Multiple active security protection method for meter and fusion terminal system

Abstract

The application discloses a kind of multiple active security protection methods of electric meter and fusion terminal system, it is related to electric energy metering protection field, including setting sampling frequency, and the real-time collection of multi-dimensional electric parameters and environmental parameters when electric energy metering and collection equipment operates, the original parameter data collected is filtered, and the data after filtering is obtained, life modeling calculation, the characteristic parameters of component in the data after filtering are analyzed, and the health state value of component is calculated.The application is based on the fusion degradation model of capacitor ESR and relay action frequency, the built-in component degradation model is used, the passive protection is changed into active early warning by life modeling and fault feature extraction, the traditional aftertreatment is changed into early warning, and then it is beneficial to effectively predict the residual life of component, to identify aging risk in advance, which can effectively prevent electrical fire and metering misalignment and other risk conditions, and improve equipment operation safety.

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 operation 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 in traditional equipment will usually be irreversibly aged 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 means for predicting the remaining life of components, it cannot identify aging risks 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 risks 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:

[0007] Step one: high-frequency data acquisition, set the sampling frequency, and collect the multi-dimensional electric parameters and environmental parameters of electric energy metering and acquisition equipment in real time during operation, and store the collected parameter data into buffer area;

[0008] Step two: signal preprocessing, filtering the original parameter data collected in step one to obtain filtered data;

[0009] 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;

[0010] 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.

[0011] 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.

[0012] 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.

[0013] 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.

[0014] 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.

[0015] Preferably, the sampling frequency in step one is set to 4kHz, the real-time acquisition in step one is realized through an ADC acquisition unit, the multi-dimensional electrical parameters in step one include voltage, current and leakage current, and the environmental parameters include temperature and humidity.

[0016] Preferably, 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:

[0017]

[0018] in, This refers to the actual sampling frequency. It refers to the highest frequency component in a signal.

[0019] Preferably, the original parameter data in step two is filtered using a finite-length unit impulse response filter, the formula of which is as follows:

[0020]

[0021] 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, is referred to as a history point index, is referred to as a coefficient of a filter, is referred to as an order of a filter, a filter output is obtained by convolution of a current and a history input signal and a filter coefficient .

[0022] Preferably, the component characteristic parameters in step three include equivalent series resistance of the capacitor and frequency of relay action, and the life modeling calculation in step three is based on a component degradation law, the component degradation law adopts an exponential decay model to calculate the change of the equivalent series resistance of the capacitor over time, and the formula of the exponential decay model is as follows:

[0023]

[0024] wherein, is referred to as time at which the equivalent series resistance is measured, is referred to as an initial value of the capacitor at factory, is referred to as a decay coefficient based on capacitor material and process, the current value is obtained by real-time calculation of the ratio of ripple voltage and current, and the capacitor health status is calculated in combination with the initial value at factory and the material decay coefficient .

[0025] Preferably, the time-frequency analysis in step four is analyzed by means of short-time Fourier transform, and the local spectral features are extracted by means of windowed segmentation; the fault feature signals in step four include fault arc signals and overcurrent signals, and the formula of the short-time Fourier transform is as follows:

[0026]

[0027] wherein, is referred to as a short-time Fourier transform result at time and frequency , is referred to as an input current signal, is referred to as a time variable, is referred to as a window function for intercepting a local segment of the signal near time , is referred to as a complex exponential function for converting a signal from time domain to frequency domain in Fourier transform, is referred to as a differential of integral variable .

[0028] Preferably, the protection threshold in step five is a leakage protection threshold, and the calculation of the protection threshold adapted to the current environment is performed by a linear interpolation model, combined with threshold setting and humidity sensitivity coefficient in dry and humid environments for real-time calculation, and the formula of the linear interpolation model is as follows:

[0029]

[0030] Among them, refers to the standard threshold in dry environment, refers to the dynamic leakage protection threshold calculated according to the current humidity, refers to the high sensitivity threshold in humid environment, refers to the preset humidity critical point.

[0031] Preferably, the fusion analysis in step six is analyzed by a multi-feature weighted decision mechanism, which includes normalizing arc fault features, current change rate and leakage current features, and giving each feature a preset weight.

[0032] When the comprehensive risk score is greater than or equal to the trip threshold, a protection action is triggered;

[0033] When the early warning threshold is less than or equal to the comprehensive risk score and less than the trip threshold, a pre-warning is triggered;

[0034] The formula of the multi-feature weighted decision mechanism is as follows:

[0035]

[0036] 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 degree of the electric energy metering and collecting equipment, refers to the trip threshold, refers to the early warning threshold.

[0037] 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:

[0038]

[0039] Among them, refers to the trip signal, refers to the comprehensive risk score, refers to the trip threshold,​ is a fault current, is a hardware instantaneous action value;

[0040] When the comprehensive risk score exceeds the tripping threshold or the fault current exceeds the hardware instantaneous action value , a high level 1 is output, driving the relay to act.

[0041] 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:

[0042]

[0043] wherein, is the value at the current moment, and is the current data point for difference calculation, is the value at the previous moment, and is the comparison data point for difference with the current data point;

[0044] The difference encoding only sends the difference between the current value and the previous value , and if the change is small, it is not sent.

[0045] Technical effects and advantages of the present application:

[0046] (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 the operation and maintenance cost by 30%, uses variable passive protection for active early warning through life modeling and fault feature extraction, changes the traditional post-disposal into pre-warning, and is thus beneficial to effectively predict the remaining life of the components, to identify the aging risk in advance, to effectively prevent the occurrence of electrical fire and measurement inaccuracy, and to improve the equipment operation safety;

[0047] (2) The present application adopts current slope and high-frequency harmonic joint analysis, identifies the omen within 500ms before overcurrent and arc fault occurs and cuts off, is convenient for improving the response speed, integrates the current change rate and high-frequency feature recognition for advanced safety protection, dynamically adjusts the leakage protection threshold through integration of environmental humidity sensing and insulation resistance monitoring, and thus builds a multiple-depth protection system integrating equipment life protection, electrical fire protection and adaptive personal protection, covers the full-dimensional safety needs of equipment, power grid and person, and solves the single defect of traditional equipment protection;

[0048] (3) The 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;

[0049] (4) The application uses double-channel switching communication, HPLC and HRF double-channel dynamic switching, which is beneficial to ensure that the power outage event reporting delay is less than 2s, improve communication reliability, ensure uninterrupted data reporting, and adapt to the setting mode of complex power grid environment;

[0050] (5) The application reduces communication traffic through difference coding, which is beneficial to optimize the static power consumption of the device, and meets the development trend of energy saving and high efficiency of the power system. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 It is a flowchart of the multiple active safety protection method of the application. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the application will be described clearly and completely 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, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0053] The application provides a multiple active safety protection method for an electric meter and a fusion terminal system, as shown in formula (I): Figure 1 The method comprises the following specific steps:

[0054] 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 device in real time, and store the collected parameter data in the buffer area, and the algorithm is realized as follows:

[0055] Sampling frequency setting logic:

[0056] A constant parameter sampling frequency is defined, and the value is 4000. The parameter represents that the electric signal will be collected 4000 times per second, that is, the sampling frequency is 4kHz. This sampling frequency is used as the time reference of the entire data acquisition system;

[0057] ADC interrupt handling program logic:

[0058] When the ADC (analog-to-digital converter) completes a conversion and triggers an interrupt, the following operations are performed:

[0059] Read the voltage data:

[0060] Read the value from the data register of the ADC module numbered 1, and store the read value in a variable named global voltage sample value;

[0061] Read the current data:

[0062] Read the value from the data register of the ADC module numbered 2, and store the read value in a variable named global current sample value;

[0063] Data buffering operation:

[0064] Store the set of sample data stored in the global voltage sample value and the global current sample value in the pre-allocated buffer according to the collection order, and ensure that these original sample data are saved for subsequent signal processing;

[0065] Interrupt preparation:

[0066] Prepare for the next ADC conversion interrupt request after completion, and ensure that the sampling process can continue at a frequency of 4kHz;

[0067] Step 2: Signal preprocessing, filtering the original parameter data collected in step 1 to obtain filtered data, which is conducive to continuously updating the buffer and participating in convolution calculation, ensuring filtering effect and reducing computational complexity, ensuring real-time performance, and the algorithm is implemented as follows:

[0068] Filter order setting logic:

[0069] Define a fixed parameter named filter order, whose value is 5, representing that the FIR (Finite Impulse Response) filter is a 5th order filter, i.e. the latest 5 times of input data and corresponding filter coefficients need to be operated in the filtering calculation process.

[0070] FIR filter function execution logic:

[0071] This function is used to receive a single original input data, and output the filtered clean data after processing by FIR filtering algorithm. The function input is the current original input data and the data buffer, and the output is the filtered data. The specific steps are as follows:

[0072] Initialize index variable Define a static index variable and initialize it to 0, which is used to record the storage position 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 position can be continuously tracked.

[0073] Store the current input data Write the current original input data received by this function to the position pointed to by the index variable in the data buffer, complete the caching of the current data.

[0074] Update buffer index to calculate new index value: new index = (current index + 1) divided by filter order and take the remainder.

[0075] When the index does not reach N-1 (e.g. N=5, index from 0 to 3), the new index is the current index +1, realizing the buffer position moving backward in turn;

[0076] When the index reaches N-1 (e.g. N=5, index is 4), the new index is reset to 0, realizing the buffer circular storage, ensuring that the latest N times of input data are always stored.

[0077] Calculate filter output data Initialize the filtered data to 0, and perform the following operations through a loop (from i=0 to i=N-1):

[0078] In each loop, multiply the historical data in the i-th position of the data buffer with the i-th coefficient of the preset FIR filter coefficient to get a single product result;

[0079] Add all single product results to the filtered data to finally get the output value of this filtering (i.e. complete the convolution operation of input data and filter coefficient).

[0080] Return the filter result The filtered data calculated is output as a function for subsequent signal analysis or processing modules, realizing the suppression of noise and interference in the original data;

[0081] 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 it is beneficial to the aging law of the core components of the device (capacitor and relay), to build a life prediction model, realize the change 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:

[0082] Capacitor equivalent series resistance calculation logic:

[0083] This part of the 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:

[0084] Function initialization and input definition:

[0085] Define a function named Calculate ESR that takes two input parameters: voltage data (i.e., the voltage signal across the capacitor) and current data (i.e., the current signal flowing through the capacitor). The function outputs the calculated ESR value as a floating-point number.

[0086] Extract ripple component:

[0087] Calculate ripple voltage: Call the Extract Ripple function with the input voltage data. This function separates the ripple component from the voltage signal (i.e., the AC fluctuation part after removing the DC or fundamental wave). The result is the ripple voltage.

[0088] Calculate ripple current: Similarly, call the Extract Ripple function with the input current data. This function separates the ripple component from the current signal. The result is the ripple current (floating-point number).

[0089] Calculate and return ESR value:

[0090] According to the physical definition of ESR (equivalent series resistance of a capacitor is equal to the ratio of the ripple voltage across it to the ripple current flowing through it), divide the ripple voltage by the ripple current to obtain the equivalent series resistance (ESR) value of the current capacitor. Return this value as the result of the Calculate ESR function.

[0091] Capacitor health degree calculation and early warning judgment logic:

[0092] This part of logic is used to evaluate the health status of the capacitor based on the calculated ESR value and trigger an early warning when the health degree is too low. The specific steps are as follows:

[0093] Calculate health score:

[0094] Define a floating-point number variable named Health Score to quantify the health status of the capacitor.

[0095] 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, then multiply 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. Therefore, the smaller the ratio of initial ESR / current ESR, the more serious the aging of the capacitor, and the lower the health degree.

[0096] Health threshold judgment and early warning trigger:

[0097] Set health threshold: A health threshold (e.g. 60 points) is pre-stored in the system, which represents the risk of capacitor aging being too high if it is below this value, according to the capacitor model, application scenario, etc.

[0098] Threshold comparison: Compare the calculated health score with the health threshold.

[0099] 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;

[0100] 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 and facilitate the realization of fault pre-identification. The algorithm is implemented as follows:

[0101] FFT initialization logic:

[0102] Define a fast real Fourier transform (RFFT) instance named "fft" to store the configuration information required for FFT calculation;

[0103] Call the fast RFFT initialization function to initialize the configuration of the above "fft" instance. The configuration parameters include the length of the FFT (specified by FFT_LENGTH), and a float-type array fft_output with a length of FFT_LENGTH is created to store the result data after FFT calculation;

[0104] FFT calculation and fault judgment logic:

[0105] 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), and the energy value of the frequency band corresponding to the 50th position in the transformed result is checked (this position corresponds to a specific frequency range);

[0106] Compare the energy value with the preset arc fault threshold. If the energy value of the frequency band exceeds the arc fault threshold, call the protection trigger function to start the corresponding protection mechanism;

[0107] Step five: Dynamic threshold adjustment, according to the environmental parameters collected in step one, the protection threshold suitable for the current environment is calculated in real time through the threshold adjustment model, which is beneficial to dynamically adjusting the leakage protection threshold according to the environmental humidity, improving the adaptability of personal safety protection, and the algorithm is implemented as follows:

[0108] Base definitions and function declaration logic:

[0109] Declare a function named read_humidity_sensor with no input parameters and a return value of a 16-bit unsigned integer. This function is used to obtain the current environmental humidity data (e.g., percentage RH value) from the humidity sensor.

[0110] Define the base leakage threshold as 30.0 mA. This value is the default environmental (e.g., 50% RH humidity) leakage protection reference value, ensuring basic personal safety protection.

[0111] Define the humidity sensitivity coefficient as 0.1. This coefficient represents the adjustment magnitude of the leakage protection threshold when the humidity changes by 1% RH, quantifying the impact of environmental humidity on the protection threshold.

[0112] Dynamic threshold calculation and setting logic:

[0113] Obtain real-time humidity: Call the read_humidity_sensor function to obtain the real-time humidity data of the current environment and store it in the humidity variable (e.g., detect the current humidity as 60% RH).

[0114] Calculate the dynamic threshold: Take 50% RH as the humidity reference point, and calculate the dynamic leakage protection threshold adapted to the current humidity using the formula: dynamic threshold = base leakage threshold - humidity sensitivity coefficient × (current humidity - 50.0). Example: If the current humidity is 60% RH, substituting into the formula gives: dynamic threshold = 30.0 - 0.1 × (60.0 - 50.0) = 29.0 mA, achieving a decrease in threshold when humidity increases (improving protection sensitivity).

[0115] Set threshold to take effect: Call the set_leakage_threshold function and pass in the calculated dynamic threshold, updating the trigger threshold of the leakage protection comparator in the system, so that the protection mechanism adapts to the current environmental humidity, avoiding false actions or protection failures.

[0116] Step six: risk judgment decision, using 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 in step five, to determine whether to trigger a warning or protection action, thereby facilitating the fusion of multi-dimensional information such as life, fault, and environment, avoiding single feature misjudgment, ensuring decision accuracy, and the algorithm formula is as follows:

[0117] Emergency fault determination and trip logic:

[0118] Judge the following three types of emergency risk conditions (any one of them will trigger an emergency action):

[0119] If the arc fault detection result is true (i.e., the system has identified a fault arc);

[0120] or the current rate of change is greater than the critical current rate of change (i.e. the current rises sharply, there is a risk of overcurrent / short circuit);

[0121] or the leakage current is greater than the dynamic leakage protection threshold (i.e. the leakage current exceeds the standard, there is a risk of electric shock);

[0122] If any of the above conditions are met, the action instruction is set to trip immediately, and the relay is driven to cut off the circuit to isolate the fault.

[0123] Potential risk determination and early warning logic:

[0124] 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 deteriorates, there is a risk of future leakage / short circuit), the action instruction is set to send a warning, generate a warning information and report to the operation and maintenance platform, and prompt early troubleshooting and maintenance.

[0125] Default logic for conditions not triggered:

[0126] If none of the above emergency fault conditions and potential risk conditions are met, the current action instruction is not changed (default to continue monitoring state, no trip or warning action);

[0127] Step seven: protection action execution, when the protection action is triggered in step six, 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, which is conducive to realizing the rapid cutting off of the fault and recording the event, ensuring the reliability of the protection, and the algorithm is implemented as follows:

[0128] Trip function execution logic:

[0129] Define a function called trigger trip, which performs the following operations:

[0130] Control the relay action:

[0131] Access the GPIO port register where the relay control pin is located, set the BSRR register of the port, and set the relay control pin to the off state, which will physically cut off the circuit and achieve circuit protection;

[0132] Record the fault event:

[0133] 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.) in the system log, which is convenient for subsequent analysis and tracing;

[0134] Step eight: data reporting communication, the health status information in step three, event information and early warning information in step seven are compressed and encoded, and reported to the operation and maintenance platform through the communication module. The communication module supports dual-channel dynamic switching, which is conducive to efficient and reliable data reporting, reduces communication costs and power consumption, and uses high-speed power line carrier HPLC (high transmission rate and good stability) under normal working conditions. When the high-speed power line carrier HPLC communication is interrupted (such as line failure), it automatically switches to high-frequency wireless HRF, ensuring communication reliability ≥ 99.99%, and for emergency information such as power failure, the reporting delay is ≤ 2s, ensuring that operation and maintenance personnel respond in a timely manner. The algorithm is implemented as follows:

[0135] Report data packet structure definition logic:

[0136] Define a data structure named report data packet to integrate the key information that needs to be reported;

[0137] The data packet contains three fields:

[0138] Health score: float type, storing the health status value of the device component;

[0139] Event type: 8-bit unsigned integer, indicating the type of reported event (such as failure, warning, etc.);

[0140] Timestamp: 32-bit unsigned integer, recording the time information of the event occurrence;

[0141] Enable compression and packaging properties for the data packet, remove padding bytes in the data structure, and minimize the data packet size;

[0142] Data sending logic:

[0143] Call the high-speed power line carrier sending function, the first parameter of the function is the address of the report data packet constructed above, indicating the data to be sent;

[0144] The second parameter of the function is the size of the data packet, used to inform the receiver of the data length;

[0145] After executing the function, the data packet will be sent to the target receiving end through the high-speed power line carrier communication method.

[0146] 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.

[0147] 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:

[0148] ;

[0149] 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.

[0150] 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:

[0151] ;

[0152] 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.

[0153] 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:

[0154] ;

[0155] 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.

[0156] 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:

[0157] ;

[0158] 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.

[0159] 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:

[0160] ;

[0161] 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.

[0162] Furthermore, the fusion analysis in step six is ​​conducted through a multi-feature weighted decision-making mechanism, which includes normalizing the characteristics of arc faults, current change rate, and leakage current, and assigning preset weights to each feature.

[0163] When the comprehensive risk score is greater than or equal to the tripping threshold, the protection action is triggered;

[0164] When the warning threshold is less than or equal to the comprehensive risk score and less than the tripping threshold, a warning is triggered.

[0165] The formula for the multi-feature weighted decision-making mechanism is as follows:

[0166] ;

[0167] 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 , it means that the device risk is extremely high, and a trip action needs to be triggered to cut off the circuit to ensure safety, when the comprehensive risk score greater than the threshold value and less than or equal to , it means that the device has a certain risk, and a warning needs to be issued , prompting relevant personnel to pay attention to the device state, when the comprehensive risk score less than or equal to , it means that the device risk is within an acceptable range, and no special action is taken.

[0168] Specifically, the fault event information in step seven includes fault type, fault timestamp, multi-dimensional electrical 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. The formula of the relay driving timing model is as follows:

[0169] ;

[0170] Among them, is a trip signal, a control signal for driving the relay action, is a comprehensive risk score, which is obtained by weighted calculation of multiple features such as arc fault characteristics, current rate of change and leakage current characteristics, and reflects the comprehensive risk degree of the device, is a trip threshold value, when the comprehensive risk score exceeds the threshold value, it is considered that the device risk reaches the degree that needs to be tripped, is a fault current, which refers to the current value when the device fails during operation, is a hardware instantaneous action value, which is a current threshold value set at the hardware level, when the fault current exceeds the value, the hardware will directly trigger the protection action;

[0171] When the comprehensive risk score exceeds the trip threshold value or the fault current exceeds the hardware instantaneous action value , output high level 1, drive the relay action;

[0172] Specifically, the compression encoding in step eight adopts a difference encoding method, and only necessary state information is transmitted, 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 in step eight is as follows:

[0173] ;

[0174] Among them, ​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, and the difference encoding only sends the current value and the difference with the previous value If the change is small, it is not sent, thereby reducing the data transmission amount and improving communication efficiency while ensuring effective reporting of key information.

[0175] A multiple active security protection system of an electric meter and a fusion terminal system, comprising:

[0176] A parameter acquisition module, which is configured to set a sampling frequency, and to collect multi-dimensional electrical parameters and environmental parameters in real time when the electric energy metering and collecting device is running, and to store the collected parameter data in a buffer;

[0177] A signal filtering preprocessing module, which is configured to perform signal preprocessing and filter the collected original parameter data to obtain filtered data;

[0178] A state monitoring and early warning module, which is configured to perform life modeling calculation, analyze the characteristic parameters of the components in the filtered data, and calculate the health state value of the components;

[0179] A fault feature recognition module, which is configured to perform fault feature extraction, perform time-frequency analysis and local frequency spectrum 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 to identify the arc fault feature;

[0180] An adaptive protection threshold calculation module, which is configured to perform dynamic threshold adjustment, and to calculate the protection threshold adaptive to the current environment in real time through a threshold adjustment model according to the collected environmental parameters;

[0181] A decision module, which is configured to adopt a multi-feature weighted decision mechanism, to perform fusion analysis on the health state value, the fault feature detection result, and the dynamic protection threshold, and to judge whether to trigger a warning or a protection action;

[0182] A protection module, which is configured to drive a relay element to cut off the circuit when it is judged to trigger a protection action, and to record fault event information, and to generate warning information when it is judged to trigger a warning;

[0183] A data reporting module, which is configured to compress and encode the health state information, the event information, and the warning information, and to report them to an operation and maintenance platform through a double communication channel dynamic switching mode.

[0184] Finally, it should be noted that the above only describes the preferred embodiments of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art will appreciate that the technical solutions described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalent features, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

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 drive 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