assessment and fusion module) and their sub-modules (such as the uncertainty assessment sub-module and the data fusion sub-module) are general descriptions of the technical solutions. In specific implementations, these functional modules are implemented through the organic combination of hardware circuits, embedded software algorithms, and stored data. Specifically: The uncertainty assessment submodule is implemented in this embodiment through the collaborative work of the environment sensing unit and the uncertainty mapping unit. The environment sensing unit is responsible for collecting raw physical quantities (such as displacement, attitude, and signal spectrum) to provide input parameters for the assessment; the uncertainty mapping unit, based on these parameters, outputs a quantified uncertainty estimate by executing pre-stored calculation logic or querying pre-set mapping relationships. ).

[0024] In this embodiment, the data fusion submodule is implemented by the fusion computing unit (usually a microprocessor) by running the described inverse variance weighting algorithm software.

[0025] The aforementioned environmental sensing unit, uncertainty mapping unit, and fusion calculation unit together constitute the physical and logical entities that realize the complete functionality of the dynamic uncertainty assessment and fusion module. The working methods of these specific units will be described in detail below.

[0026] The hardware measurement unit includes independent but synchronously triggered potential drop method and electromagnetic induction method sub-modules. The potential drop method module is equipped with a constant current source (output frequency of 55Hz to avoid power frequency interference, current I0=5A±0.1%), a high-precision voltage acquisition circuit (24-bit ADC, input impedance>10MΩ), and a relay matrix for switching the current electrode (C) and voltage electrode (P). The electromagnetic induction method module uses a clamp-on current sensor based on a dual-coil differential structure (sensitivity 1mV / A, bandwidth 10Hz-1kHz), a lock-in amplifier circuit, and a signal conditioning circuit. The environmental perception unit is equipped with a nine-axis inertial measurement unit (IMU), which uses dead reckoning combined with initial GPS positioning information to calculate the movement trajectory of the measurement electrodes in real time and accurately calculate the actual wiring path length. d aCtual The angle of deviation from the ideal straight line provides key input parameters for uncertainty assessment. The uncertainty mapping unit internally stores two experimentally calibrated mapping functions. and ,in By simulating different grounding fields in a controlled grounding field in the laboratory Routing conditions with values ​​(0.5 to 1.2), each Perform at least 100 repeated measurements at the specified value and calculate the standard deviation of the potential drop method measurement results. Then use the least squares method to ( The data points were fitted to piecewise linear or quadratic functions, and the calibration results show... <0.8 Follow Decrease the approximate exponential increase; In an experimental environment with a controllable source of interference, the signal-to-interference-plus-noise ratio is changed. Record the standard deviation of the electromagnetic induction measurement values. By fitting, an inverse proportional function relationship is established ( The lower the value, the stronger the interference. (The larger the value). The fusion computing unit uses an embedded microprocessor (such as an ARM Cortex-M4 core) to receive raw values ​​from the hardware measurement unit in real time. R v , R c and the output of the mapping unit The weights are calculated using the following formula: in, This is an estimate of the standard deviation of the measurement error. The standard deviation of the measurement error is the estimated value; the output fusion result is shown in the following formula: At the same time, the original data, The process data is packaged together for subsequent transmission.

[0027] This embodiment provides a specific implementation method. Assuming the system is deployed in a substation with undulating terrain, after installation, the main control unit reads 60% of the diagonal length of the substation's grounding grid from preset parameters as... (For example, 120 meters), the implementation process is as follows: After the device starts a measurement, a constant current source injects a 55Hz, 5A current into the ground electrode (G) and the current electrode (C). The voltage acquisition circuit measures the potential difference V between G and the voltage electrode (P) and calculates it. R v The specific formula is as follows: In the formula, V represents the potential difference and I represents the current.

[0028] Simultaneously, the clamp sensor is attached to the grounding lead, and the 55Hz current signal I is extracted through phase-locked loop amplification. clamp Calculate based on the known test voltage inside the device R c Meanwhile, the IMU records the trajectory of the device moving from point G to point P and then to point C, and calculates the total length of the actual path. d aCtual =78 meters.

[0029] Then calculate As shown in the following formula: Query the pre-stored mapping function through the uncertainty mapping unit. get (0.65) = 0.48Ω. The signal processing unit analyzes the clamp meter signal spectrum to obtain the current... =12, query get (12) = 0.12Ω.

[0030] The fusion computing unit performs weight calculation: =(1 / 0.48 2 ) / (1 / 0.48 2 +1 / 0.12 2 )≈0.06, =1-0.06=0.94, if this time R v =1.8Ω R c =2.5Ω, then R fused=0.06×1.8+0.94×2.5≈2.43Ω. During this process, the system automatically identified that terrain limitations caused the uncertainty of the potential drop method to be much higher than that of the electromagnetic induction method. By dynamically adjusting the weights, the system effectively corrected the significant negative bias of the potential drop method caused by insufficient wiring, obtaining measurement results closer to the true value.

[0031] Example 2 This embodiment provides a specific implementation of a communication and energy joint optimization scheduling module.

[0032] like Figure 3 As shown, the core of this module lies in maximizing the value of data transmission under multiple physical constraints. It consists of a data cache manager, a value density calculator, a constraint monitor, an optimization solver, and a communication mode switching submodule. The data cache manager maintains a circular queue of 100 records, each containing a timestamp. , Fusion resistance value Device status words and related metadata required for calculation (such as corresponding The value density calculator implements the value model as shown below: in This represents the rolling median of the resistance values ​​at this monitoring point over the past 30 days. It is the Pearson correlation coefficient with the j-th point in the preset list of adjacent points (updated online based on a window of historical data from the past 7 days), and the weighting coefficient is set to a default value. =0.4、 =0.5、 =0.1, =0.1 (corresponding to a half-life of approximately 6.93 time units), and all parameters can be configured remotely via the cloud.

[0033] The constraint monitor monitors three hard constraints in real time: length constraint (The maximum length of the short message user information segment of the Beidou RDSS module is 120 bytes after encryption), energy constraint E (the precise remaining power is obtained by reading the coulomb counter of a battery management chip such as the TIBQ series; the typical energy consumption e of a single transmission by the Beidou module has been experimentally calibrated to be 50mAh), and frequency constraint N (a daily transmission counter is maintained in the device's non-volatile memory and compared with the daily available frequency quota of the Beidou card number). The optimization solver uses a heuristic method combining greedy algorithms and dynamic programming to solve the 0-1 knapsack problem, adapting to the computing power of embedded devices: first, the value density of all cached data is calculated, as shown in the following formula: Then press After sorting in descending order, a greedy strategy is used to initially select items until the length constraint is met. Then, dynamic programming is used to check the feasibility of local replacements to improve the total value, obtaining a near-optimal feasible solution within a finite time. The communication mode switching submodule continuously samples the RSRP value of the 4G module and sets T. low =-100dBm, T high =-90dBm, t1=30 seconds, t2=10 seconds, when RSRP <T low Switching to BeiDou mode is triggered after 30 seconds, when RSRP > T. high Switch back to 4G mode after 10 seconds.

[0034] Assuming the device is in BeiDou mode and has 5 data entries in its cache, their value Vᵢ and compressed size are given. and value density They are: D1 (Vᵢ=0.92, =45 bytes =0.0204), D2 (Vᵢ=0.45, =30 bytes =0.0150), D3 (Vᵢ=0.78, =55 bytes, =0.0142), D4 (Vᵢ=0.20, =20 bytes, =0.0100), D5 (Vᵢ=0.60, =0.0150), =0.0150, =0.0150, =0.0150, =0.0150, =0.0150, =0.0150, =0.0150, =0.0150, =0.0150, =0.0142, =0.0150 ... =35 bytes =0.0171), the current constraint is =120 bytes, E=1500mAh, e=50mAh, N=3, the implementation process is as follows: First calculate The data is sorted in descending order to obtain D1, D5, D2, D3, and D4. A greedy strategy is initially used to select D1 (45B), D5 (35B), and D2 (30B), with a total size of 110B < 120B and a total value of 0.92 + 0.60 + 0.45 = 1.97. Dynamic programming is used for optimization. An attempt is made to replace D5 with D3 to obtain the combination D1 (45B) + D3 (55B) = 100B, leaving 20B to accommodate D4. However, the total value decreases to 0.92 + 0.78 + 0.20 = 1.90. Replacing D2 would result in a combination size exceeding 120B, which is not feasible. Therefore, the original greedy solution {D1, D5, D2} is retained. The selected data is packaged, compressed, and encrypted before being transmitted via the Beidou module. The remaining energy is updated to E = 1500 - 50 = 1450mAh, and the remaining transmission attempts for the day are N = 3 - 1 = 2. This process ensures that, under strict byte capacity constraints, limited communication resources are used to transmit the most information-intensive content, prioritizing the reliable transmission of critical and abnormal data.

[0035] Example 3 This embodiment provides a specific implementation of a multi-stage risk warning engine based on failure physics.

[0036] This engine is deployed within the computing and alarm microservices of a cloud service platform. Its core function is to extract features directly related to physical failures from mixed time-series signals. It consists of a data preprocessing unit, a sequence decomposition unit, a feature extraction unit, and a rule-based decision unit. The data preprocessing unit receives standardized resistance data from a message queue (such as RabbitMQ) and maintains a circular storage area based on a time-series database (such as TDengine) for each monitoring point, retaining aggregated data at the minute / hour / day level for the most recent three years. The sequence decomposition unit uses the Seasonal Decomposition (STL) algorithm to implement R(t) = T(t) + S(t) + ... The decomposition takes a daily average resistance sequence as input, with a seasonal period of 365 days (annual cycle), a trend smoothing window of 91 days (one quarter), and a robust iteration count of 5. It effectively handles nonlinear trends and variable seasonality, and outputs a smoothed trend term T(t) (characterizing structural degradation of the grounding conductor), a periodic term S(t) (characterizing the periodic influence of environmental factors), and residuals. .

[0037] The feature extraction unit performs multi-dimensional feature calculations on the decomposed sequence: trend acceleration a t The second derivative is calculated by applying the central difference method to the T(t) sequence, as shown below: a t =(T t+1 -2T t +T t-1 ) / (Δt) 2 Δt = 1 day Take the most recent 30 days a t The moving average is used as the current trend acceleration; the deperiodized CUSUM statistic C t First, the residual sequence The equation =R(t)-S(t) is standardized as follows: in, and for The mean and standard deviation of historical data are used to calculate CUSUM, as shown in the following formula: in, The reference value is set to 0, and the tolerance k is set to 0.5. As a statistic; mutation detection statistic J t The sequential probability ratio test (SPRT) is used, and the hypothesis H0 is set (the data comes from distribution N( H1 (data from distribution N) ), Δ is the preset minimum detectable offset (e.g., 0.5Ω), calculate the cumulative log-likelihood ratio J. t And set upper and lower thresholds A and B (corresponding to target false alarm rate and false negative rate).

[0038] The rule-determination unit executes tiered early warning logic. An example of the threshold values ​​is shown below: Th a1 =0.003Ω / day 2 ,Th a2 =0.006Ω / day 2 ;Th c1 =5.0 (standardized unit); preset duration T hold =10 days; Absolute safety threshold Th abs The current setting should be dynamically adjusted according to the grounding grid design specifications (e.g., 5.0Ω).

[0039] Taking the monitoring point "GT-05" in a chemical plant area as an example, the cloud platform analyzes its daily data over the past year. The implementation process is as follows: The sequence decomposition unit calls the STL algorithm to input the daily average resistance sequence R(t) of GT-05 over the past 365 days. After decomposition, a clear upward trend term T(t), a periodic term S(t) related to annual rainfall, and residuals are obtained; the feature extraction unit calculates the trend acceleration moving average a over the past 30 days. t =0.0048Ω / day², deperiodic residual CUSUM statistic C t The SPRT module's current J value has steadily increased from 2.1 to 6.8 over the past 15 days. t No sudden changes were detected in the value within the control limits; the rule determination unit makes determinations according to hierarchical logic, because |a t |(0.0048)>Th a1 (0.003) and C t (6.8)>Th c1 (5.0) Triggers a Level 1 warning. The warning triggering module simultaneously executes multiple actions: turns the GT-05 icon yellow on the GIS map of the operation and maintenance platform, adds a trend warning indicator to the device details page, and generates a warning log: "The grounding resistance of monitoring point GT-05 shows a continuous upward trend (acceleration 0.0048Ω / day)". 2 Furthermore, a persistent positive residual exists, suggesting increased attention. This point was automatically added to the "Key Focus List," and the data monitoring frequency was increased from once per hour to once every 15 minutes. Two weeks later, a... t Increased to 0.007Ω / day 2 And C t Continued growth, satisfying "|a t |>Tha2 And C t Continued growth exceeding T hold "When the conditions are met, a level-two early warning is triggered, and the system automatically generates an inspection work order and sends an SMS notification to the responsible person. This mechanism, through decomposition and feature extraction, identifies the slow deterioration trend caused by grounding corrosion about 40 days in advance, even before the absolute resistance value exceeds the standard (only 2.8Ω). This realizes the transformation from passive threshold alarm to proactive trend prediction, providing sufficient decision-making time for planned maintenance."

[0040] The above embodiments detail the specific implementation methods, parameter examples, and workflows of each technical feature, fully demonstrating the feasibility of the present invention. Those skilled in the art can implement the present invention without inventive effort based on the disclosed content.

[0041] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0042] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An online monitoring system for changes in grounding grid resistance, characterized in that: This includes on-site monitoring devices, communication networks, and cloud service platforms; The field monitoring device is used to collect and process grounding resistance data. It integrates a dynamic uncertainty assessment and fusion module as well as a communication and energy joint optimization scheduling module. A communication network is used to provide a data transmission channel between field monitoring devices and cloud service platforms; The cloud service platform is used to receive and store data and perform risk analysis and early warning. It has a multi-stage risk warning engine based on failure physics deployed inside. in, The dynamic uncertainty assessment and fusion module is configured to simultaneously perform measurements using the potential drop method and the electromagnetic induction method. Based on the measurement uncertainties of the two methods assessed in real time, it uses an inverse variance weighted model to perform dynamic data fusion and outputs the optimal estimated resistance value. The communication and energy joint optimization scheduling module is configured to, in the BeiDou short message communication mode, decide the content of the data to be sent by solving an optimization problem based on the constructed data value density model and under the constraints of message length, number of transmissions and remaining energy. The multi-stage risk warning engine based on failure physics is configured to perform hybrid model decomposition on grounding resistance time series data, extract trend acceleration, deperiodic residual cumulative sum and mutation detection statistics, and trigger graded warnings based on the logical combination of the above feature parameters.

2. The online monitoring system for changes in grounding resistance value of a grounding grid according to claim 1, characterized in that: The dynamic uncertainty assessment and fusion module includes: The uncertainty assessment submodule is configured as follows: Calculate the spacing distortion factor using the potential drop method: in The actual wiring length is obtained through the environmental sensing unit. To satisfy the theoretical minimum length of the semi-infinite field assumption, and based on the pre-defined mapping relationship f v Obtain the estimated standard deviation of the measurement error using the potential drop method. ; Calculate the signal-to-interference-plus-noise ratio (SIR) of the electromagnetic induction method: in The signal current of the circuit under test, Interference current from adjacent circuits, The background noise equivalent current is determined according to a preset mapping relationship. f c Obtain the estimated standard deviation of the measurement error using the electromagnetic induction method. ; The data fusion submodule is configured to calculate dynamic weights and perform fusion using the following formula: in, R v and R c These are the measured values ​​using the potential drop method and the electromagnetic induction method, respectively. R fused This represents the optimal estimated resistance value after fusion.

3. The online monitoring system for changes in grounding resistance value of a grounding grid according to claim 2, characterized in that: The on-site monitoring device also includes an environmental sensing unit, which includes at least a tilt sensor and an odometer for real-time estimation of the actual wiring length. d aCtual Electromagnetic induction measurement employs a dual-coil differential clamp structure.

4. The online monitoring system for changes in grounding resistance value of a grounding grid according to claim 1, characterized in that: The joint optimization scheduling module for communication and energy includes: The value calculation submodule is configured to calculate the information value density for each piece of data i to be transmitted in the cache queue. V i The calculation formula is: in, For data delay time, This is the resistance measurement value. This is the historical baseline value. The historical correlation coefficient with the data of the j-th geographically or electrically adjacent monitoring point. , These are configurable weighting coefficients. It is the attenuation constant; The optimization decision submodule is configured to solve the following 0-1 knapsack problem to determine the sending set each time a wake-up call is sent: Objective function: ; Constraints: C1： ； C2: e≤E; C3: N>0; C4:x i ∈{0,1}; in, Let i be the number of bytes after compression. Here, e represents the maximum capacity of the BeiDou short message service, e represents the energy consumption of the BeiDou module per transmission, E represents the current remaining energy of the device, N represents the remaining number of transmissions available for the day, and x represents the maximum capacity of the BeiDou short message service. i These are decision variables.

5. The online monitoring system for changes in grounding resistance value of a grounding grid according to claim 4, characterized in that: The joint optimization scheduling module for communication and energy also includes a communication mode switching submodule, configured to: monitor the reference signal received power (RSRP) of the 4G network in real time; when RSRP remains below a first threshold T low When the preset duration t1 is reached, the control field monitoring device switches to BeiDou short message communication mode and activates the optimization decision submodule; when RSRP recovers to a level higher than the second threshold T... high When the preset duration t2 is reached, the control field monitoring device switches back to 4G communication mode, where T... high >T low .

6. The online monitoring system for changes in grounding resistance value of a grounding grid according to claim 1, characterized in that: A multi-stage risk warning engine based on failure physics includes: The sequence decomposition submodule is configured to decompose the resistance time series data R(t) into a trend term T(t), a periodic term S(t), and a residual term. Satisfying: R(t) = T(t) + S(t) + The trend term T(t) is used to characterize the structural deterioration of the grounding electrode, and the periodic term S(t) is used to characterize the periodic influence of environmental factors. The feature extraction submodule is configured as follows: Calculate the second discrete derivative of the trend term T(t) as the trend acceleration a. t ; Calculate the cumulative sum CUSUM statistic C for the deperiodic residual sequence R(t)−S(t). t ; The sequential probability ratio test (SPRT) was performed on the original resistance time series data R(t) to obtain the mutation detection statistic J. t .

7. The online monitoring system for changes in grounding resistance value of a grounding grid according to claim 6, characterized in that: The multi-stage risk warning engine based on failure physics also includes a warning triggering submodule, which is configured to execute hierarchical warning logic: When the condition is met: |a t ∣>Th a1 Or C t >Th c1 At that time, a Level 1 warning was triggered; When the condition is met: |a t ∣>Th a2 And C t When the growth continues for more than a preset duration, a level-two warning is triggered, in which Th a2 >Th a1 ; When condition: J t Exceeding the preset mutation threshold or R(t)>Th abs At that time, a level 3 alarm was triggered.

8. The online monitoring system for changes in grounding resistance value of a grounding grid according to claim 1, characterized in that: The field monitoring device has a built-in power supply unit and a main control unit. The main control unit supports low-power sleep and timed wake-up modes and works in conjunction with the communication and energy joint optimization scheduling module to ensure communication optimization scheduling under energy constraints.

9. The online monitoring system for changes in grounding resistance value of a grounding grid according to claim 1, characterized in that: The cloud service platform adopts a microservice architecture, which includes at least: The device access gateway group is used to concurrently receive and parse data from different communication protocols, and to standardize the data based on a predefined object model; Message middleware is used to asynchronously decouple data streams and distribute standardized data to downstream services. The computing and alerting service, as a containerized microservice, runs a multi-stage risk warning engine based on failure physics. The hierarchical data storage cluster includes Redis for caching real-time data, MySQL for storing business relationship data, and TDengine for storing historical time-series data.

10. The online monitoring system for changes in grounding resistance value of a grounding grid according to claim 1, characterized in that: The cloud service platform specifically includes: The device access gateway group is used to concurrently receive and parse data from different communication protocols, and to standardize the data based on a predefined object model; The message middleware connects to the device access gateway group and is used for asynchronous distribution of standardized data; The computing and alerting service, connected to the message middleware, is a containerized microservice that runs a multi-stage risk warning engine based on failure physics. A tiered data storage cluster, connected to computing and alarm services, includes a cache database for storing real-time data, a relational database for storing business relationship data, and a time-series database for storing historical time-series data.

Images