Multiple active safety protection method for ammeter and fusion terminal system

Abstract

The invention discloses a multiple active safety protection method for an electric meter and a fusion terminal system, which relates to the field of electric energy metering protection, and comprises the following steps: setting a sampling frequency, collecting multi-dimensional electrical parameters and environmental parameters in real time when electric energy metering and collecting equipment operates, filtering the collected original parameter data, and determining the safety of the electric meter and the fusion terminal system. And carrying out life modeling calculation, analyzing characteristic parameters of the components in the filtered data, and calculating health state values of the components. According to the method, a capacitor ESR and relay operation frequency-based fusion degradation model and a built-in element degradation model are adopted, passive protection is changed into active early warning, and through life modeling and fault feature extraction, traditional postmortem treatment is converted into beforehand early warning, so that effective prediction of the residual life of the component is facilitated, the aging risk is identified in advance, and the reliability of the component is improved. Risk conditions such as electrical fire and metering inaccuracy can be effectively prevented, and the operation safety of equipment is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electric energy metering protection, in particular to a kind of multiple active security protection method of electric meter and fusion terminal system. BACKGROUND

[0002] Electric energy metering and acquisition equipment is usually the core terminal for realizing electric energy metering, data acquisition and remote communication in power system, and its running stability and safety directly affect the reliability of power supply, user power safety and power company operation efficiency.

[0003] At present, the functions of traditional electric energy metering and acquisition equipment are mainly concentrated on basic metering and communication, but the key components such as capacitors, chips and relays inside traditional equipment usually will occur irreversible aging with the extension of use time, resulting in the increase of equivalent series resistance of capacitor, the attenuation of chip performance or the wear of relay contact. Since the existing technology lacks effective prediction means for the remaining life of components, it cannot identify aging risk in advance, and can only be handled passively after failure occurs, thereby causing inaccurate metering during metering and acquisition, and more likely causing electrical fire due to overheating of components.

[0004] Therefore, it is necessary to use a safety protection method with active early warning, multiple protection and accurate life management to identify aging risk in advance. SUMMARY

[0005] The purpose of the present application is to provide a multiple active security protection method of electric meter and fusion terminal system to solve the problems raised in the background art.

[0006] To achieve the above purpose, the present application provides the following technical scheme: a multiple active security protection method of electric meter and fusion terminal system, comprising the following specific steps: Step one: data high-frequency acquisition, set sampling frequency, and collect multi-dimensional electric parameters and environmental parameters in real time when electric energy metering and acquisition equipment is running, and store the collected parameter data into buffer area; Step two: signal preprocessing, filtering the original parameter data collected in step one to obtain filtered data; Step three: life modeling calculation, analyze the component characteristic parameters in the filtered data in step two, calculate the health state value of the components, and if the health state value is lower than the preset health threshold, trigger life warning; Step four: fault feature extraction, time-frequency analysis and local frequency spectrum feature extraction are performed on the current signal in the filtered data to detect whether the energy of specific frequency band exceeds the preset arc fault threshold to identify arc fault features; Step five: dynamic threshold adjustment, according to the environmental parameters collected in step one, the protection threshold value suitable for the current environment is calculated in real time through the threshold adjustment model; Step six: risk research decision, using a multi-feature weighted decision mechanism, the health state value in step three, the fault feature detection result in step four and the dynamic protection threshold value in step five are fused and analyzed to determine whether to trigger a warning or protection action; Step seven: protection action execution, when it is determined to trigger a protection action in step six, the relay element is driven to cut off the circuit, and the fault event information is recorded, when it is determined to trigger a warning, the warning information is generated; Step eight: data reporting communication, the health state information in step three, the event information and the warning information in step seven are compressed and encoded, and reported to the operation and maintenance platform through the communication module, and the communication module supports dual-channel dynamic switching.

[0007] Preferably, the sampling frequency in step one is set to 4kHz, the real-time collection in step one is realized by an ADC collection unit, the multi-dimensional electrical parameters in step one include voltage, current and leakage current, and the environmental parameters include temperature and humidity.

[0008] Preferably, the sampling frequency in step one is set based on the Shannon sampling theorem, and the formula of the Shannon sampling theorem is as follows:

[0009] Among them, is the actual sampling frequency, is the highest frequency component in the signal.

[0010] Preferably, the original parameter data in step two is processed by a finite-length unit impulse response filter, and the formula of the finite-length unit impulse response filter is as follows:

[0011] Among them, is the output value of the first point after filtering, is the original input signal, where is the current point, is the historical point index, is the coefficient of the filter, is the order of the filter, and the filter output is obtained by convolution of the current and historical input signals and the filter coefficients .

[0012] ​Preferably, the component characteristic parameters in step three include the equivalent series resistance of the capacitor and the relay operating frequency. The lifetime modeling calculation in step three is based on the component degradation law, which uses an exponential decay model to calculate the change of the equivalent series resistance of the capacitor over time. The formula for the exponential decay model is as follows:

[0013] in, Refers to time The measured value of the equivalent series resistance at time t. This refers to the initial condition of the capacitor when it leaves the factory. value, This refers to the attenuation coefficient based on capacitor materials and processes, obtained by calculating the ratio of ripple voltage to current in real time. Value, combined with factory initial value and material attenuation coefficient Calculate the health status of the capacitor.

[0014] Preferably, the time-frequency analysis in step four is performed using a short-time Fourier transform, and the local spectral features are extracted using a windowed segmentation method; the fault feature signals in step four include fault arc signals and overcurrent signals, and the formula for the short-time Fourier transform is as follows:

[0015] in, It refers to time and frequency The short-time Fourier transform result at the point, This refers to the input current signal. It refers to the time variable. It refers to a window function used to capture signals. In time Nearby local segments, It refers to the complex exponential function, used in Fourier transform to convert a signal from the time domain to the frequency domain. Refers to the integral variable The differential.

[0016] Preferably, the protection threshold in step five is a leakage current protection threshold. The protection threshold adapted to the current environment in step five is calculated in real time using a linear interpolation model, combined with threshold settings for dry and humid environments and a humidity sensitivity coefficient. The formula for the linear interpolation model is as follows:

[0017] in, This refers to the current humidity. computed dynamic leakage protection threshold, refers to the standard threshold in dry environment, refers to the high sensitivity threshold in humid environment, refers to the preset humidity critical point.

[0018] Preferably, the fusion analysis in step six is analyzed by a multi-feature weighted decision mechanism, which includes normalizing arc fault features, current rate of change and leakage current features, and giving each feature a preset weight; When the comprehensive risk score is greater than or equal to the trip threshold, trigger the protection action; When the early warning threshold is less than or equal to the comprehensive risk score and less than the trip threshold, trigger the sending of early warning; The formula of the multi-feature weighted decision mechanism is as follows:

[0019] Among them, refers to the action taken according to the comprehensive risk score, which is the output result of the decision mechanism, refers to the comprehensive risk score, reflecting the current risk level of the electric energy metering and collecting equipment, refers to the trip threshold, refers to the early warning threshold.

[0020] Preferably, the fault event information in step seven includes fault type, fault occurrence timestamp, multi-dimensional electric parameter and environmental parameter data at the time of fault, and the protection action execution process in step seven follows a relay driving timing model, and the formula of the relay driving timing model is as follows:

[0021] Among them, refers to the trip signal, refers to the comprehensive risk score, refers to the trip threshold, refers to the fault current, refers to the hardware instantaneous action value; When the comprehensive risk score exceeds the trip threshold or the fault current exceeds the hardware instantaneous action value , output high level 1 to drive the relay action.

[0022] Preferably, the compression encoding in step eight adopts a difference encoding mode, only transmitting necessary state information including health degree, event type and timestamp or difference data, the communication module in step eight includes high-speed power line carrier and high-frequency wireless, and the formula included in step eight is as follows:

[0023] wherein, is the value at the current moment, is the current data point for difference calculation, is the value at the previous moment, is the comparison data point for difference with the current data point; Difference encoding only sends the difference between the current value and the previous value If the change is small, it is not sent.

[0024] Technical effects and advantages of the present application: (1) The present application realizes a life prediction accuracy of >95% through a fusion degradation model based on the capacitance ESR and the relay action frequency, an internal element degradation model, supports early warning by 720 hours, reduces operation and maintenance costs by 30%, uses passive protection as active early warning through life modeling and fault feature extraction, changes traditional post-disposal into pre-warning, and is thus beneficial to effectively predicting the remaining life of components, identifying aging risks in advance, effectively preventing electrical fires and measurement inaccuracies, and improving equipment operation safety; (2) The present application adopts current slope and high-frequency harmonic joint analysis, identifies premonitory signs within 500ms before overcurrent and arc fault occurs and cuts off, is beneficial to improving response speed, integrates current change rate and high-frequency feature recognition for advanced safety protection, dynamically adjusts the leakage protection threshold through integrated environmental humidity sensing and insulation resistance monitoring, and thus builds a multi-depth protection system integrating device life protection, electrical fire protection and adaptive personal protection, covers equipment, power grid and personal safety needs in all dimensions, and solves the single defect of traditional device protection; (3) The present application improves response efficiency through filtering, feature extraction, decision-making and edge intelligence, all data processing is completed locally on the device, the response time is millisecond level, much faster than cloud processing, improves protection reliability, reduces cloud computing load, and reduces cloud construction cost; (4) The present application uses dual-channel switching communication, HPLC and HRF dual-channel dynamic switching, and is thus beneficial to ensuring that the power outage event reporting delay is <2s, improving communication reliability, ensuring uninterrupted data reporting, and adapting to the setting mode of complex power grid environment; (5) The application is adapted to the power grid demand with high efficiency and low consumption, reduces the communication flow through difference coding, and is beneficial to optimizing the static power consumption of equipment, and meets the development trend of energy saving and high efficiency of the power system. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 A flowchart of the multiple active safety protection method of the application. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the application.

[0027] The application provides a multiple active safety protection method for an electric meter and a fusion terminal system, as shown in the formula (I): Figure 1 The multiple active safety protection method comprises the following specific steps: Step 1: High-frequency data acquisition, set the sampling frequency, and collect the multi-dimensional electric parameters and environmental parameters of the electric energy metering and acquisition equipment in real time during operation, and store the collected parameter data into the buffer area, and the algorithm is realized as follows: Sampling frequency setting logic: Define a constant parameter sampling frequency, and the value is 4000. The parameter represents that 4000 times of collection will be performed on the electric signal per second, that is, the sampling frequency is 4 kHz. This sampling frequency is used as the time reference of the entire data acquisition system. ADC interrupt handling program logic: When the ADC (analog-to-digital converter) completes a conversion and triggers an interrupt, the following operations are performed: Read the voltage data: Read the value from the data register of the ADC module numbered 1, and store the read value in the variable named global voltage sampling value. Read the current data: Read the value from the data register of the ADC module numbered 2, and store the read value in the variable named global current sampling value. Data buffering operation: Store the set of sampling data stored in the global voltage sampling value and the global current sampling value in the pre-allocated buffer area according to the collection sequence, so as to ensure that these original sampling data are saved for subsequent signal processing. Interrupt preparation: Prepare for the interrupt request after the next ADC conversion is completed, so that the sampling process can continue at a frequency of 4 kHz. Step two: signal preprocessing, filtering the original parameter data collected in step one to obtain filtered data, which is conducive to continuously updating the buffer and participating in convolution calculation, ensuring filtering effect, reducing calculation complexity, ensuring real-time performance, and the algorithm is implemented as follows: Filter order setting logic: Define a fixed parameter called filter order, whose value is 5, representing that the FIR (Finite Impulse Response) filter is a 5th order filter, i.e. the latest 5 input data and corresponding filter coefficients need to be called for operation in the filtering calculation process.

[0028] FIR filter function execution logic: This function is used to receive a single raw input data and output clean data after processing by the FIR filtering algorithm. The function input is the current raw input data and the data buffer, and the output is the filtered data. The specific steps are as follows: Initialize index variable Define a static index variable and initialize it to 0. This variable is used to record the storage location of the current input data in the buffer, and because of the static attribute, the index value will not be reset when the function is called multiple times, and the buffer storage location can be continuously tracked.

[0029] Store the current input data Write the current raw input data received by this function into the data buffer at the position pointed to by the index variable, completing the caching of the current data.

[0030] Update buffer index Calculate the new index value: new index = (current index + 1) divided by filter order and take the remainder.

[0031] When the index does not reach N-1 (e.g. N=5, index from 0 to 3), the new index is current index + 1, which realizes the movement of the buffer position to the rear one by one. When the index reaches N-1 (e.g. N=5, index is 4), the new index is reset to 0, which realizes the circular storage of the buffer and ensures that the latest N input data is always stored.

[0032] Calculate the filtered output data Initialize the filtered data to 0, and perform the following operations through a loop (from i=0 to i=N-1): In each loop, multiply the historical data at the i-th position of the data buffer by the i-th coefficient of the preset FIR filter coefficient to obtain a single product result. Add all single product results to the filtered data to obtain the output value of this filtering (i.e. complete the convolution operation of input data and filter coefficients).

[0033] The return filtering result outputs the calculated filtered data as a function for subsequent signal analysis or processing modules to use, thereby achieving suppression of noise and interference in the original data. Step three: life modeling calculation, analyze the characteristic parameters of the components in the filtered data in step two, calculate the health state value of the components, if the health state value is lower than the preset health threshold, trigger the life warning, and then build a life prediction model based on the aging law of the core components of the equipment (capacitors and relays), realize the transition from replacement after failure to early warning, the life of the relay is positively related to the action frequency (the more the action times, the more serious the contact wear), by recording the cumulative action times of the relay, combining with the rated action times calibrated at the factory, calculating the health state value of the relay, when the health state value is lower than the threshold, it is included in the life warning at the same time, the algorithm is as follows: Capacitor equivalent series resistance calculation logic: This part of logic is used to calculate the equivalent series resistance (ESR) of the capacitor through voltage and current data, the core is to use the ratio of capacitor ripple voltage and ripple current to derive ESR, the specific steps are as follows: Function initialization and input definition: Define a function named calculate ESR, which receives two input parameters: voltage data (i.e. voltage signal across the capacitor) and current data (i.e. current signal flowing through the capacitor), the function output result is the calculated ESR value (data type is float).

[0034] Extract ripple component: Calculate ripple voltage: call the extract ripple function, pass the input voltage data into the function, separate the ripple component in the voltage signal (i.e. the alternating fluctuation part in the voltage signal except for the direct current or fundamental wave) through the function, get the ripple voltage.

[0035] Calculate ripple current: also call the extract ripple function, pass the input current data into the function, separate the ripple component in the current signal, get the ripple current (float type).

[0036] Calculate and return ESR value: According to the physical definition of ESR (the equivalent series resistance of the capacitor is equal to the ratio of its ripple voltage and ripple current), divide the ripple voltage by the ripple current to get the equivalent series resistance (ESR) value of the current capacitor, return it as the result of the calculate ESR function.

[0037] Capacitor health degree calculation and warning judgment logic: This part of logic is used to evaluate the health state of the capacitor based on the calculated ESR value, and trigger the warning when the health degree is too low, the specific steps are as follows: Calculate health score: Define a floating-point variable named health score to quantify the health status of the capacitor.

[0038] Calculation logic: Take the initial ESR value of the capacitor when it leaves the factory (the standard ESR value of a new capacitor, pre-stored in the system) as the reference, divide it by the current calculated ESR value, and multiply it by 100 to get the health score expressed in percentage. For example: If the initial ESR is 10Ω and the current ESR is 12Ω, the health score is 100×(10 / 12)≈83.3 points. The principle is that during the aging process of the capacitor, ESR will increase with the degree of aging, so the smaller the ratio of initial ESR / current ESR, the more serious the capacitor aging, and the lower the health score.

[0039] Health threshold judgment and early warning triggering: Set health threshold: Pre-store a health threshold in the system (for example, 60 points, set according to capacitor model and application scenario, below which the capacitor aging risk is too high).

[0040] Threshold comparison: Compare the calculated health score with the health threshold.

[0041] Trigger warning: If the health score is less than the health threshold, call the send warning function and pass in the warning information that the capacitor health is low. This warning information will be reported to the operation and maintenance platform or pushed to the user end through the system communication module, prompting the staff to check or replace the capacitor in time; Step four: Fault feature extraction, time-frequency analysis and local frequency spectrum feature extraction are performed on the current signal in the filtered data to detect whether the energy of a specific frequency band exceeds the preset arc fault threshold, so as to identify the arc fault feature, which is conducive to realizing fault identification in advance. The algorithm is implemented as follows: FFT initialization logic: Define a fast real Fourier transform (RFFT) instance named "fft" to store the configuration information required for FFT calculation; Call the fast RFFT initialization function to initialize the configuration of the above "fft" instance. The configuration parameters include the length of FFT (specified by FFT_LENGTH), and create a floating-point array fft_output with a length of FFT_LENGTH to store the result data after FFT calculation; FFT calculation and fault judgment logic: Call the fast RFFT calculation function to perform Fourier transform on the current signal data in the current buffer (current_buffer). The calculation result is stored in the fft_output array (the last parameter 0 represents performing forward FFT transform). Check the energy value of the frequency band corresponding to the 50th position in the transformed result (this position corresponds to a specific frequency range). The energy value is compared with a preset arc fault threshold value, and if the frequency band energy value exceeds the arc fault threshold value, a protection trigger function is called to start the corresponding protection mechanism; Step five: dynamic threshold adjustment, according to the environmental parameters collected in step one, the protection threshold value suitable for the current environment is calculated in real time through the threshold adjustment model, which is beneficial to dynamically adjust the leakage protection threshold value according to the environmental humidity, improve the adaptability of personal safety protection, and the algorithm is realized as follows: Basic definition and function declaration logic: Declare a function named read humidity sensor, which has no input parameters and returns a 16-bit unsigned integer, used to obtain the humidity data (such as percentage RH value) of the current environment from the humidity sensor.

[0042] Define the reference leakage threshold value as 30.0mA, which is the leakage protection reference value in the default environment (such as 50% RH humidity), to ensure basic personal safety protection.

[0043] Define the humidity sensitivity coefficient as 0.1, which represents the adjustment amplitude of the leakage protection threshold value when the humidity changes by 1% RH, used to quantify the influence of environmental humidity on the protection threshold value.

[0044] Dynamic threshold calculation and setting logic: Get real-time humidity: call the read humidity sensor function to get the real-time humidity data of the current environment and store it in the humidity variable (such as detecting that the current humidity is 60% RH).

[0045] Calculate dynamic threshold: take 50% RH as the humidity reference point, and calculate the dynamic leakage protection threshold value suitable for the current humidity through the formula: dynamic threshold = reference leakage threshold value - humidity sensitivity coefficient × (current humidity - 50.0) Example: if the current humidity is 60% RH, substitute it into the formula: dynamic threshold = 30.0 - 0.1 × (60.0 - 50.0) = 29.0mA, which realizes the threshold value reduction (improves protection sensitivity) when the humidity rises.

[0046] Set threshold to take effect: call the set leakage threshold value function to pass the calculated dynamic threshold value into the function, update the trigger threshold value of the leakage protection comparator in the system, make the protection mechanism adapt to the current environmental humidity, and avoid false action or protection failure; Step six: risk judgment and decision, adopt a multi-feature weighted decision mechanism to fuse and analyze the health state value in step three, the fault feature detection result in step four, and the dynamic protection threshold value in step five, to judge whether to trigger an early warning or protection action, which is beneficial to fuse multi-dimensional information of life, fault and environment, avoid single feature misjudgment, and ensure decision accuracy, and the algorithm formula is as follows: Emergency fault judgment and trip logic: The following three types of emergency risk conditions are detected or determined (any one of them triggers emergency action): If the arc fault detection result is true (i.e. the system has identified a fault arc); Or the current rate of change is greater than the critical value of the current rate of change (i.e. the current rises sharply, there is an overcurrent / short circuit risk); Or the leakage current is greater than the dynamic leakage protection threshold (i.e. the leakage current is excessive, there is a risk of electric shock); If any of the above conditions are met, the action command is set to trip immediately, driving the relay to cut off the circuit to isolate the fault.

[0047] Potential risk determination and early warning logic: If none of the above emergency fault conditions are met, further determine the potential risk: if the insulation resistance is less than the lower limit of the insulation resistance (i.e. the insulation performance of the device is deteriorating, there is a potential risk of future leakage / short circuit), the action command is set to send a warning, generate a warning information and report to the operation and maintenance platform, and prompt early troubleshooting and maintenance.

[0048] Default logic for non-triggered conditions: If none of the above emergency fault conditions and potential risk conditions are met, the current action command is not changed (default to continue monitoring state, no trip or warning action); Step seven: protection action execution, when the protection action is triggered in step six, drive the relay element to cut off the circuit, and record the fault event information, when the warning is triggered, generate a warning information, which is conducive to realize the fault rapid cut-off and event recording, ensure the protection reliability, the algorithm is implemented as follows: Trip function execution logic: Define a function called trigger trip, which performs the following operations: Control the relay action: Access the GPIO port register where the relay control pin is located, set the BSRR register of this port, and set the relay control pin to the off state, which will physically cut off the circuit and achieve circuit protection; Record the fault event: Call the record event function and pass in the fault event parameter, which will store the relevant information of this trip event (such as time, fault type, etc.) into the system log for subsequent analysis and tracing; Step 8: Data Reporting Communication. The health status information from Step 3, the event information from Step 7, and the early warning information are compressed and encoded, and reported to the operation and maintenance platform through the communication module. The communication module supports dynamic switching between dual channels, which facilitates efficient and reliable data reporting, reduces communication costs and power consumption. Under normal operating conditions, high-speed power line carrier HPLC (high transmission rate and good stability) is used. When the high-speed power line carrier HPLC communication is interrupted (such as by line fault), it automatically switches to high-frequency wireless HRF to ensure communication reliability ≥99.99%. For emergency information such as power outages, the reporting delay is ≤2 seconds to ensure timely response by operation and maintenance personnel. The algorithm implementation is as follows: Reporting data packet structure definition logic: Define a data structure named Reporting Data Packet to integrate the key information that needs to be reported; This data packet contains three fields: Health score: Floating-point value, representing the health status of storage device components; Event type: An 8-bit unsigned integer that identifies the type of event being reported (such as fault, warning, etc.). Timestamp: A 32-bit unsigned integer that records the time information of an event. Enable compression and packing attributes on data packets to remove padding bytes from data structures and minimize data packet size; Data sending logic: Call the function to send via high-speed power line carrier. The first parameter of the function is the address of the report data packet constructed above, which specifies the data to be sent. The second parameter of the function is the size of the data packet, which is used to inform the receiver of the data length; After the function is executed, the data packet will be sent to the target receiver via high-speed power line carrier communication.

[0049] Furthermore, the sampling frequency in step one is set to 4kHz, and the real-time acquisition in step one is realized through the ADC acquisition unit. The multi-dimensional electrical parameters in step one include voltage, current and leakage current. Voltage reflects the stability of the power grid voltage, current reflects the monitoring of overcurrent and short circuit trends, leakage current reflects the identification of leakage risk, and environmental parameters include temperature and humidity. Temperature monitors the overheating of the monitoring element, and humidity affects the risk of electric shock and insulation performance. The acquired parameters are stored in the buffer in real time through the ADC interrupt service routine to avoid data loss and ensure the continuity of subsequent processing.

[0050] The sampling frequency set in step one is based on Shannon's sampling theorem, and the formula for Shannon's sampling theorem is as follows: ; in, This refers to the actual sampling frequency. This refers to the highest frequency component in the signal. The actual sampling frequency is more than twice that of the highest frequency component in the signal to ensure that the signal is recovered without distortion, which facilitates the accuracy of subsequent data analysis.

[0051] Furthermore, the raw parameter data in step two is filtered using a finite-length unit impulse response filter. The formula for the finite-length unit impulse response filter is as follows: ; in, This refers to the filtered first... The output value of each point, It refers to the original input signal, where For the current point, This refers to the historical point index. It refers to the coefficients of the filter. This refers to the order of the filter, the filter output. From current and historical input signals With filter coefficients Convolution is advantageous because it can be implemented through a circular buffer, which is computationally efficient and easy to implement. This allows for the filtering of the original parameter data through convolution operations between the current and historical input signals and the filter coefficients, removing noise interference, improving data quality, and providing more reliable basic data for subsequent analysis.

[0052] Specifically, the component characteristic parameters in step three include the equivalent series resistance of the capacitor and the relay operating frequency. The lifetime modeling calculation in step three is based on the component degradation law. The component degradation law uses an exponential decay model to calculate the change of the equivalent series resistance of the capacitor over time. The formula for the exponential decay model is as follows: ; in, Refers to time The measured value of the equivalent series resistance at time t. This refers to the initial condition of the capacitor when it leaves the factory. value, This refers to the attenuation coefficient based on capacitor materials and processes, obtained by calculating the ratio of ripple voltage to current in real time. Value, combined with factory initial value and material attenuation coefficient Calculate the health status of the capacitor.

[0053] Specifically, the time-frequency analysis in step four is performed using short-time Fourier transform. Local spectral features are extracted through windowing and segmentation. A movable window function (such as a Hanning window) is used to extract local segments of the current signal, and a Fourier transform is performed on each segment to obtain the spectral distribution at different time points, thereby capturing the transient characteristics of the fault. Fault arc identification is based on the fact that a fault arc will generate a specific high-frequency spectrum (such as 10kHz-1MHz). By detecting whether the energy in this frequency band exceeds a preset threshold, it is determined whether an arc fault exists. Overcurrent trend identification is based on the analysis of the current change rate. If the current change rate exceeds a critical value, an overcurrent trend is determined. The entire fault feature extraction process has a response time of ≤500ms, which is much faster than traditional mechanical protection, achieving early warning. The fault feature signals in step four include fault arc signals and overcurrent signals. The formula for short-time Fourier transform is as follows: ; in, It refers to time and frequency The short-time Fourier transform result is the core output of time-frequency analysis, used to observe how the frequency components of a signal change over time. This refers to the input current signal. This refers to the time variable, representing the value of a signal at different times, and is the object of time-frequency analysis. It refers to a window function used to capture signals. In time Nearby local segments, It refers to the complex exponential function, used in Fourier transform to convert a signal from the time domain to the frequency domain. It is the imaginary unit. It is a frequency variable; through this exponential term, the signal in the time domain is... Transform to the frequency domain to obtain the signal in the frequency domain. The components at the site, Refers to the integral variable The derivative is used to perform integration over the entire time domain to obtain the result of the short-time Fourier transform.

[0054] Furthermore, the protection threshold in step five is the leakage current protection threshold. The protection threshold adapted to the current environment in step five is calculated in real time using a linear interpolation model, combined with the threshold settings for dry and humid environments and the humidity sensitivity coefficient. The formula for the linear interpolation model is as follows: ; in, This refers to the current humidity. The calculated dynamic leakage current protection threshold, This refers to the standard threshold under dry conditions. This refers to the high sensitivity threshold in humid environments. This refers to a preset humidity threshold, which facilitates real-time calculation of the protection threshold applicable to the current humidity level, thereby enhancing the system's adaptability to the environment.

[0055] Furthermore, the fusion analysis in step six is ​​conducted through a multi-feature weighted decision mechanism, which includes normalizing the characteristics of arc faults, current change rate, and leakage current, and assigning preset weights to each feature. When the comprehensive risk score is greater than or equal to the tripping threshold, the protection action is triggered; When the warning threshold is less than or equal to the comprehensive risk score and less than the tripping threshold, a warning is triggered. The formula for the multi-feature weighted decision-making mechanism is as follows: ; in, This refers to the actions taken based on the comprehensive risk score; it is the output of this decision-making mechanism. This refers to a comprehensive risk score, which reflects the current risk level of electricity metering and data acquisition equipment. This refers to the tripping threshold. This refers to the early warning threshold, which is determined by the comprehensive risk score. Greater than When the overall risk score is high, it means the equipment is at extremely high risk and a tripping action needs to be triggered to disconnect the circuit and ensure safety. Greater than the threshold and less than or equal to This indicates that the equipment poses a certain risk and a warning needs to be issued. This prompts relevant personnel to pay attention to the equipment status, and the overall risk score is... Less than or equal to When this indicates that the equipment risk is within an acceptable range, no action is taken. Special actions.

[0056] Specifically, the fault event information in step seven includes the fault type, fault occurrence timestamp, and multi-dimensional electrical and environmental parameter data at the time of the fault. The protection action execution process in step seven follows the relay drive timing model, and the formula for the relay drive timing model is as follows: ; in, This refers to the trip signal, a control signal used to drive the relay to operate. This refers to a comprehensive risk score, calculated by weighting multiple characteristics such as arc fault features, current change rate, and leakage current characteristics. It reflects the overall risk level of the equipment. This refers to the tripping threshold. When the comprehensive risk score exceeds this threshold, the equipment risk is considered to have reached a level requiring tripping. This refers to fault current, the current value when a fault occurs during equipment operation. It refers to the instantaneous action value of the hardware, which is the current threshold set at the hardware level. When the fault current exceeds this value, the hardware will directly trigger the protection action. When the comprehensive risk score Exceeding the trip threshold or fault current Exceeding the instantaneous motion value of the hardware When the output is high (1), it drives the relay to operate. Specifically, the compression encoding in step eight uses differential encoding, transmitting only necessary status information, including health status, event type, and timestamp or differential data. The communication module in step eight has dual channels, including high-speed power line carrier and high-frequency radio. The formulas included in step eight are as follows: ; in, This refers to the value at the current moment, which is the current data point used for difference calculation. This refers to the value from the previous moment, which is the data point compared to the current data point by subtracting it. Difference encoding only sends the current value. Compared with the previous value If the difference is small, it will not be sent, which helps to reduce the amount of data transmitted, improve communication efficiency, and ensure the effective reporting of key information.

[0057] A multi-active safety protection system for electricity meters and integrated terminal systems includes: The parameter acquisition module is used to set the sampling frequency and collect multi-dimensional electrical parameters and environmental parameters in real time during the operation of the power metering and acquisition equipment, and store the collected parameter data in a buffer. The signal filtering preprocessing module is used for signal preprocessing, which filters the acquired raw parameter data to obtain filtered data. The status monitoring and early warning module is used for life modeling calculation, analyzing the component characteristic parameters in the filtered data, and calculating the health status value of the components. The fault feature identification module is used to extract fault features, perform time-frequency analysis and local spectral feature extraction on the current signal in the filtered data, and detect whether the energy of a specific frequency band exceeds a preset arc fault threshold in order to identify arc fault features. An adaptive protection threshold calculation module is used to perform dynamic threshold adjustment. Based on the collected environmental parameters, the module calculates the protection threshold that is adapted to the current environment in real time through a threshold adjustment model. The decision module is used to perform a multi-feature weighted decision mechanism to fuse and analyze health status values, fault feature detection results and dynamic protection thresholds to determine whether to trigger an early warning or protection action. The protection module is used to drive the relay element to cut off the circuit and record the fault event information when it is determined that a protection action is triggered; and to generate a warning message when it is determined that a warning is triggered. The data reporting module is used to compress and encode health status information, event information, and early warning information, and report them to the operation and maintenance platform through dynamic switching of dual communication channels.

[0058] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is 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. A method for multiple active security protections of an electricity meter and a converged terminal system, characterized in that: The specific steps include the following: Step 1: High-frequency data acquisition. Set the sampling frequency and collect multi-dimensional electrical and environmental parameters in real time during the operation of the power metering and acquisition equipment. Store the collected parameter data in the buffer. Step 2: Signal preprocessing. The raw parameter data collected in Step 1 is filtered to obtain filtered data. Step 3: Lifetime modeling and calculation. Analyze the component characteristic parameters in the filtered data from Step 2, calculate the health status value of the components, and trigger a lifetime warning if the health status value is lower than the preset health threshold. Step 4: Fault feature extraction. Time-frequency analysis and local spectral feature extraction are performed on the current signal in the filtered data to detect whether the energy of a specific frequency band exceeds the preset arc fault threshold in order to identify arc fault characteristics. Step 5: Dynamic threshold adjustment. Based on the environmental parameters collected in Step 1, the protection threshold adapted to the current environment is calculated in real time through the threshold adjustment model. Step Six: Risk Assessment and Decision Making. A multi-feature weighted decision-making mechanism is adopted to integrate and analyze the health status value in Step Three, the fault feature detection results in Step Four, and the dynamic protection threshold in Step Five to determine whether to trigger an early warning or protection action. Step 7: Protection action execution. When the protection action is triggered as determined in Step 6, the relay element is driven to cut off the circuit and the fault event information is recorded. When the warning is triggered, the warning information is generated. Step 8: Data reporting and communication. The health status information from Step 3, the event information from Step 7, and the early warning information are compressed and encoded, and reported to the operation and maintenance platform through the communication module. The communication module supports dynamic switching between dual channels.

2. The method for multiple active security protection of an electricity meter and a converged terminal system according to claim 1, characterized in that: The sampling frequency in step one is set to 4kHz. The real-time acquisition in step one is achieved through an ADC acquisition unit. The multi-dimensional electrical parameters in step one include voltage, current and leakage current. The environmental parameters include temperature and humidity.

3. The method for multiple active security protection of an electricity meter and a converged terminal system according to claim 1, characterized in that: The sampling frequency set in step one is based on Shannon's sampling theorem, and the formula of Shannon's sampling theorem is as follows: in, This refers to the actual sampling frequency. It refers to the highest frequency component in a signal.

4. The method for multiple active security protection of an electricity meter and a converged terminal system according to claim 1, characterized in that: The raw parameter data in step two is filtered using a finite-length unit impulse response filter. The formula for the finite-length unit impulse response filter is as follows: in, This refers to the filtered first... The output value of each point, It refers to the original input signal, where For the current point, This refers to the historical point index. It refers to the coefficients of the filter. This refers to the order of the filter, the filter output. From current and historical input signals With filter coefficients Obtained through convolution.

5. The method for multiple active security protection of an electricity meter and a converged terminal system according to claim 1, characterized in that: The component characteristic parameters in step three include the equivalent series resistance of the capacitor and the operating frequency of the relay. The lifetime modeling calculation in step three is based on the component degradation law. The component degradation law uses an exponential decay model to calculate the change of the equivalent series resistance of the capacitor over time. The formula for the exponential decay model is as follows: in, Refers to time The measured value of the equivalent series resistance at time t. This refers to the initial condition of the capacitor when it leaves the factory. value, This refers to the attenuation coefficient based on capacitor materials and processes, obtained by calculating the ratio of ripple voltage to current in real time. Value, combined with factory initial value and material attenuation coefficient Calculate the health status of the capacitor.

6. The method for multiple active security protection of an electricity meter and a converged terminal system according to claim 1, characterized in that: The time-frequency analysis in step four is performed using short-time Fourier transform, and the local spectral features are extracted using windowing and segmentation. The fault feature signals in step four include fault arc signals and overcurrent signals, and the formula for the short-time Fourier transform is as follows: in, It refers to time and frequency The short-time Fourier transform result at the point, This refers to the input current signal. It refers to the time variable. It refers to a window function used to capture signals. In time Nearby local segments, It refers to the complex exponential function, used in Fourier transform to convert a signal from the time domain to the frequency domain. Refers to the integral variable The differential.

7. The method for multiple active security protection of an electricity meter and a converged terminal system according to claim 1, characterized in that: The protection threshold in step five is the leakage current protection threshold. The protection threshold adapted to the current environment in step five is calculated in real time using a linear interpolation model, combined with the threshold settings for dry and humid environments and the humidity sensitivity coefficient. The formula for the linear interpolation model is as follows: in, This refers to the current humidity. The calculated dynamic leakage current protection threshold, This refers to the standard threshold under dry conditions. This refers to the high sensitivity threshold in humid environments. This refers to the preset humidity threshold.

8. The method for multiple active security protection of an electricity meter and a converged terminal system according to claim 1, characterized in that: The fusion analysis in step six is ​​performed through a multi-feature weighted decision mechanism, which includes normalizing the characteristics of arc faults, current change rate and leakage current, and assigning preset weights to each feature. When the comprehensive risk score is greater than or equal to the tripping threshold, the protection action is triggered; When the warning threshold is less than or equal to the comprehensive risk score and less than the tripping threshold, a warning is triggered. The formula for the multi-feature weighted decision-making mechanism is as follows: in, This refers to the actions taken based on the comprehensive risk score; it is the output of this decision-making mechanism. This refers to a comprehensive risk score, which reflects the current risk level of electricity metering and data acquisition equipment. This refers to the tripping threshold. This refers to the warning threshold.

9. The method for multiple active security protection of an electricity meter and a converged terminal system according to claim 1, characterized in that: The fault event information in step seven includes the fault type, fault occurrence timestamp, and multi-dimensional electrical and environmental parameter data at the time of the fault. The protection action execution process in step seven follows a relay-driven timing model, the formula of which is as follows: in, This refers to the trip signal. This refers to the comprehensive risk score. This refers to the tripping threshold. This refers to the fault current. This refers to the instantaneous action value of the hardware; When the comprehensive risk score Exceeding the trip threshold or fault current Exceeding the instantaneous motion value of the hardware When the output is high (1), it drives the relay to operate.

10. The method for multiple active security protection of an electricity meter and a converged terminal system according to claim 1, characterized in that: The compression encoding in step eight uses differential encoding to transmit only necessary status information, including health status, event type, and timestamp or differential data. The communication module in step eight has dual channels, including high-speed power line carrier and high-frequency radio. The formulas included in step eight are as follows: in, This refers to the value at the current moment, which is the current data point used for difference calculation. It refers to the value at the previous moment, which is the comparison data point with the current data point by subtracting it; Differential encoding only sends the current value. Compared with the previous value If the difference is small, then no data will be sent.

Images