High-precision measurement method and system for electric energy meter

Abstract

The invention relates to an electric energy meter high-precision measurement method and system, and relates to the technical field of electric energy metering, and the method comprises the steps: extracting a combined load feature vector of a real-time current signal, and comparing the combined load feature vector with a preset feature library to accurately determine a current load mode, thereby adaptively selecting an optimal metering strategy parameter set; the active power of fundamental waves is accurately calculated through a hybrid harmonic separation technology so as to eliminate interference of harmonic components on metering; and carrying out dynamic compensation on the fundamental wave power by using the selected strategy parameters, and carrying out integration to obtain final electric energy. In addition, the system periodically utilizes an internal high-stability reference source to calibrate and compensate system-level errors of the whole measurement link through a parallel closed-loop self-correction process. According to the invention, the metering accuracy, adaptability and long-term stability of the electric energy meter under various complex load conditions can be significantly improved.

Description

Technical Field

[0001] This application relates to the technical field of electricity metering, and in particular to a high-precision measurement method and system for electricity meters. Background Technology

[0002] With the rapid development of electronic technology and digital signal processing technology, modern electricity meters have shifted from traditional mechanical induction type to electronic and intelligent type. Electronic electricity meters use dedicated metering chips or microprocessors to process the collected voltage and current signals in real time to calculate power grid parameters such as active power, reactive power, and electrical energy, achieving higher measurement accuracy and richer functional integration.

[0003] However, current electricity metering technology still reveals its inherent limitations when facing increasingly complex electricity usage environments. On the one hand, the composition of modern power loads has undergone profound changes. The widespread use of nonlinear loads such as switching power supplies, frequency converters, and energy-saving lighting fixtures injects a large amount of harmonic current into the power grid. These harmonic components not only distort the current waveform but also make the power characteristics of the load diverse and time-varying. Traditional electricity meters typically use a fixed set of measurement algorithms and compensation parameters. This strategy is often optimized for standard linear loads during design. When the actual load is a nonlinear load with huge differences in harmonic characteristics, this fixed metering strategy is difficult to adapt, resulting in significant errors in power calculation under harmonic interference and a sharp decline in metering accuracy.

[0004] On the other hand, the performance of electronic components that make up the measurement chain of an electricity meter, such as transformers, amplifiers, and analog-to-digital converters, changes gradually over time. This process is known as component aging. Simultaneously, the operating temperature of the electricity meter is not constant; temperature fluctuations also cause drift in the parameters of key analog components. These slow changes caused by aging and environmental factors collectively introduce a systematic and cumulative measurement error, leading to a gradual deterioration in the accuracy of the electricity meter during long-term operation. Current solutions mainly rely on rigorous calibration before shipment. Once deployed in the field, there is a lack of effective means to monitor and compensate for this long-term accuracy drift in real time, making it difficult to guarantee high-precision metering throughout its entire lifespan. Summary of the Invention

[0005] The purpose of this application is to provide a high-precision measurement method and system for electricity meters, which aims to solve the problems of decreased accuracy of existing electricity meters when measuring nonlinear and harmonic loads due to fixed metering strategies, as well as long-term drift errors caused by component aging and environmental changes.

[0006] Firstly, the high-precision measurement method for electricity meters provided in this application adopts the following technical solution: The method includes: Simultaneous acquisition of voltage and current signals yields discrete voltage and current signals. Joint feature extraction is performed on the discretized current signal to construct a load feature vector that can characterize the current load size and harmonic characteristics; The load feature vector is matched with a preset load pattern feature library to determine the current load pattern, and the metering strategy parameter set corresponding to the load pattern is selected. The discretized voltage and current signals are subjected to mixed harmonic separation to separate the fundamental signal, and the fundamental active power is calculated based on the fundamental signal. The fundamental active power is dynamically compensated using the metering strategy parameter set to obtain the final energy value. The parallel closed-loop self-calibration steps include: periodically or when triggered by preset conditions, switching the input of the metering channel to an internal reference source, measuring the internal reference source to calculate the current measurement error of the system, and updating the global calibration coefficients according to the measurement error to compensate for systematic errors introduced by component aging or environmental changes.

[0007] The step of constructing the load feature vector includes: calculating the effective value of the current, the total harmonic distortion of the current, and the harmonic composition feature vector based on the discretized current signal, and using the three together as the load feature vector.

[0008] The total harmonic distortion of the current (D) THDi It is determined by the following formula: Among them, D THDi I represents the total harmonic distortion of the current. fund I is the effective value of the fundamental current. h Here, H represents the effective value of the h-th harmonic current, H is the highest harmonic order analyzed, and h is the harmonic order index.

[0009] The harmonic characteristic vector includes the ratio of the effective value of the current of at least one selected harmonic to the effective value of the fundamental current.

[0010] The step of matching the load feature vector with a preset load pattern feature library includes: The current load mode is determined by calculating the Euclidean distance between the load feature vector and each reference feature vector in the load mode feature library, and selecting the load mode corresponding to the one with the smallest distance.

[0011] The steps of the hybrid harmonic separation include: The discrete voltage and current signals are subjected to Fast Fourier Transform to obtain their spectra; In the spectrum, the amplitudes of all frequency components except the fundamental frequency are set to zero to obtain the fundamental frequency spectrum; The fundamental frequency spectrum is subjected to an inverse fast Fourier transform to reconstruct the fundamental frequency signal without harmonics.

[0012] The fundamental active power (P) is calculated based on the fundamental signal. fund The steps are determined by the following formula: Among them, P fund The fundamental active power; u fund [n] represents the discretized fundamental voltage signal, i fund [n] represents the discretized fundamental current signal, N T is the number of sampling points within the fundamental frequency period, and n is the discrete-time index.

[0013] The step of applying the metering strategy parameter set for dynamic compensation includes: Gain correction coefficients and bias correction coefficients are extracted from the set of metering strategy parameters, and the fundamental active power is linearly corrected.

[0014] The triggering condition for the closed-loop self-calibration step is: The closed-loop self-calibration is automatically triggered according to the preset calibration cycle, or when the environmental sensor detects that the change in environmental parameters exceeds the preset threshold.

[0015] Secondly, the high-precision measurement system for electricity meters provided in this application adopts the following technical solution: The signal acquisition module is used to synchronously acquire voltage and current signals to obtain discrete voltage and current signals; A joint feature extraction module, connected to the signal acquisition module, is used to perform joint feature extraction on the discretized current signal to construct a load feature vector that can characterize the current load size and harmonic characteristics. The metering strategy selection module, connected to the joint feature extraction module, is used to match the load feature vector with a preset load mode feature library to determine the current load mode and select the metering strategy parameter set corresponding to the load mode; the fundamental power calculation module, connected to the signal acquisition module, is used to perform mixed harmonic separation on the discretized voltage signal and current signal to separate the fundamental signal and calculate the fundamental active power based on the fundamental signal. A dynamic compensation module, connected to the metering strategy selection module and the fundamental power calculation module, is used to apply the metering strategy parameter set to dynamically compensate the fundamental active power in order to obtain the final energy value. The closed-loop self-calibration module, connected to the signal acquisition module, is used to periodically or when triggered by preset conditions to switch the input of the metering channel to the internal reference source, measure the internal reference source to calculate the current measurement error of the system, and update the global calibration coefficient according to the measurement error to compensate for systematic errors introduced by component aging or environmental changes.

[0016] In summary, this application includes at least one of the following beneficial technical effects: 1. This invention constructs a load feature vector that includes the effective value of current, total harmonic distortion, and harmonic composition characteristics. This enables accurate determination of the real-time operating condition of the load. Then, based on the determined different load modes, it selectively calls the matching metering strategy parameter set for dynamic compensation. This targeted compensation method avoids the errors caused by fixed parameter metering when the load changes, and ensures the measurement accuracy under wide dynamic range and different load characteristics. 2. This invention employs hybrid harmonic separation technology to separate the fundamental signal from the acquired voltage and current signals, and the calculation of active power is based solely on this fundamental signal. This method eliminates the interference of each harmonic component on the power integration process from the source of calculation, so that the measurement results reflect the transmission and consumption of fundamental electrical energy, thereby improving the reliability of the measurement results in environments with strong harmonic sources such as industrial frequency converters. 3. This invention sets up a parallel closed-loop self-calibration mechanism, which can be triggered periodically or automatically when environmental conditions change. It calculates the current measurement error of the system by measuring the internal reference source and updates the global calibration coefficient accordingly. This active, closed-loop calibration method can effectively compensate for the accuracy drift caused by factors such as aging of electronic components or changes in ambient temperature, ensuring that the electricity meter maintains stable metering performance throughout its entire life cycle. Attached Figure Description

[0017] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a system architecture diagram of the present invention; Figure 3 This is a detailed structural diagram of the signal acquisition module of the present invention; Figure 4 This is a schematic diagram of the joint load feature vector construction process of the present invention; Figure 5 This is a frequency domain schematic diagram of the hybrid harmonic separation process of the present invention; Figure 6 This is a schematic diagram illustrating the working principle of the closed-loop self-calibration module of the present invention; Explanation of reference numerals in the attached figures: 10, Signal acquisition module; 20, Joint feature extraction module; 30, Metrology strategy selection module; 40, Fundamental power calculation module; 50, Dynamic compensation module; 60, Closed-loop self-calibration module. Detailed Implementation

[0018] The following is in conjunction with the appendix Figure 1 - Appendix Figure 6 This application will be described in further detail below.

[0019] Example 1: High-precision measurement method for electricity meters, referring to... Figures 1-6 This includes the following steps: S100: Synchronously acquire and preprocess the voltage and current signals of the power grid.

[0020] In step S100, the voltage and current signals of the power grid are synchronously acquired and preprocessed. This step provides high-fidelity discretized data input for subsequent feature extraction and power calculation.

[0021] Specifically, the high-voltage signal of the power grid is first converted into a low-voltage analog voltage signal suitable for the input range of the analog-to-digital converter by a voltage transformer or a high-precision voltage divider resistor network. At the same time, the current signal in the line is converted into a low-voltage analog voltage signal proportional to the line current by a current transformer or a shunt.

[0022] Before analog-to-digital conversion, the two low-voltage analog signals pass through independent analog anti-aliasing filters. These filters are low-pass filters, and their cutoff frequency is set to the sampling frequency f of the subsequent analog-to-digital converter. s Less than half of f, that is, less than f s / 2, this step is used to filter out frequency components in the signal that are higher than the Nyquist frequency, to prevent frequency aliasing during sampling and to ensure the effectiveness of the discretized signal.

[0023] The two filtered analog signals are fed into a high-resolution, dual-channel synchronous analog-to-digital converter (ADC). This ADC is driven by a unified, highly stable clock signal to ensure that the sampling operations of the voltage and current channels occur precisely at the same time, thus guaranteeing time alignment of the sampling points. The ADC resolution is set to 24 bits to ensure the quantization accuracy of the signal amplitude, especially its ability to capture weak current signals under light loads. The sampling frequency f of the ADC is... s The selection of f must satisfy f s >2·f h,max The condition, where f h,max This is the frequency of the highest harmonic that this method needs to analyze.

[0024] Through the above process, the continuous analog voltage signal u(t) and current signal i(t) are converted into discretized digital sequences, and the outputs are discrete-time voltage signal u[n] and discrete-time current signal i[n], whose relationship can be expressed as: u[n]=u(n·T s ); i[n]=i(n·T s ); Among them, T s The sampling period is 1 / f. s ; n is the discrete-time index, which is a non-negative integer. These two synchronized digital sequences will serve as data inputs for subsequent steps S200 and S400.

[0025] S200, Perform joint feature extraction operation on the discretized current signal i[n] output in step S100.

[0026] The purpose of this step is to generate a multi-dimensional load feature vector that can comprehensively and quantitatively describe the current load operating state. This vector will serve as the direct basis for decision-making in the subsequent step S300.

[0027] First, a data analysis window containing N sampling points is extracted from the input discrete-time current signal sequence i[n]. The signal within the data window is then transformed from the time domain to the frequency domain using the Fast Fourier Transform algorithm to obtain the complex spectrum sequence I[k].

[0028] The mathematical expression for this transformation process is: Where i[n] is the sampled value of the discrete-time current signal, N is the number of points in the Fast Fourier Transform algorithm, i.e., the length of the data analysis window, k is the frequency domain index (k = 0, 1, ..., N-1), j is the imaginary unit, and the amplitude |I[k]| of the complex spectrum I[k] represents the intensity of the current signal at the corresponding frequency component.

[0029] Based on the time-domain signal i[n] and the frequency-domain spectrum I[k], the load feature vector V is calculated. L The various components of the load feature vector V. L It is defined as a vector containing three components: V L =[I rms D THDi V HCF ].

[0030] The first component is the effective value of the current, I. rms This is used to characterize the overall scale of energy consumption by a load. It is obtained by averaging and taking the square root of the sum of the squares of discrete current samples within an analysis window, as shown in the following formula: Where i[n] is the sampled value of the discrete-time current signal, and N is the total number of sampling points within the analysis window.

[0031] The second component is the total harmonic distortion (THD) of the current. THDi It is used to quantify the degree of distortion of the current waveform relative to a pure sine wave. This value is obtained by calculating the ratio of the total effective value of all harmonic components to the effective value of the fundamental component. The calculation formula is as follows: Among them, I fund I is the effective value of the fundamental current. h Here, I represents the effective value of the h-th harmonic current, where h is the harmonic order index, HH is the highest harmonic order analyzed, and I... fund and I h The values ​​can all be calculated from the amplitude of the spectrum I[k] at the corresponding frequency point.

[0032] The third component is the harmonic eigenvector V. HCF This vector, used to provide detailed indicators of load nonlinearity to distinguish different types of nonlinear loads, consists of the normalized amplitudes of several key subharmonics. Specifically, specific subharmonics with high load type differentiation are selected, and their effective current value I is calculated. h With the fundamental current effective value I fund The ratio, for example, the vector can be constructed as: Through this normalization process, the eigenvector mainly reflects the relative distribution structure of harmonics, and its influence on changes in the absolute magnitude of the load is reduced.

[0033] Finally, the calculated effective value of the current I rms Total harmonic distortion of current D THDi Harmony constitutes the eigenvector V HCF Combined in a predetermined order, a complete load feature vector V is formed. L The vector is then output to step S300.

[0034] S300, the load feature vector V generated in step S200 L The process is performed to determine the current load pattern and select the corresponding set of metering strategy parameters.

[0035] This step is the core link connecting load status perception and metering behavior adjustment.

[0036] This method pre-builds a load pattern feature library in memory. This library consists of multiple entries, each representing a typical load pattern M that has been pre-calibrated and characterized. j Each entry contains two parts of data: One part is the reference feature vector V of this load mode. ref,j The other part is the set of measurement strategy parameters P associated with this model. cfg,j Reference feature vector V ref,j It is standard data obtained by measuring and statistically analyzing standard, known load types under different operating conditions.

[0037] When step S200 outputs the real-time load feature vector V L Next, this step uses a matching algorithm to determine its load pattern. This algorithm traverses all reference feature vectors V in the load pattern feature library. ref,j And calculate the real-time load feature vector V one by one. L With each reference feature vector V ref,j The Euclidean distance d between them j The formula for calculating this distance is: Among them, I rms D THDi and V HCF It is the real-time load feature vector V L The component; I rms,j D THDi,j , and V HCF,j It is the j-th reference eigenvector V ref,j The amount.

[0038] After completing all distances d j After calculation, the system compares the results and determines the entry with the smallest distance value, which corresponds to the load mode M. opt This means that the load mode of the current power grid is determined, and the determination process can be expressed as follows:

[0039] Once the optimal matching load pattern M is determined opt The system then retrieves and loads the metering strategy parameter set P associated with the load pattern from the load pattern feature library. cfg,opt This parameter set is a data structure containing multiple preset parameters, including but not limited to: the gain correction coefficient G used to correct the scaling error. opt The bias correction factor O used to correct fixed deviations opt , and a set of digital filter coefficients suitable for this load mode.

[0040] Ultimately, the selected measurement strategy parameter set P cfg,opt It is output to step S500 for precise dynamic compensation of the subsequently calculated fundamental active power.

[0041] S400: Perform hybrid harmonic separation on the discretized voltage signal u[n] and current signal i[n] output in step S100, and calculate the fundamental active power based on the separated fundamental signal.

[0042] This step aims to eliminate the influence of harmonic components on active power measurement.

[0043] This step receives complete discrete voltage signals u[n] and current signals i[n] containing the fundamental frequency and harmonics. First, fast Fourier transform is applied to these two signal sequences to obtain their respective complex spectra U[k] and I[k].

[0044] In the frequency domain, harmonic separation is performed by determining the frequency domain index k corresponding to the fundamental frequency based on the nominal fundamental frequency of the power grid and the parameters of the fast Fourier transform. fund Subsequently, two new spectral sequences, U fund [k] and I fund [k], the rules for generating these two new sequences are: Only the original spectrum is retained at the fundamental frequency index k fund and its conjugate frequency index Nk fund The complex value at the specified position is set to zero, while the complex values ​​at all other frequency index positions are set to zero.

[0045] After completing frequency domain filtering, the new fundamental frequency spectrum sequence U fund [k] and I fund [k] Perform the inverse fast Fourier transform, which converts the frequency domain signal back to the time domain, outputting a clean, harmonic-free discretized fundamental voltage signal u. fund [n] and discretized fundamental current signal i fund [n].

[0046] After obtaining the pure fundamental signal sequence, the fundamental active power is calculated. First, at each discrete time point n, the in-phase fundamental voltage sample value is multiplied by the fundamental current sample value to obtain the instantaneous fundamental power p at that moment. fund [n]: p fund [n] = u fund [n]·i fund [n]; Among them, u fund [n] represents the discretized fundamental voltage signal, i fund [n] represents the discretized fundamental current signal.

[0047] Fundamental active power p fundDefined as the average instantaneous fundamental power over a complete fundamental period, the instantaneous fundamental power is summed at all sampling points within the fundamental period, and then divided by the total number of sampling points N in that period. T The fundamental active power p is obtained. fund The calculation formula is as follows: Where, N T Let n be the number of sampling points within the fundamental frequency period, and n be the discrete-time index. The calculated fundamental active power P... fund It is output to step S500.

[0048] In step S500, the fundamental active power P output in step S400 is... fund Dynamic compensation is performed, and the compensated power value is integrated over time to obtain the cumulative electrical energy.

[0049] This step is the final execution stage for achieving high-precision measurement.

[0050] This step receives two data inputs: the fundamental active power P calculated in step S400. fund And the metering strategy parameter set P selected by step S300 that corresponds to the current load mode. cfg,opt The gain correction coefficient G in this parameter set opt and bias correction coefficient O opt It is used for linear correction of fundamental active power.

[0051] The compensation operation is implemented through a dynamic compensation function, which calculates the input fundamental active power P. fund Adjustments are made based on the correction coefficients to obtain the corrected active power P. corrected The compensation process is determined by the following formula: P corrected =G opt ·P fund +O opt ; Among them, P corrected G represents the corrected active power. opt P is the selected gain correction factor used to compensate for scaling errors that may exist under the current load conditions. fund The input fundamental active power; O opt The selected bias correction factor is used to compensate for any fixed deviations that may exist under the current load conditions.

[0052] Obtain the corrected active power P corrected The system then performs discrete-time integration to calculate and update the cumulative energy value. This integration process is performed periodically within a fixed time interval Δt. The update of the cumulative energy EE is accomplished using the following formula: E new =E old +P corrected ·Δt; Among them, E new This is the updated cumulative energy value after this calculation; E old This represents the cumulative electrical energy value stored in the previous calculation; P corrected Δt represents the corrected active power obtained in this calculation; Δt is the time interval between two consecutive power calculations, which is a preset constant value of the system.

[0053] Finally, the new cumulative energy value E was calculated. new The energy register, stored in non-volatile memory, is used for subsequent display, recording, and data communication.

[0054] S600 is a parallel closed-loop self-calibration process.

[0055] This step operates independently of the main measurement process from S100 to S500. Its function is to actively detect and compensate for the overall, slowly changing errors in the measurement link caused by aging of system components or changes in environmental conditions.

[0056] The self-calibration step is triggered under two conditions: one is periodic triggering, which is controlled by the system's internal timer and automatically starts a calibration process at a preset time interval; the other is event triggering, which starts the calibration process immediately when the system's built-in environmental sensor detects that the change in the current environmental parameter compared to the recorded value at the time of the last calibration exceeds a preset threshold. In addition, the system will also execute this self-calibration process once every time it is powered on.

[0057] When the self-calibration process is triggered, the system control unit first disconnects the input terminal of the signal acquisition module 10 from the external voltage and current transformers via an analog multiplexer, and simultaneously connects it to an internal high-stability reference source. This reference source generates a precise DC reference voltage V that does not change with time or temperature. ref This voltage is applied simultaneously to the input terminals of both the voltage measurement channel and the current measurement channel.

[0058] The signal acquisition module 10 then immediately applied the reference voltage V ref A complete measurement was performed to obtain the digital measurement results in the voltage and current channels, respectively, D. u,meas and D i,meas Meanwhile, the system uses the known reference voltage V ref Using the precise value of the analog-to-digital converter and the ideal transfer function, we can calculate the theoretically ideal digital quantity D that should be obtained. ideal .

[0059] Subsequently, the current system gain error ∈ is calculated based on the measured value and the ideal value. G This error is defined as the normalized deviation of the measured value from the ideal value, and is calculated using the following formula: This gain error ∈ G The overall gain drift of the entire measurement link was quantified.

[0060] The system calculates the gain error ∈ G Update the global gain calibration coefficient C stored in non-volatile memory. G The update algorithm for this coefficient aims to offset the currently measured error. The update formula is: Among them, C G,new It is the updated global gain calibration coefficient, C G,old This is the coefficient value before the update. In this way, the new calibration coefficient can compensate for any existing system gain drift. This updated global gain calibration coefficient C G,new It will be used in all subsequent main measurement processes as a final correction factor applied to the energy calculation results.

[0061] Example 2: High-precision measurement system for electricity meters, refer to Figure 2 It includes a signal acquisition module 10, a joint feature extraction module 20, a metering strategy selection module 30, a fundamental power calculation module 40, a dynamic compensation module 50, and a closed-loop self-calibration module 60.

[0062] The signal acquisition module 10 has its input terminals connected to voltage and current transformers for acquiring power grid signals. The module integrates an anti-aliasing filter and a high-precision synchronous analog-to-digital converter. When performing step S100, the module provides its output discretized voltage signal u[n] and current signal i[n] to the fundamental power calculation module 40, and provides the discretized current signal i[n] to the joint feature extraction module 20.

[0063] The joint feature extraction module 20, whose input terminal is connected to the current signal output terminal of the signal acquisition module 10, is configured to perform the operation of step S200. By performing fast Fourier transform, current RMS value calculation, total harmonic distortion calculation, and harmonic feature vector calculation on the input current signal i[n], the load feature vector V is finally constructed. L This module generates the load feature vector V. L Output to metering strategy selection module 30.

[0064] The metering strategy selection module 30, whose input is connected to the output of the joint feature extraction module 20, contains a storage unit storing a load pattern feature library. This module is configured to perform the operation of step S300 by calculating the input real-time load feature vector V. L The Euclidean distance between the current load pattern and the reference feature vectors in the database is used to determine the current load pattern, and the corresponding metering strategy parameter set P is retrieved from the database. cfg,opt This module will select the measurement strategy parameter set P cfg,opt Output to dynamic compensation module 50.

[0065] The fundamental power calculation module 40 has its input terminal connected to the voltage and current signal output terminals of the signal acquisition module 10. This module is configured to perform the operation of step S400, which involves performing frequency domain transformation, fundamental component extraction, and inverse frequency domain transformation on the input voltage signal u[n] and current signal i[n] to separate the pure fundamental signal u. fund [n] and i fund [n], and based on this, the fundamental active power P is calculated. fund This module will calculate the fundamental active power P. fund Output to dynamic compensation module 50.

[0066] The dynamic compensation module 50, whose input terminals are connected to the output terminals of the metering strategy selection module 30 and the fundamental power calculation module 40 respectively, is configured to perform the operation of step S500. It applies the input metering strategy parameter set P. cfg,opt For the input fundamental active power P fund Perform linear correction to obtain the corrected active power P. corrected The power value is then integrated over time to update the cumulative energy value stored in the non-volatile memory.

[0067] The closed-loop self-calibration module 60 includes an internal reference source and an analog multiplexer. The output of the analog multiplexer is connected to the input of the signal acquisition module 10. The module is configured to execute the parallel calibration process of step S600. When a trigger signal is received, the module controls the analog multiplexer to switch the input of the signal acquisition module 10 to the internal reference source and initiates a measurement. Based on the deviation between the measurement result and the nominal value of the reference source, the module calculates and updates the global calibration coefficient. This coefficient is used by the system to perform the final system-level calibration of all measurement results.

[0068] In one implementation, the functions of the joint feature extraction module 20, the metering strategy selection module 30, the fundamental power calculation module 40, and the dynamic compensation module 50 can all be implemented by executing a preset firmware program through a central microprocessor or digital signal processor.

Claims

1. A high-precision measurement method for an electric energy meter, characterized in that, The method comprises the following steps: synchronously collecting voltage signals and current signals to obtain discretized voltage signals and current signals; jointly extracting features from the discretized current signals to construct a load feature vector capable of representing current load size and harmonic characteristics; matching the load feature vector with a preset load mode feature library to determine a current load mode and select a metering strategy parameter set corresponding to the load mode; separating mixed harmonics from the discretized voltage signals and current signals to separate fundamental signals and calculate fundamental active power based on the fundamental signals; applying the metering strategy parameter set to dynamically compensate the fundamental active power to obtain a final electric energy value; a parallel closed-loop self-correction step, which comprises periodically or upon a preset condition being triggered, switching input of a metering channel to an internal reference source, measuring the internal reference source to calculate a current measurement error of the system, and updating a global calibration coefficient based on the measurement error to compensate for systematic errors introduced by component aging or environmental changes.

2. The high-precision measurement method of the electric energy meter according to claim 1, characterized in that, The step of constructing the load feature vector comprises: calculating current effective values, total harmonic distortion of current, and a harmonic composition feature vector based on the discretized current signals, and taking the three as the load feature vector.

3. The high-precision measurement method of the electric energy meter according to claim 2, characterized in that, The total harmonic distortion of current is determined by the following formula: where D THDi is the total harmonic distortion of the current; I fund is the fundamental current effective value, I h is the hth harmonic current effective value, H is the highest harmonic number analyzed, and h is the harmonic number index.

4. The high-precision measurement method of an electric energy meter according to claim 1, characterized in that, The step of matching the load feature vector with the preset load mode feature library comprises: determining the current load mode by calculating Euclidean distances between the load feature vector and each reference feature vector in the load mode feature library and selecting a load mode corresponding to the smallest distance.

5. The method of claim 1, wherein, The step of separating mixed harmonics comprises: performing fast Fourier transform on the discretized voltage signals and current signals to obtain their frequency spectrums; in the frequency spectrums, setting amplitudes of all frequency components other than the fundamental frequency point to zero to obtain a fundamental frequency spectrum; performing inverse fast Fourier transform on the fundamental frequency spectrum to reconstruct the fundamental signal without harmonics.

6. The method of claim 1, wherein, The step of calculating the fundamental active power based on the fundamental signal is determined by the following formula: where P fund is the fundamental active power; u fund [n] is the discretized fundamental voltage signal, i fund [n] is the discretized fundamental current signal, N T is the number of sampling points within the fundamental period, and n is the discrete time index.

7. The method of claim 1, wherein, The step of applying the metering strategy parameter set to perform dynamic compensation comprises: extracting a gain correction coefficient and a bias correction coefficient from the metering strategy parameter set and linearly correcting the fundamental active power.

8. The method of claim 1, wherein, The trigger condition of the closed-loop self-correction step is: automatically triggering the closed-loop self-correction according to a preset calibration period or when an environmental sensor detects that an environmental parameter changes by more than a preset threshold.

9. The method of claim 2, wherein, The harmonic composition feature vector comprises a ratio of current effective values of at least one selected harmonic to a fundamental current effective value.

10. A high-precision measurement system of an electric energy meter, according to the high-precision measurement method of an electric energy meter according to any one of claims 1-9, characterized in that, The method comprises: a signal collection module for synchronously collecting voltage signals and current signals to obtain discretized voltage signals and current signals; a joint feature extraction module for jointly extracting features from the discretized current signals to construct a load feature vector capable of representing current load size and harmonic characteristics; a metering strategy selection module configured to match the load feature vector with a preset load mode feature library to determine a current load mode and select a metering strategy parameter set corresponding to the load mode; a fundamental power calculation module configured to separate a fundamental signal from the discretized voltage signal and current signal through mixed harmonic separation, and calculate a fundamental active power based on the fundamental signal; a dynamic compensation module configured to apply the metering strategy parameter set to dynamically compensate the fundamental active power to obtain a final power value; a closed-loop self-correction module configured to periodically or upon a preset condition, switch an input of a metering channel to an internal reference source, measure the internal reference source to calculate a current measurement error of the system, and update a global calibration coefficient according to the measurement error to compensate for systematic errors introduced by component aging or environmental changes.

Images