High-precision voltage measurement method of intelligent voltage power meter

Abstract

The invention discloses a high-precision voltage measurement method for an intelligent voltage power meter, and relates to the technical field of electrical sensors, and the method comprises the steps: reckoning the time sequence state of a trend evolution vector, constructing a voltage behavior virtual body, and predicting the voltage characteristic change trend of a voltage waveform through the voltage behavior virtual body; comparing the time sequence consistency of the voltage characteristic change trend and the historical voltage record to obtain a consistency offset, jointly presuming the range crossing risk of the voltage waveform through the consistency offset and the trend evolution vector, and generating a range prediction indication quantity; and dynamically switching the measuring range of the voltage waveform according to the measuring range prediction indicating quantity, generating a correction waveform, executing period aggregation and amplitude-phase joint reconstruction on the correction waveform, and outputting high-precision measurement voltage. According to the innovative method of trend evolution mapping and construction of the voltage behavior virtual body, accurate prediction of voltage waveform characteristics is realized, dynamic range switching in the voltage measurement process is optimized, and the flexibility and precision of measurement are improved.

Description

Technical Field

[0001] This invention relates to the field of electrical sensor technology, and in particular to a high-precision voltage measurement method for an intelligent voltage power meter. Background Technology

[0002] With the continuous development of power equipment, voltage measurement technology has evolved from traditional analog voltmeters to modern digital voltage and power meters. Intelligent voltage and power meters utilize digital technology and embedded computing to accurately monitor grid voltage and analyze its waveform in real time, providing more precise energy metering and fault detection capabilities. Currently, the main technological directions of intelligent voltage and power meters include: high-precision voltage acquisition, noise and harmonic suppression, state quantification feature extraction, and evolution analysis. With the development of industrial automation and smart grids, voltage measurement technology is gradually moving towards higher precision, lower error, and stronger dynamic adaptability.

[0003] However, existing technologies still have some shortcomings, especially when voltage signals are affected by noise and harmonic interference. Traditional filtering methods often struggle to balance efficient filtering with signal fidelity, leading to a loss of voltage waveform accuracy. Although existing technologies reduce the impact of noise and harmonics on measurement results through filtering algorithms, they usually cannot completely eliminate complex nonlinear noise in external voltage signals, affecting the accuracy of voltage waveforms. Most existing methods rely on simple linear filtering or standard filters, which cannot adapt to the complex variations in different voltage waveforms, especially in environments with significant high-frequency waveforms and transient interference. Another technical bottleneck is the range selection problem. Most smart voltage and power meters rely on fixed ranges and cannot dynamically adjust the measurement range according to changes in the voltage signal. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a high-precision voltage measurement method for intelligent voltage and power meters to solve the problems of not being able to effectively suppress complex nonlinear noise and dynamically adjust the measurement range.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a high-precision voltage measurement method for an intelligent voltage and power meter, comprising:

[0008] External voltage signals and historical voltage records are collected, and noise and harmonics in the external voltage signals are suppressed by time-frequency domain filtering to obtain the voltage waveform;

[0009] Extract the state quantization features of the voltage waveform, perform state evolution mapping on the state quantization features, and generate a trend evolution vector;

[0010] The temporal state of the trend evolution vector is calculated, a voltage behavior virtual body is constructed, and the voltage characteristic change trend of the voltage waveform is predicted through the voltage behavior virtual body;

[0011] By comparing the voltage characteristic change trend with the time sequence consistency of historical voltage records, the consistency offset is obtained. The range out-of-bounds risk of the voltage waveform is estimated by combining the consistency offset with the trend evolution vector, and a range prediction indicator is generated.

[0012] The range of the voltage waveform is dynamically switched based on the range prediction indication, a corrected waveform is generated, and the cycle aggregation and amplitude-phase joint reconstruction are performed on the corrected waveform to output a high-precision measured voltage.

[0013] As a preferred embodiment of the high-precision voltage measurement method for the intelligent voltage and power meter described in this invention, the specific steps for obtaining the voltage waveform by suppressing noise and harmonics of the external voltage signal through time-frequency domain filtering are as follows.

[0014] The frequency density of the external voltage signal is statistically analyzed and the frequency range is divided to form a noise energy distribution spectrum;

[0015] Harmonic amplitude, harmonic energy gradient and harmonic contribution ratio of noise energy distribution spectrum are extracted to generate harmonic component indicator group;

[0016] Identify the high-frequency energy density and energy concentration of the noise energy distribution spectrum to form a transient sensitivity index;

[0017] The external voltage signal is differentially suppressed based on the harmonic component indicator group and the transient sensitivity index to generate a voltage waveform.

[0018] As a preferred embodiment of the high-precision voltage measurement method for the intelligent voltage and power meter described in this invention, the specific steps for extracting the state quantization features of the voltage waveform are as follows:

[0019] By statistically analyzing the zero-crossing period parameters, peak amplitude parameters, phase drift parameters, and rise slope parameters of the voltage waveform, a four-dimensional behavioral parameter set is constructed.

[0020] Behavioral coupling mapping and nonlinear scaling are performed on the four-dimensional behavioral parameter set to generate state quantization features.

[0021] As a preferred embodiment of the high-precision voltage measurement method for the intelligent voltage and power meter described in this invention, the specific steps for performing state evolution mapping on the state quantization features to generate a trend evolution vector are as follows.

[0022] The amplitude change increment, frequency drift increment, and phase shift increment of the state quantization characteristics are obtained to form a state evolution sequence;

[0023] Calculate the behavioral sensitivity component of the state quantification feature and generate a set of behavioral sensitivity coefficients;

[0024] The state evolution trajectory of the state evolution sequence is derived and trend synthesis is performed with the behavior sensitivity coefficient group to generate a trend evolution vector.

[0025] As a preferred embodiment of the high-precision voltage measurement method for the intelligent voltage and power meter described in this invention, the specific steps for calculating the temporal state of the trend evolution vector and constructing a voltage behavior virtual entity are as follows:

[0026] Extract the evolution direction of the trend evolution vector, perform direction consistency analysis and progressive intensity calibration, and generate trend-driven state groups;

[0027] Deconstruct and reorganize the behavioral and temporal hierarchies of trend-driven state groups and state quantification features to generate multi-time state mapping columns;

[0028] The amplitude, frequency, and phase patterns are extracted from the multi-time state mapping series, and a virtual voltage behavior is constructed through stability dominance analysis and trend persistence analysis.

[0029] As a preferred embodiment of the high-precision voltage measurement method for the intelligent voltage and power meter described in this invention, the specific steps for predicting the voltage characteristic change trend of the voltage waveform through a voltage behavior virtual entity are as follows:

[0030] By decomposing the composite amplitude shape, composite frequency shape, and composite phase shape of the voltage behavior virtual object into a time-consistent form, a behavior pattern sequence is generated.

[0031] The dominant relationship and temporal relationship of behavioral patterns are screened out and cross-behavioral influence coupling is carried out to construct a trend inference matrix;

[0032] The trend projection matrix is ​​combined with the trend trajectory according to the time dimension to generate the voltage characteristic change trend.

[0033] As a preferred embodiment of the high-precision voltage measurement method for the intelligent voltage and power meter described in this invention, the steps for comparing the voltage characteristic change trend with the temporal consistency of historical voltage records to obtain the consistency offset are as follows:

[0034] Perform characteristic deconstruction mapping on historical voltage records, and parse the historical voltage records into historical amplitude change trajectory, historical frequency change trajectory and historical phase change trajectory in time order to generate historical characteristic trajectory group;

[0035] The amplitude change trajectory, frequency drift trajectory, and phase shift trajectory are extracted from the voltage characteristic change trend, and compared with the time series differences of the historical characteristic trajectory group to generate the corresponding amplitude difference sequence, frequency difference sequence, and phase difference sequence.

[0036] The cross-dimensional consistency of the amplitude difference sequence, frequency difference sequence, and phase difference sequence is inferred separately, and then integrated to generate a consistency offset.

[0037] As a preferred embodiment of the high-precision voltage measurement method for the intelligent voltage and power meter described in this invention, the specific steps for estimating the range out-of-bounds risk of the voltage waveform by jointly using consistency offset and trend evolution vector to generate a range prediction indication are as follows.

[0038] Deconstruct the range sensitivity of the trend evolution vector and generate range sensitivity component sets;

[0039] By performing joint risk mapping between the consistency offset and the range sensitivity component group, a range out-of-bounds risk assessment vector is generated.

[0040] The risk levels of the range out-of-range risk assessment vector are divided and the risk levels are aggregated to generate the range prediction indicator.

[0041] As a preferred embodiment of the high-precision voltage measurement method for the intelligent voltage and power meter described in this invention, the specific steps for dynamically switching the voltage waveform range based on the range prediction indication to generate a corrected waveform are as follows:

[0042] The local amplitude envelope of the voltage waveform is inferred and compared with the range prediction indication to generate a range switching control parameter set.

[0043] Based on the range of the segmented mapped voltage waveform of the range switching control parameter group, a switching waveform segment is generated;

[0044] Amplitude continuity matching and transition segment interpolation reconstruction are performed on the switched waveform segments to generate the corrected waveform.

[0045] As a preferred embodiment of the high-precision voltage measurement method for the intelligent voltage and power meter described in this invention, the specific steps for performing period aggregation and amplitude-phase joint reconstruction on the corrected waveform to output a high-precision measured voltage are as follows.

[0046] Locate the periodic interval of the corrected waveform, and extract the amplitude sequence and phase sequence for adjacent periodic intervals respectively to generate periodic segmentation feature groups;

[0047] By comparing the steady-state amplitude and phase continuity of the periodic segmented characteristic group, an amplitude-phase consistency coefficient group is formed.

[0048] Based on the amplitude-phase consistency coefficient set, the corrected waveform is subjected to amplitude calibration and reshaping and phase continuity reconstruction to generate an amplitude-phase reconstructed waveform.

[0049] The amplitude and phase reconstruction waveform is subjected to periodic aggregation and multi-cycle averaging fusion to output a high-precision measured voltage.

[0050] The beneficial effects of this invention are as follows: By combining time-frequency domain filtering with harmonic component indication, noise and harmonics in the external voltage signal are successfully suppressed, improving the accuracy and stability of the voltage waveform; through the innovative method of trend evolution mapping and voltage behavior virtual body construction, accurate prediction of voltage waveform characteristics is achieved, enabling the voltage power meter to dynamically adapt to voltage changes and provide early warning of range overrun risks, thereby optimizing the dynamic range switching in the voltage measurement process and improving the flexibility and accuracy of measurement. Attached Figure Description

[0051] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a flowchart of a high-precision voltage measurement method for intelligent voltage and power meters.

[0053] Figure 2 A flowchart for constructing state quantization features, trend evolution vectors, and voltage behavior virtual entities.

[0054] Figure 3 This is a flowchart for generating high-precision voltage measurements.

[0055] Figure 4 A flowchart for preprocessing external voltage signals. Detailed Implementation

[0056] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0057] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0059] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a high-precision voltage measurement method for a smart voltage power meter, comprising the following steps:

[0060] S1. Acquire external voltage signals and historical voltage records, suppress noise and harmonics in the external voltage signals through time-frequency domain filtering, and obtain the voltage waveform.

[0061] S1.1 Statistically determine the frequency density of the external voltage signal and divide the frequency range to form a noise energy distribution spectrum.

[0062] It should be noted that the external voltage signal originates from the voltage input terminal of the smart voltage power meter and is connected to the sampling contact through an isolated sampling path. The sampling process uses a fixed sampling frequency to convert the continuous voltage waveform into a discrete sampling sequence. The historical voltage record consists of voltage amplitude records, frequency records, phase records, and corresponding time stamps within the operating cycle of the smart voltage power meter. The historical voltage record can be stored locally or obtained from synchronous archived data on the background recording platform.

[0063] A fixed sampling frequency and analysis length are set for the discrete sampling sequence obtained from the external voltage signal acquisition, and the discrete sampling sequence is divided into analysis segments of equal length. A window function is applied to each analysis segment to make the analysis segment transition smoothly at the time domain boundary and avoid frequency domain leakage. The analysis segments after window function processing are transformed from time coordinates to frequency coordinates through discrete Fourier transform and arranged in frequency coordinate order to obtain the amplitude data sequence. The amplitude data sequence is divided into several frequency groups according to a fixed frequency step size. Within each frequency group, the mean amplitude, peak amplitude, and frequency group differential momentum (root mean square of amplitude difference) between adjacent frequency points are statistically analyzed as a frequency band density index.

[0064] For all frequency groups, compare the mean amplitude in the band density index sequentially from low to high according to the frequency coordinates. Identify the set of continuous frequency groups whose mean amplitude shows a continuous upward trend, and mark the set of all frequency groups before the end of the continuous upward trend as the candidate set for the low-frequency interval. In the full frequency groups following the candidate set for the low-frequency interval, find the frequency position where the amplitude peak reaches the amplitude peak of the full frequency group, and form the candidate set for the main frequency interval with one directly adjacent frequency group before and after it. In the full frequency groups following the candidate set for the main frequency interval, find the set of all continuous frequency groups whose mean amplitude shows a continuous downward trend and whose frequency group differential momentum of adjacent frequency points shows a continuous upward trend, and mark them as the candidate set for the high-frequency interval. Combine the candidate set for the low-frequency interval, the candidate set for the main frequency interval, and the candidate set for the high-frequency interval in frequency order to form an interval set. Use the mean amplitude, peak amplitude, and frequency group differential momentum of each frequency group in the interval set as energy density descriptions. Arrange the energy density descriptions in frequency order to form a noise energy distribution spectrum.

[0065] S1.2 Extract the harmonic amplitude, harmonic energy gradient and harmonic contribution ratio of the noise energy distribution spectrum, and generate a harmonic component indicator group.

[0066] It should be noted that the average amplitude is read at integer multiples of frequency positions in the noise energy distribution spectrum, and the average amplitude at integer multiples of frequency positions is taken as the harmonic amplitude. In the sequential arrangement of integer multiples of frequency positions, the average amplitude of two adjacent integer multiples of frequency positions is taken, and the difference between the average amplitudes is taken as the harmonic energy gradient. The ratio of the peak value at each integer multiple of frequency position to the peak value across the entire frequency range of the noise energy distribution spectrum is taken as the harmonic contribution ratio. The harmonic amplitude, harmonic energy gradient, and harmonic contribution ratio are combined in integer multiples of frequency order to form a harmonic component indicator group.

[0067] S1.3 Identify the high-frequency energy density and energy concentration of the noise energy distribution spectrum to form a transient sensitivity index.

[0068] It should be noted that the amplitude peak value of each frequency group in the candidate set of the high-frequency interval in the noise energy distribution spectrum is read and used as the high-frequency energy density; the frequency group differential momentum of all frequency groups in the candidate set of the high-frequency interval in the noise energy distribution spectrum is read, and the first-order difference and second-order difference of the frequency group differential momentum are obtained in frequency order; the frequency position where the sign changes from positive to negative or from negative to positive is identified from the second-order difference, and all continuous frequency groups before the sign change are used as the high-energy differential set; all frequency group sets with adjacent relationships are searched in the high-energy differential set, and the frequency span of each frequency group set is used as the energy concentration; the product of the high-frequency energy density and the energy concentration is used as the transient sensitivity value representing the transient interference intensity of the frequency group, and the transient sensitivity values ​​of all frequency groups are arranged in frequency order to form the transient sensitivity index.

[0069] S1.4. Based on the harmonic component indicator group and transient sensitivity index, perform differentiated suppression on the external voltage signal to generate a voltage waveform.

[0070] It should be noted that, in the noise energy distribution spectrum, the amplitude data corresponding to each frequency group is read in frequency order to form a frequency domain amplitude sequence; for positions with integer multiples of frequency, the corresponding harmonic contribution ratio is read from the harmonic component indicator group, and the corresponding transient sensitivity value is read from the transient sensitivity index; the product of the harmonic contribution ratio and the transient sensitivity value is used as the harmonic suppression factor for the corresponding integer multiple of frequency position; for positions with non-integer multiples of frequency, the ratio of the transient sensitivity value to the maximum value of all transient sensitivity values ​​in the candidate set of the high-frequency interval is used as the transient suppression factor; the product of the amplitude data with the harmonic suppression factor and the transient suppression factor is used as the amplitude reduction amount, and the amplitude reduction amount is subtracted from the amplitude data to generate the suppressed amplitude data; all suppressed amplitude data are recombined in frequency order to generate a voltage waveform.

[0071] S2. Extract the state quantization features of the voltage waveform, perform state evolution mapping on the state quantization features, and generate a trend evolution vector.

[0072] S2.1 Statistically analyze the zero-crossing period parameters, peak amplitude parameters, phase drift parameters, and rising slope parameters of the voltage waveform to construct a four-dimensional behavioral parameter set.

[0073] It should be noted that the voltage waveform is scanned point by point in chronological order. Between every two adjacent sampling points, it is determined whether the voltage polarity changes. The time marker corresponding to the sampling point where the change occurs is recorded as the zero-crossing time point. All zero-crossing time points from negative to positive and all zero-crossing time points from positive to negative are recorded in chronological order. In each type of zero-crossing time point, the time difference between two adjacent zero-crossing time points is counted in chronological order. Each time difference is recorded as a zero-crossing cycle, and all zero-crossing cycles are arranged in chronological order as the zero-crossing cycle parameter.

[0074] Read the amplitude of all sampling points covered by each zero-crossing cycle, find the sampling point where the maximum amplitude is located, record the amplitude corresponding to the sampling point as the peak amplitude parameter of the corresponding zero-crossing cycle, and arrange all peak amplitude parameters in the order of zero-crossing cycles to form a peak amplitude parameter sequence.

[0075] A phase analysis operation is performed on the voltage waveform for all sampling points to obtain the phase angle value of each sampling point. The difference between the phase angle value corresponding to each sampling point and the phase angle value of the first sampling point of the zero-crossing cycle is obtained. Each difference is recorded as a phase drift parameter, and all phase drift parameters are arranged in chronological order to form a phase drift parameter sequence.

[0076] In each zero-crossing cycle, the amplitude change of adjacent sampling points is read from all sampling points covered, and the ratio of the amplitude change to the time interval between the corresponding sampling points is obtained as the rising slope parameter. All rising slope parameters in each zero-crossing cycle are arranged in chronological order to form a rising slope parameter sequence. The zero-crossing cycle parameter, peak amplitude parameter sequence, phase drift parameter sequence and rising slope parameter sequence are combined in the order of appearance to form a four-dimensional behavior parameter group.

[0077] S2.2 Perform behavior coupling mapping and nonlinear scaling on the four-dimensional behavior parameter set to generate state quantization features.

[0078] It should be noted that within each zero-crossing cycle, the product between the peak amplitude parameter and the rising slope parameter, the difference between the peak amplitude parameter and the phase drift parameter, and the ratio between the rising slope parameter and the phase drift parameter are obtained and arranged in the order of acquisition to form a behavioral coupling sequence. The minimum and maximum values ​​in the behavioral coupling sequence are obtained respectively, and the interval between the minimum and maximum values ​​is used as the rounding interval. The behavioral coupling sequence is rounded according to the rounding interval to obtain rounded values ​​in a uniform scale interval. All rounded values ​​are arranged in the time order of the zero-crossing cycle to generate state quantization features.

[0079] S2.3 Obtain the amplitude change increment, frequency drift increment, and phase shift increment of the state quantization characteristics to form a state evolution sequence.

[0080] It should be noted that, sequentially, each normalized value after the first normalized value in the state quantization feature is read, and the numerical difference between the current normalized value and the previous normalized value is recorded as the amplitude change increment at the current moment; according to the position of each normalized value within the normalization interval, the corresponding frequency offset position is obtained, and the difference between every two adjacent frequency offset positions is recorded as the frequency drift increment, and all frequency drift increments are arranged in the order of appearance to form a frequency drift increment sequence; according to the position of each normalized value within the normalization interval, the corresponding phase offset position is obtained, and the difference between every two adjacent phase offset positions is recorded as the phase offset increment, and all phase offset increments are arranged in the order of appearance to form a phase offset increment sequence; the amplitude change increment, the frequency drift increment sequence, and the phase offset increment sequence are combined in chronological order to form a state evolution sequence.

[0081] S2.4 Calculate the behavioral sensitivity component of the state quantification feature and generate a behavioral sensitivity coefficient set.

[0082] It should be noted that the amplitude change increment, frequency drift increment, and phase shift increment are read from the state evolution sequence in chronological order and normalized. The behavioral sensitivity components of the state quantization features are calculated, and the behavioral sensitivity components corresponding to all time indices are arranged in chronological order to form a behavioral sensitivity coefficient group.

[0083] The expression for calculating the behavioral sensitivity component is:

[0084] ;

[0085] in, Indicates the first Behavioral sensitivity components at each time index Indicates the first Normalized amplitude change increment at each time index Indicates the first Normalized frequency drift increment at each time index Indicates the first The phase offset increment after normalization at each time index. This represents a calculation constant used to prevent the denominator from being zero. Its value ranges from 0.000000001 to 0.00001. Among them, 0.000000001 is obtained by statistically analyzing the minimum effective range of all non-zero increments. It is used in scenarios where the normalized increment has high accuracy and extremely low noise, ensuring that the denominator is always non-zero and will not affect the increment structure. 0.00001 is obtained by statistically analyzing the random jitter range of the normalized increment. It is used in scenarios where the normalized increment is greatly affected by quantization noise or sensor error, so that the behavior sensitivity calculation is not dominated by noise.

[0086] S2.5. Derive the state evolution trajectory of the state evolution sequence and synthesize it with the behavior sensitivity coefficient group to generate a trend evolution vector.

[0087] It should be noted that the amplitude change increment, frequency drift increment, and phase shift increment in the state evolution sequence are arranged by time index to form a state change segment sequence. Using a fixed-length time window, window statistics (including amplitude window statistics, frequency window statistics, and phase window statistics) for all increments in the state change segment sequence are obtained and combined to form the state evolution trajectory. The minimum and maximum values ​​are determined for each behavioral sensitivity component in the behavioral sensitivity coefficient group. The difference between each behavioral sensitivity component and the minimum value is used as the corresponding de-shift sensitivity component. The de-shift sensitivity is then determined from all de-shift sensitivity components. The maximum value of each component is used as the weight coefficient. When the maximum value of the de-offset sensitivity component is greater than zero, the ratio of each de-offset sensitivity component to its maximum value is used as the corresponding weight coefficient. When the maximum value of the de-offset sensitivity component is equal to zero, the corresponding weight coefficient is recorded as zero. The arithmetic mean of all weight coefficients is obtained within each time window and recorded as the corresponding behavioral sensitivity weight. Within each time window, the product of all incremental window statistics and the corresponding behavioral sensitivity weight is used as a window state segment. All window state segments are rearranged according to the starting time index of the time window to form a trend evolution vector.

[0088] S3. Calculate the temporal state of the trend evolution vector, construct a voltage behavior virtual body, and predict the voltage characteristic change trend of the voltage waveform through the voltage behavior virtual body.

[0089] S3.1 Extract the evolution direction of the trend evolution vector, perform direction consistency analysis and progressive intensity calibration, and generate trend-driven state groups.

[0090] It should be noted that amplitude window statistics, frequency window statistics, and phase window statistics are extracted from the trend evolution vector. The difference between adjacent time indices for each type of statistic is obtained. A difference greater than zero is marked as an upward direction, a difference less than zero is marked as a downward direction, and a difference equal to zero is marked as a hold-off direction. The maximum absolute value of the differences of all statistics is obtained as the corresponding direction mark (including amplitude direction mark, frequency direction mark, and phase direction mark). Under each time index, the absolute value of the differences of all statistics is read and compared with the respective direction mark to obtain the amplitude progression intensity, frequency progression intensity, and phase progression intensity. All direction marks and all progression intensities are combined in a fixed order of "amplitude direction mark → amplitude progression intensity → frequency direction mark → frequency progression intensity → phase direction mark → phase progression intensity" to form a trend-driven state record, and arranged in chronological order to form a trend-driven state group.

[0091] S3.2 Deconstruct and reorganize the behavioral and temporal levels of trend-driven state groups and state quantification features to generate multi-time state mapping columns.

[0092] It should be noted that under each time index, the zero-crossing period parameter, peak amplitude parameter, amplitude direction marker, and amplitude progression intensity are combined to form an amplitude behavior level; the zero-crossing period parameter, frequency direction marker, and frequency progression intensity are combined to form a frequency behavior level; and the phase drift parameter, phase direction marker, and phase progression intensity are combined to form a phase behavior level. These are all registered together with the time index as time state mapping records. The amplitude direction marker, frequency direction marker, and phase direction marker in adjacent time state mapping records are compared in time index order. When the three types of direction markers are completely identical, the corresponding time index is assigned to the same time sequence number, and the time sequence number is written into their respective time state mapping records. All time state mapping records with time sequence numbers are arranged in ascending order of time index to generate a multi-time state mapping column.

[0093] S3.3 Extract the amplitude shape, frequency shape and phase shape from the multi-time state mapping column, and construct a voltage behavior virtual body through stability dominance analysis and trend persistence analysis.

[0094] It should be noted that, at the amplitude behavior level, zero-crossing period parameters, peak amplitude parameters, amplitude direction markers, and amplitude progression intensities are extracted and arranged by time index to form an amplitude pattern sequence; at the frequency behavior level, zero-crossing period parameters, frequency direction markers, and frequency progression intensities are extracted and arranged by time index to form a frequency pattern sequence; at the phase behavior level, phase drift parameters, phase direction markers, and phase progression intensities are extracted and arranged by time index to form a phase pattern sequence; the consistency of all direction markers at adjacent time indices in the amplitude pattern sequence, frequency pattern sequence, and phase pattern sequence is compared to see if all progression intensities maintain the same sign; continuous segments with consistent direction markers and consistent progression intensities are recorded as stable segments, and the duration of these continuous segments is recorded as a trend persistence parameter; all stable segments and trend persistence parameters are recombined by time index to construct a voltage behavior virtual body.

[0095] S3.4. Generate a sequence of behavioral patterns by decomposing the composite amplitude, composite frequency, and composite phase patterns of the voltage behavior virtual body through time-series consistency decomposition.

[0096] It should be noted that, under the time index, the zero-crossing period parameter, peak amplitude parameter, amplitude direction marker, amplitude progression intensity, frequency direction marker, frequency progression intensity, phase drift parameter, phase direction marker, and phase progression intensity corresponding to the time index in the amplitude pattern sequence, frequency pattern sequence, and phase pattern sequence are combined in a fixed order: "zero-crossing period parameter → peak amplitude parameter → amplitude direction marker → amplitude progression intensity → frequency direction marker → frequency progression intensity → phase drift parameter → phase direction marker → phase progression intensity," forming composite amplitude pattern, composite frequency pattern, and composite phase pattern records corresponding to the time index. In all composite pattern records, the time index is used to... The system sequentially compares whether the amplitude direction marker, frequency direction marker, and phase direction marker of adjacent records are all consistent, and compares whether the signs of the amplitude progression intensity, frequency progression intensity, and phase progression intensity are all consistent. When all direction markers and progression intensity signs are consistent, the current composite morphology record is assigned to the same time sequence group as the previous composite morphology record. When any direction marker or any progression intensity sign changes, the current composite morphology record is assigned to a new time sequence group. Within each time sequence group, the composite morphology records are arranged in time index order as behavioral morphology sequence segments, and all behavioral morphology sequence segments are combined in time sequence group starting time index order to generate a behavioral morphology sequence.

[0097] S3.5. Screen out the dominant relationship and temporal relationship of the behavior pattern sequence and perform cross-behavioral influence coupling to construct a trend inference matrix.

[0098] It should be noted that, within each behavioral pattern sequence segment, the frequency direction marker, frequency direction marker, and phase direction marker are counted to determine their frequency, and the direction marker with the highest frequency is recorded as the dominant direction marker. The arithmetic mean of all amplitude progression intensity, frequency progression intensity, and phase progression intensity within the current behavioral pattern sequence segment is calculated and used as the dominant progression intensity value. Between two adjacent behavioral pattern sequence segments, the dominant direction markers are compared to determine if they are the same, and the difference between the corresponding dominant progression intensity values ​​is obtained. The position where the dominant direction marker changes is recorded as a direction switching record, and the position where the difference in the dominant progression intensity value is positive or negative is recorded as a progression intensity change record. The dominant direction marker, dominant progression intensity value, direction switching record, and progression intensity change record corresponding to each time index are combined in a fixed order of "dominant direction marker → dominant progression intensity value → direction switching record → progression intensity change record" to form a trend inference factor record, and all trend inference factor records are arranged in time index order to form a trend inference matrix.

[0099] S3.6 Combine the trend projection matrix with trend trajectories according to the time dimension to generate the voltage characteristic change trend.

[0100] It should be noted that the dominant directional markers in each trend projection factor record are arranged in the order of amplitude direction marker, frequency direction marker, and phase direction marker to form directional trajectory segments. The dominant progressive intensity values ​​are arranged in the order of amplitude progressive intensity, frequency progressive intensity, and phase progressive intensity to form progressive trajectory segments. All directional trajectory segments and all progressive trajectory segments are combined in time index order to form directional trajectory sequences and progressive trajectory sequences, respectively. The changes of various directional markers in the directional trajectory sequences are compared between adjacent time indices. Combined with the increase and decrease relationships of various progressive intensities in the progressive trajectory sequences, time segments with consistent directions and continuous progressive relationships are combined to form trend trajectory segments. All trend trajectory segments are arranged in time index to generate voltage characteristic change trends.

[0101] S4. Compare the voltage characteristic change trend with the time sequence consistency of historical voltage records to obtain the consistency offset. Use the consistency offset and trend evolution vector to jointly estimate the range out-of-bounds risk of the voltage waveform and generate the range prediction indicator.

[0102] S4.1 Perform characteristic deconstruction mapping on the historical voltage records, and parse the historical voltage records into historical amplitude change trajectory, historical frequency change trajectory and historical phase change trajectory in chronological order to generate historical characteristic trajectory group.

[0103] It should be noted that the voltage amplitude record, frequency record, and phase record in the historical voltage record are read separately. The amplitude difference of the voltage amplitude record is calculated according to the adjacent time position, and each amplitude difference is arranged in chronological order to form a historical amplitude change trajectory. The frequency difference in the frequency record is calculated according to the adjacent time position, and each frequency difference is arranged in chronological order to form a historical frequency change trajectory. The phase difference in the phase record is calculated according to the adjacent time position, and each phase difference is arranged in chronological order to form a historical phase change trajectory. The historical amplitude change trajectory, historical frequency change trajectory, and historical phase change trajectory are combined in a fixed order of "historical amplitude change trajectory → historical frequency change trajectory → historical phase change trajectory" to form a historical characteristic trajectory group.

[0104] S4.2 Extract the amplitude change trajectory, frequency drift trajectory, and phase shift trajectory from the voltage characteristic change trend, and compare them with the time sequence differences of the historical characteristic trajectory group to generate the corresponding amplitude difference sequence, frequency difference sequence, and phase difference sequence.

[0105] It should be noted that the amplitude change trajectory, frequency drift trajectory, and phase shift trajectory are read from the voltage characteristic change trend by time index, and aligned with the historical amplitude change trajectory, historical frequency change trajectory, and historical phase change trajectory in the historical characteristic trajectory group by time marker. Under each time marker, the difference between the amplitude change trajectory of the voltage characteristic change trend and the historical amplitude change trajectory is taken as the amplitude difference value; the difference between the frequency drift trajectory and the historical frequency change trajectory is taken as the frequency difference value; and the difference between the phase shift trajectory and the historical phase change trajectory is taken as the phase difference value. The amplitude difference value, frequency difference value, and phase difference value are sorted in chronological order to generate the corresponding amplitude difference sequence, frequency difference sequence, and phase difference sequence.

[0106] S4.3. Infer the cross-dimensional consistency of the amplitude difference sequence, frequency difference sequence and phase difference sequence respectively, and integrate them to generate a consistency offset.

[0107] It should be noted that normalization is performed on the amplitude difference sequence, frequency difference sequence, and phase difference sequence respectively. Under each time index, it is determined whether the increase or decrease direction of the difference values ​​such as amplitude difference value, frequency difference value, and phase difference value is the same. When the increase or decrease direction is all the same, the time index is recorded as the consistency time index. Under each time index, the sum of the absolute values ​​of all difference values ​​is obtained to obtain the comprehensive difference intensity. The maximum value of the comprehensive difference intensity is found in the entire time index range, and the ratio of the comprehensive difference intensity corresponding to each time index to the maximum value is recorded as the cross-dimensional difference weight. The arithmetic mean of the cross-dimensional difference weights corresponding to all consistent time indices is recorded as the consistency offset.

[0108] S4.4 Deconstruct the range sensitivity of the trend evolution vector and generate a range sensitivity component group.

[0109] It should be noted that, in the trend evolution vector, the three types of window statistics (amplitude window statistics, frequency window statistics, and phase window statistics) contained in each window state segment are read sequentially according to the time index. Under each time index, the difference between the three types of window statistics and their respective average values ​​in all time indices is obtained and recorded as the amplitude range sensitivity, frequency range sensitivity, and phase range sensitivity of the corresponding time index. The minimum and maximum values ​​of the three types of range sensitivity are determined in all time indices, and the three types of range sensitivity are interval-rounded according to the interval between the minimum and maximum values ​​to make the range sensitivity of each type of range sensitivity fall within a uniform scale interval. The rounded amplitude range sensitivity, frequency range sensitivity, and phase range sensitivity are combined in a fixed order of "amplitude range sensitivity → frequency range sensitivity → phase range sensitivity" to form the range sensitivity component record of the corresponding time index. All range sensitivity component records are arranged in time index order to generate a range sensitivity component group.

[0110] S4.5. Perform joint risk mapping between the consistency offset and the range sensitivity component group to generate a range out-of-bounds risk assessment vector.

[0111] It should be noted that, under each time index, the amplitude offset, frequency offset, and phase offset in the consistency offset are multiplied by the range sensitivity component of the corresponding time index, and the multiplication results are recorded as the amplitude offset risk component, frequency offset risk component, and phase offset risk component. Within each time index, the three types of offset risk components are arithmetically averaged as the range out-of-bounds risk value. Within all time indices, the range out-of-bounds risk values ​​corresponding to each time index are recorded sequentially and arranged in ascending order of time index to form the range out-of-bounds risk evaluation vector.

[0112] S4.6 Divide the risk levels of the range out-of-range risk assessment vector and consolidate the risk levels to generate the range prediction indicator.

[0113] It should be noted that all range out-of-bounds risk values ​​are sorted in ascending order to obtain a sorted risk sequence. Within this sequence, the risk sequence is divided into three segments using equal division methods. The range out-of-bounds risk value corresponding to the end of the first segment is used as the first segmentation value, and the range out-of-bounds risk value corresponding to the end of the second segment is used as the second segmentation value. The first segmentation value distinguishes between low-risk and medium-risk intervals, and the second segmentation value distinguishes between medium-risk and high-risk intervals. Each range out-of-bounds risk value is read from the entire time index, and a value less than or equal to the first segmentation value is marked. The risk level is classified as low-risk. When the range out-of-bounds risk value is greater than the first segmentation value but less than or equal to the second segmentation value, it is marked as medium-risk. When the range out-of-bounds risk value is greater than the second segmentation value, it is marked as high-risk. The occurrence frequency of low-risk, medium-risk, and high-risk levels is counted in the entire time index sequence, and the risk level with the most occurrences is taken as the dominant risk level. The dominant risk level is converted into the corresponding range adjustment indication according to the mapping rules, where low-risk level corresponds to range maintenance, medium-risk level corresponds to range fine-tuning, and high-risk level corresponds to range increase, which is output as the range prediction indication.

[0114] S5. Dynamically switch the voltage waveform range based on the range prediction indication, generate a corrected waveform, perform period aggregation and amplitude-phase joint reconstruction on the corrected waveform, and output a high-precision measured voltage.

[0115] S5.1. Infer the local amplitude envelope of the voltage waveform and compare it with the range prediction indication to generate a range switching control parameter set.

[0116] It should be noted that the voltage waveform is divided into several voltage time windows in chronological order. In each voltage time window, the amplitude of all sampling points is read and the maximum amplitude is taken. These values ​​are then arranged in chronological order to form a local amplitude envelope. From the range prediction indication, the corresponding risk level is read sequentially according to the time window index. The upper limit voltage value of each range is determined based on the hardware input range and nominal range interval of the smart voltage power meter. All upper limit voltage values ​​are then arranged in ascending order to form a range set.

[0117] The difference between the local amplitude envelope and the upper limit voltage value of all ranges in the range set is obtained. The range with the smallest difference that is greater than zero is recorded as the range close to the limit. The difference of the range close to the limit is compared with the historical amplitude fluctuation range of the local amplitude envelope in the same time period (determined by the difference between the maximum and minimum amplitude values ​​of the same length voltage time window in the historical voltage record). When the difference of the range close to the limit is less than the historical amplitude fluctuation range, it is marked as a range switching candidate window. All range switching candidate windows are merged according to the continuity of time index and combined with the corresponding range close to the limit to form a range switching control parameter group.

[0118] S5.2. Generate the switching waveform segment based on the range of the segmented mapped voltage waveform of the range switching control parameter group.

[0119] It should be noted that, in the voltage waveform, all sampling points are scanned in the order of sampling time. The time stamp of the sampling point is compared with the starting time position of the range switching candidate window. The sampling point time that first reaches the starting time of a certain range switching candidate window is taken as the range switching position. The voltage waveform is divided into a pre-switching waveform segment and a post-switching waveform segment using the range switching position as the dividing point. At each range switching position, the last sampling point of the pre-switching waveform segment and the first sampling point of the post-switching waveform segment are saved respectively. All pre-switching waveform segments and post-switching waveform segments are combined in the order of appearance to generate the switching waveform segment.

[0120] S5.3 Perform amplitude continuity matching and transition segment interpolation reconstruction on the switched waveform segment to generate the corrected waveform.

[0121] It should be noted that, at the start time index of the switched waveform segment, the amplitude of the last sampling point before the switch and the amplitude of the first sampling point after the switch are read respectively, and the amplitude difference is obtained. The amplitude difference is evenly distributed according to the number of sampling points in the switched waveform segment to obtain the amplitude transition compensation amount corresponding to each sampling point. The amplitude transition compensation amount is superimposed on the amplitude data of each sampling point in the switched waveform segment to ensure a smooth transition of amplitude change, forming a sampling point amplitude sequence. Between the switched waveform segment and the adjacent unswitched waveform segments before and after it, the amplitudes of two adjacent sampling points at the switching boundary are used as interpolation endpoints. The total number of sampling points is counted, and the ratio between the amplitude difference of the interpolation endpoints and the total number of sampling points is used as the amplitude step size between sampling points. The interpolation amplitudes of each sampling point are accumulated point by point according to the sampling order to form a linearly increasing interpolation amplitude. The interpolation amplitudes of each sampling point are arranged according to the sampling order to form a transition segment sampling point sequence. The sampling point amplitude sequence, the transition segment sampling point sequence, and the original sampling point sequence of the unswitched waveform segment are recombined according to the time index to generate the corrected waveform.

[0122] S5.4 Locate the periodic interval of the corrected waveform, and extract the amplitude sequence and phase sequence according to the adjacent periodic intervals to generate periodic segmentation feature groups.

[0123] It should be noted that the corrected waveform is scanned point by point in chronological order. The signs of the amplitudes are compared between adjacent sampling points. When the sign changes from negative to positive or from positive to negative, the time mark of the corresponding sampling point is recorded as the corrected zero-crossing time point. The set of sampling points between two adjacent corrected zero-crossing time points is divided into a periodic interval in chronological order. The amplitudes of all sampling points in the periodic interval are read and arranged in the sampling order to form a periodic amplitude sequence. Phase analysis is performed on all sampling points in the same periodic interval to obtain the corresponding phase angle values ​​and arrange them in the sampling order to form a periodic phase sequence. The periodic amplitude sequence and the periodic phase sequence are combined in a fixed order of "periodic amplitude sequence → periodic phase sequence" to form a periodic segmented feature record. All periodic segmented feature records are arranged in the time start order of the periodic interval to generate a periodic segmented feature group.

[0124] S5.5. Compare the amplitude steady state and phase continuity of the periodic segmented characteristic group to form an amplitude-phase consistency coefficient group.

[0125] It should be noted that, in the amplitude sequence, the amplitude change between two adjacent sampling points is obtained, and the maximum and average values ​​of all amplitude changes are calculated. The ratio of the maximum value to the average value is used as the amplitude steady-state index for the corresponding period interval. In the phase sequence, the phase difference between two adjacent sampling points is obtained, and the sum of the absolute values ​​of all phase differences is calculated. The sum of the absolute values ​​is used as the phase continuity index for the corresponding period interval. In each period interval, the amplitude steady-state index and the phase continuity index are combined in a fixed order of "amplitude steady-state index → ​​phase continuity index" to form an amplitude-phase consistency record. All amplitude-phase consistency records are arranged in the order of period intervals to form an amplitude-phase consistency coefficient group.

[0126] S5.6. Based on the amplitude and phase consistency coefficient group, perform amplitude calibration and reshaping and phase continuity reconstruction on the corrected waveform to generate the amplitude and phase reconstructed waveform.

[0127] It should be noted that, in the amplitude-phase consistency coefficient group, the amplitude steady-state index and phase continuity index corresponding to each period interval are read in the order of period intervals, and all sampling points covered by the period interval are located in the corrected waveform; in the period interval where the amplitude steady-state index is greater than zero, the amplitude of all sampling points within the period interval is stretched or compressed according to the reciprocal of the amplitude steady-state index, so that the distribution of amplitude change is consistent with the change level corresponding to the steady-state index; in the period interval where the phase continuity index is greater than zero, the phase difference is adjusted according to the time order of the sampling points within the period interval, and the phase difference is proportionally corrected according to the reciprocal of the phase continuity index, so that the phase difference between adjacent sampling points maintains a consistent continuous change; the sampling points of all period intervals after amplitude calibration and phase continuity processing are recombined in time order to generate the amplitude-phase reconstructed waveform.

[0128] S5.7 Performs periodic aggregation and multi-period average fusion on the amplitude and phase reconstruction waveform to output a high-precision measured voltage.

[0129] It should be noted that the voltage polarity change position between every two adjacent sampling points in the amplitude-phase reconstruction waveform is detected, and all zero-crossing time points from negative to positive and from positive to negative are recorded. Within each type of zero-crossing time point, the time difference between adjacent zero-crossing time points is calculated sequentially, and each time difference is recorded as an aggregation interval. Within each aggregation interval, the amplitude and phase sequences of all sampling points are extracted. The mean of the amplitude sequence and the mean of the phase sequence in each aggregation interval are obtained and used as the representative amplitude and phase quantities for the corresponding aggregation interval. The arithmetic mean of all representative amplitude quantities and all representative phase quantities are obtained sequentially to obtain the multi-cycle amplitude fusion quantity and the multi-cycle phase fusion quantity. The multi-cycle amplitude fusion quantity and the multi-cycle phase fusion quantity are combined in a fixed order of "multi-cycle amplitude fusion quantity → multi-cycle phase fusion quantity" to output a high-precision measured voltage.

[0130] In summary, this invention successfully suppresses noise and harmonics in external voltage signals by combining time-frequency domain filtering with a harmonic component indication group, thereby improving the accuracy and stability of the voltage waveform. Furthermore, through an innovative method of trend evolution mapping and the construction of a voltage behavior virtual body, it achieves accurate prediction of voltage waveform characteristics, enabling the voltage power meter to dynamically adapt to voltage changes and provide early warnings of range exceedance risks. This optimizes the dynamic range switching during voltage measurement, improving measurement flexibility and accuracy.

[0131] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A high-precision voltage measurement method for an intelligent voltage and power meter, characterized in that: include, External voltage signals and historical voltage records are collected, and noise and harmonics in the external voltage signals are suppressed by time-frequency domain filtering to obtain the voltage waveform; Extract the state quantization features of the voltage waveform, perform state evolution mapping on the state quantization features, and generate a trend evolution vector; The temporal state of the trend evolution vector is calculated, a voltage behavior virtual body is constructed, and the voltage characteristic change trend of the voltage waveform is predicted through the voltage behavior virtual body; By comparing the voltage characteristic change trend with the time sequence consistency of historical voltage records, the consistency offset is obtained. The range out-of-bounds risk of the voltage waveform is estimated by combining the consistency offset with the trend evolution vector, and a range prediction indicator is generated. The range of the voltage waveform is dynamically switched based on the range prediction indication, a corrected waveform is generated, and the cycle aggregation and amplitude-phase joint reconstruction are performed on the corrected waveform to output a high-precision measured voltage.

2. The high-precision voltage measurement method for an intelligent voltage and power meter as described in claim 1, characterized in that: The specific steps for suppressing noise and harmonics in the external voltage signal and obtaining the voltage waveform through time-frequency domain filtering are as follows. The frequency density of the external voltage signal is statistically analyzed and the frequency range is divided to form a noise energy distribution spectrum; Harmonic amplitude, harmonic energy gradient and harmonic contribution ratio of noise energy distribution spectrum are extracted to generate harmonic component indicator group; Identify the high-frequency energy density and energy concentration of the noise energy distribution spectrum to form a transient sensitivity index; The external voltage signal is differentially suppressed based on the harmonic component indicator group and the transient sensitivity index to generate a voltage waveform.

3. The high-precision voltage measurement method for an intelligent voltage and power meter as described in claim 2, characterized in that: The specific steps for extracting the state quantization features of the voltage waveform are as follows. By statistically analyzing the zero-crossing period parameters, peak amplitude parameters, phase drift parameters, and rise slope parameters of the voltage waveform, a four-dimensional behavioral parameter set is constructed. Behavioral coupling mapping and nonlinear scaling are performed on the four-dimensional behavioral parameter set to generate state quantization features.

4. The high-precision voltage measurement method for an intelligent voltage and power meter as described in claim 3, characterized in that: The specific steps for performing state evolution mapping on the state quantification features to generate trend evolution vectors are as follows. The amplitude change increment, frequency drift increment, and phase shift increment of the state quantization characteristics are obtained to form a state evolution sequence; Calculate the behavioral sensitivity component of the state quantification feature and generate a set of behavioral sensitivity coefficients; The state evolution trajectory of the state evolution sequence is derived and trend synthesis is performed with the behavior sensitivity coefficient group to generate a trend evolution vector.

5. The high-precision voltage measurement method for an intelligent voltage and power meter as described in claim 4, characterized in that: The temporal state of the calculated trend evolution vector is used to construct a virtual voltage behavior entity. The specific steps are as follows: Extract the evolution direction of the trend evolution vector, perform direction consistency analysis and progressive intensity calibration, and generate trend-driven state groups; Deconstruct and reorganize the behavioral and temporal hierarchies of trend-driven state groups and state quantification features to generate multi-time state mapping columns; The amplitude, frequency, and phase patterns are extracted from the multi-time state mapping series, and a virtual voltage behavior is constructed through stability dominance analysis and trend persistence analysis.

6. The high-precision voltage measurement method for an intelligent voltage and power meter as described in claim 5, characterized in that: The specific steps for predicting the voltage characteristic change trend of the voltage waveform using a voltage behavior virtual object are as follows: By decomposing the composite amplitude shape, composite frequency shape, and composite phase shape of the voltage behavior virtual object into a time-consistent form, a behavior pattern sequence is generated. The dominant relationship and temporal relationship of behavioral patterns are screened out and cross-behavioral influence coupling is carried out to construct a trend inference matrix; The trend projection matrix is ​​combined with the trend trajectory according to the time dimension to generate the voltage characteristic change trend.

7. The high-precision voltage measurement method for an intelligent voltage and power meter as described in claim 6, characterized in that: The consistency offset is obtained by comparing the trend of voltage characteristic changes with the time sequence of historical voltage records. The specific steps are as follows: Perform characteristic deconstruction mapping on historical voltage records, and parse the historical voltage records into historical amplitude change trajectory, historical frequency change trajectory and historical phase change trajectory in time order to generate historical characteristic trajectory group; The amplitude change trajectory, frequency drift trajectory, and phase shift trajectory are extracted from the voltage characteristic change trend, and compared with the time series differences of the historical characteristic trajectory group to generate the corresponding amplitude difference sequence, frequency difference sequence, and phase difference sequence. The cross-dimensional consistency of the amplitude difference sequence, frequency difference sequence, and phase difference sequence is inferred separately, and then integrated to generate a consistency offset.

8. The high-precision voltage measurement method for an intelligent voltage and power meter as described in claim 7, characterized in that: The method of jointly estimating the range out-of-bounds risk of the voltage waveform by combining the consistency offset and the trend evolution vector, and generating a range prediction indicator, involves the following specific steps. Deconstruct the range sensitivity of the trend evolution vector and generate a range sensitivity component set; By performing joint risk mapping between the consistency offset and the range sensitivity component group, a range out-of-bounds risk assessment vector is generated. The risk levels of the range out-of-range risk assessment vector are divided and the risk levels are aggregated to generate the range prediction indicator.

9. The high-precision voltage measurement method for an intelligent voltage and power meter as described in claim 8, characterized in that: The specific steps for dynamically switching the voltage waveform range based on the range prediction indication to generate a corrected waveform are as follows: The local amplitude envelope of the voltage waveform is inferred and compared with the range prediction indication to generate a range switching control parameter set. Based on the range of the segmented mapped voltage waveform of the range switching control parameter group, a switching waveform segment is generated; Amplitude continuity matching and transition segment interpolation reconstruction are performed on the switched waveform segments to generate the corrected waveform.

10. The high-precision voltage measurement method for an intelligent voltage and power meter as described in claim 9, characterized in that: The process of performing period aggregation and amplitude-phase joint reconstruction on the corrected waveform to output a high-precision measured voltage involves the following steps: Locate the periodic interval of the corrected waveform, and extract the amplitude sequence and phase sequence for adjacent periodic intervals respectively to generate periodic segmentation feature groups; By comparing the steady-state amplitude and phase continuity of the periodic segmented characteristic group, an amplitude-phase consistency coefficient group is formed. Based on the amplitude-phase consistency coefficient set, the corrected waveform is calibrated and reorganized for amplitude and phase continuity reconstruction to generate an amplitude-phase reconstructed waveform. The amplitude and phase reconstruction waveform is subjected to periodic aggregation and multi-cycle averaging fusion to output a high-precision measured voltage.

Images