A high-precision voltage fluctuation real-time monitoring method and system based on an electric energy meter

Abstract

The application relates to the technical field of power system monitoring, and discloses a high-precision voltage fluctuation real-time monitoring method and system based on an electric energy meter. The method comprises the following steps: collecting voltage waveform data and carrying out denoising and zero point detection; calculating adjacent zero point intervals and identifying phase offset events; analyzing frequency components to determine phase change trends and classify offset events; establishing a state correlation relationship; triggering a response signal; adjusting a control priority; verifying waveform stability and generating a recovery log. The application realizes high-precision real-time monitoring and intelligent control of voltage phase offset, and improves the stability of a power system and power supply quality under new energy access.

Description

Technical Field

[0001] This invention relates to the field of power system monitoring technology, and in particular to a high-precision real-time monitoring method and system for voltage fluctuations based on an electricity meter. Background Technology

[0002] With the large-scale integration of intermittent renewable energy sources such as wind and solar power, and the increasing complexity of user-side loads, voltage fluctuations in the power grid are becoming increasingly prominent. In particular, voltage phase shift has become a key factor affecting power supply reliability. Phase shift not only leads to decreased efficiency and shortened lifespan of electrical equipment, but in severe cases, it can also trigger system-level faults such as grid resonance and protection malfunctions.

[0003] Currently, traditional voltage fluctuation monitoring methods mainly rely on amplitude monitoring and simple zero-crossing detection techniques. In a typical existing technical solution, the input AC signal is filtered, and then a comparator converts the filtered signal into a square wave output. This is combined with a deviation voltage correction circuit and a control circuit. The control circuit calculates the phase difference between the comparator output square waves and generates a control signal to drive the deviation voltage correction circuit, correcting the deviation voltage at the comparator base, thereby avoiding phase detection errors caused by the comparator's "dead zone".

[0004] However, existing technical solutions mainly rely on local compensation of hardware circuits, lacking the ability to analyze and track dynamic changes in phase angle in complex power grid environments in real time using software algorithms. Furthermore, they only correct the comparator dead zone and do not consider the impact of power grid noise, harmonic interference, and waveform distortion on the accuracy of zero-crossing detection, resulting in decreased monitoring accuracy under nonlinear loads or new energy access scenarios. In summary, existing technologies result in low accuracy in real-time monitoring of voltage fluctuations. Summary of the Invention

[0005] This invention provides a high-precision real-time voltage fluctuation monitoring method and system based on an electricity meter to solve the problem of low accuracy in real-time voltage fluctuation monitoring caused by existing technologies.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a high-precision real-time monitoring method for voltage fluctuations based on an electricity meter, comprising:

[0007] Collect voltage waveform data, perform noise reduction processing on the voltage waveform data and extract the zero-point position to determine the zero-point detection result;

[0008] Based on the zero-point detection results, the interval time between adjacent zero points is calculated, and based on the deviation between the interval time and the preset standard period, potential phase shift events are identified, and a list of shift events is generated.

[0009] Analyze the frequency components corresponding to the events in the offset event list, determine the phase angle change value and trend, and classify the phase offset events based on the trend to obtain the classified offset type;

[0010] Based on the classified offset type, obtain the associated state parameters and establish the state association relationship between the offset type and the state parameters;

[0011] Based on the state correlation, highly correlated parameter groups are selected, and when the phase offset amplitude of the highly correlated parameter groups exceeds a preset warning threshold, a real-time response signal is triggered to generate a response signal sequence.

[0012] Based on the response signal sequence, the priority of the control operation is adjusted, and an operation suggestion queue is generated;

[0013] Verify the waveform sequence changes after executing the operation suggestion queue. If the interval after the change meets the preset standard period, it is marked as a stable recovery event, and a recovery event log is generated.

[0014] In one optional implementation, the acquisition of voltage waveform data, the denoising processing of the voltage waveform data, the extraction of zero-point positions, and the determination of zero-point detection results include:

[0015] The voltage waveform data is denoised to obtain a denoised waveform sequence;

[0016] Zero-point positions are extracted from the denoised waveform sequence to obtain a set of zero-point positions;

[0017] The similarity between the denoised waveform sequence and the pre-established reference sine wave is calculated, and the precise moment when the waveform crosses the zero axis is determined by combining the zero position set, so as to obtain the zero detection sequence as the zero detection result.

[0018] In one optional implementation, the step of calculating the interval time between adjacent zero points based on the zero-point detection results, and identifying potential phase shift events based on the deviation of the interval time from a preset standard period, and generating a list of shift events, includes:

[0019] Based on the zero-point detection results, the timestamps of adjacent zero points are obtained, and the time difference between each pair of adjacent zero points is calculated to generate a set of zero-point intervals;

[0020] If the interval time in the zero-point interval set deviates from the preset standard period by more than a preset deviation threshold, then the interval time is marked as a potential phase shift event.

[0021] Summarize all the marked potential phase offset events to generate a list of offset events.

[0022] In one optional implementation, the step of analyzing the frequency components corresponding to the events in the offset event list, determining the phase angle change value and trend, and classifying the phase offset events based on the trend to obtain the classified offset type includes:

[0023] Based on the list of offset events, the start time and offset magnitude data of each event are obtained, and the frequency components corresponding to each event are calculated by Fourier transform to generate a set of frequency components.

[0024] Based on the set of frequency components, compare the frequency spectra within the time window before and after the event, calculate the phase difference, and obtain the set of phase angle change values.

[0025] Extract the trend of change from the set of phase angle change values;

[0026] If the degree of matching between the changing trend and the preset typical load switching characteristic pattern exceeds the preset threshold, the corresponding event is classified as an offset caused by load switching, and a preliminary offset type set is obtained.

[0027] Based on the initial set of offset types, the starting time and the offset magnitude data are linked, and the classified offset types are obtained through event tagging linking operations.

[0028] In one optional implementation, the step of obtaining associated state parameters based on the classified offset type and establishing a state association relationship between the offset type and the state parameters includes:

[0029] Based on the classified offset type, the event timestamp and classification tag are extracted. The event timestamp is then time-aligned with the current load data and new energy access data through a time-series matching operation to obtain a preliminary time-series association set.

[0030] Calculate the correlation coefficient between the current load data and the new energy access data and the offset type, filter and retain parameters with correlation coefficients greater than a preset correlation threshold, and form a set of highly correlated parameters;

[0031] The parameters in the highly correlated parameter set are clustered and grouped according to their time-series characteristics to obtain a parameter group set;

[0032] Based on the matching relationship between the parameter group set and the offset type, an association analysis operation is performed between the state parameters and the offset type to obtain the state association relationship.

[0033] In one optional implementation, the step of filtering out highly correlated parameter groups based on the state correlation, and triggering a real-time response signal and generating a response signal sequence when the phase offset amplitude of the highly correlated parameter groups exceeds a preset warning threshold, includes:

[0034] Obtain the phase offset amplitude data and timing features from the state association relationship;

[0035] The phase offset amplitude data is compared with a preset warning threshold. When the warning threshold is exceeded, a real-time response signal is generated, forming a set of real-time response signals.

[0036] After sorting the signals in the real-time response signal set according to their timestamps, they are matched and aligned with the time sequence features to generate the response signal sequence.

[0037] In one optional implementation, adjusting the priority of the control operation and generating an operation suggestion queue based on the response signal sequence includes:

[0038] Based on the response signal sequence, obtain the fluctuation characteristic information in the new energy access data;

[0039] Extract fluctuation records whose amplitude exceeds a preset fluctuation amplitude threshold from the fluctuation feature information to generate a candidate fluctuation event set;

[0040] The time identifiers of the candidate fluctuation event set are matched with the start times of the offset event list. When the time deviation is less than a preset time matching threshold, the time matching is considered successful.

[0041] Based on the severity parameters of each event in the successfully matched set of related events, the priority weight of the control operation is recalculated;

[0042] In one optional implementation, the waveform sequence change after the verification of the operation suggestion queue is marked as a stable recovery event if the interval after the change meets a preset standard period, and a recovery event log is generated, including:

[0043] Collect the real-time voltage waveform after the control operation is performed, and extract the waveform sequence to be analyzed;

[0044] Calculate the similarity between the waveform sequence and the preset reference waveform. When the similarity value exceeds the preset similarity threshold, it is marked as a candidate stable recovery event.

[0045] Calculate the time interval between the candidate stable recovery event and its adjacent events to generate the actual periodic sequence;

[0046] The actual periodic sequence is compared with the standard period. When the periodic deviation is less than a preset tolerance threshold, it is confirmed as a stable recovery event and a recovery event log is generated.

[0047] Secondly, the present invention provides a high-precision real-time voltage fluctuation monitoring system based on an electricity meter, comprising:

[0048] The waveform acquisition and preprocessing module acquires voltage waveform data, performs noise reduction on the voltage waveform data, extracts the zero-point position, and determines the zero-point detection result.

[0049] The phase shift identification module calculates the interval time between adjacent zero points based on the zero point detection results, and identifies potential phase shift events based on the deviation between the interval time and a preset standard period, and generates a list of shift events.

[0050] The event classification and analysis module analyzes the frequency components corresponding to the events in the offset event list, determines the phase angle change value and trend, and classifies the phase offset events based on the trend to obtain the classified offset type.

[0051] The state association establishment module obtains the associated state parameters based on the classified offset type and establishes the state association relationship between the offset type and the state parameters.

[0052] The response signal generation module, based on the state correlation, filters out highly correlated parameter groups, and triggers a real-time response signal to generate a response signal sequence when the phase offset amplitude of the highly correlated parameter groups exceeds the warning threshold.

[0053] The operation queue optimization module adjusts the priority of the control operations based on the response signal sequence and generates an operation suggestion queue.

[0054] The stability verification feedback module verifies the waveform sequence changes after executing the operation suggestion queue. If the interval after the change meets the preset standard period, it is marked as a stable recovery event and a recovery event log is generated.

[0055] Thirdly, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the high-precision real-time voltage fluctuation monitoring method based on an electricity meter as described in any one of the above.

[0056] Fourthly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the high-precision real-time voltage fluctuation monitoring method based on an electricity meter as described above.

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] (1) The present invention uses wavelet transform algorithm to denoise voltage waveform data, effectively preserving the phase characteristics of the original waveform, reducing noise interference, and improving the accuracy of zero-point detection;

[0059] (2) This invention uses a cross-correlation algorithm to compare with a reference sine wave to accurately identify the zero-point crossing time, further improving the accuracy of phase shift detection;

[0060] (3) This invention combines Fourier transform analysis of frequency components and phase angle change trends, which can effectively distinguish between phase shift caused by load switching and faults, and improve the reliability of event classification.

[0061] (4) By establishing state correlation and filtering highly correlated parameter groups, the present invention realizes intelligent identification and response to phase shift events, thereby improving the real-time performance and adaptability of the system. Attached Figure Description

[0062] Figure 1 This is a schematic diagram of a high-precision real-time voltage fluctuation monitoring method based on an electricity meter provided in the first embodiment of the present invention;

[0063] Figure 2 This is a schematic diagram of a high-precision real-time voltage fluctuation monitoring system based on an electricity meter, provided in the second embodiment of the present invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] Reference Figure 1 The first embodiment of the present invention provides a high-precision real-time monitoring method for voltage fluctuations based on an electricity meter, comprising the following steps:

[0066] S101, Collect voltage waveform data, perform noise reduction processing on the voltage waveform data and extract the zero point position to determine the zero point detection result;

[0067] S102, Based on the zero-point detection results, calculate the interval time between adjacent zero points, and based on the deviation between the interval time and the preset standard period, identify potential phase shift events and generate a list of shift events;

[0068] S103, Analyze the frequency components corresponding to the events in the offset event list, determine the phase angle change value and trend, and classify the phase offset events based on the trend to obtain the classified offset type;

[0069] S104, Based on the classified offset type, obtain the associated state parameters and establish the state association relationship between the offset type and the state parameters;

[0070] S105, based on the state correlation, select a high correlation parameter group, and when the phase offset amplitude of the high correlation parameter group exceeds a preset warning threshold, trigger a real-time response signal and generate a response signal sequence;

[0071] S106, Based on the response signal sequence, adjust the priority of the control operation and generate an operation suggestion queue;

[0072] S107, verify the waveform sequence change after executing the operation suggestion queue. If the interval after the change meets the preset standard period, mark it as a stable recovery event and generate a recovery event log.

[0073] In step S101, voltage waveform data is acquired, the voltage waveform data is denoised, and the zero-point position is extracted to determine the zero-point detection result, including:

[0074] Step S1011: Denoise the voltage waveform data to obtain a denoised waveform sequence;

[0075] Step S1012: Extract the zero-point positions from the denoised waveform sequence to obtain a set of zero-point positions;

[0076] Step S1013: Calculate the similarity between the denoised waveform sequence and the pre-established reference sine wave, and determine the precise moment when the waveform crosses the zero axis by combining the zero position set, and obtain the zero detection sequence as the zero detection result.

[0077] In step S1011, the voltage waveform data is denoised to obtain a denoised waveform sequence.

[0078] It should be noted that the grid voltage waveform data is acquired through a data acquisition device (such as a high-precision voltage sensor), with a sampling rate preferably of 10kHz to capture waveform details. It is also worth noting that the metering chip front end of the energy meter needs to sample the voltage through a voltage divider resistor or voltage transformer; this sampling element can be understood as an implementation of the aforementioned high-precision voltage sensor. The original voltage waveform data is denoised using a wavelet transform algorithm, preferably employing the Daubechies wavelet basis (such as db4), as it has good adaptability to non-stationary signals in power systems.

[0079] In one specific implementation, a decomposition of 3 to 5 layers is typically performed, and thresholds are generated using unbiased risk estimation or heuristic thresholding rules. Then, a soft thresholding function is used to process the detail coefficients of each layer. The above parameter settings are suitable for sampling rate environments of 6.4 kHz to 12.8 kHz to ensure that the fundamental frequency and low-order harmonic components of the power frequency are preserved while effectively separating high-frequency noise.

[0080] In step S1012, zero-point positions are extracted from the denoised waveform sequence to obtain a set of zero-point positions.

[0081] It should be noted that a threshold judgment method is used to extract the zero-point position from the first waveform sequence. The preset zero-point judgment threshold range is set based on the system rated voltage (Un) in per-unit or percentage form to accommodate different voltage levels. Preferably, the threshold range is set to -0.002Un to +0.002Un (i.e., ±0.2%Un). For example, for a system with a rated voltage of 220V, the threshold range is approximately -0.44V to +0.44V; for a 10kV system, the threshold range is approximately -20V to +20V. When the voltage value in the sequence falls within this threshold range, it is determined to be a zero-point position.

[0082] The threshold range can be dynamically adjusted based on the device's measurement accuracy, background noise level, and system operating status in the actual application scenario to achieve an optimal balance between detection sensitivity and anti-interference capability. The principle of threshold setting is to filter out false zero-crossing signals caused by noise as much as possible, while ensuring that no true zero-crossing points are missed.

[0083] In step S1013, the similarity between the denoised waveform sequence and the pre-established reference sine wave is calculated, and the precise moment when the waveform crosses the zero axis is determined by combining the zero position set, so as to obtain the zero detection sequence as the zero detection result.

[0084] It should be noted that the reference sine wave data is dynamically generated: Upon system initialization or when a step change in the grid frequency is detected, the reconstruction process of the reference waveform is automatically triggered. This process selects a stable voltage waveform under the current state (e.g., a frequency deviation of less than 0.05Hz for one second), accurately extracts its fundamental frequency, amplitude, and initial phase using Fast Fourier Transform (FFT), and uses these parameters to generate an ideal sine wave synchronized with the actual frequency of the current grid as the reference waveform in real time. A cross-correlation algorithm is then used to calculate the similarity between the first waveform sequence and this reference sine wave.

[0085] In one specific implementation, a sliding window comparison is used to calculate the correlation coefficient. The time point corresponding to the peak of the correlation coefficient is the precise moment when the waveform crosses the zero axis. For example, if the correlation coefficient is maximum at t=0.0102s, this moment is determined as the zero-point crossing moment, and a zero-point detection sequence is generated. This method effectively corrects zero-point offsets caused by noise or equipment deviations, improving the accuracy and reliability of phase detection.

[0086] In step S102, based on the zero-point detection results, the interval time between adjacent zero points is calculated, and based on the deviation of the interval time from a preset standard period, potential phase shift events are identified, and a list of shift events is generated, including:

[0087] Step S1021: Based on the zero-point detection results, obtain the timestamps of adjacent zero points, calculate the time difference between each pair of adjacent zero points, and generate a zero-point interval set;

[0088] Step S1022: If the interval time in the zero-point interval set deviates from the preset standard period by more than a preset deviation threshold, then the interval time is marked as a potential phase shift event.

[0089] Step S1023: Summarize all the marked potential phase shift events to generate the offset event list.

[0090] In step S1021, based on the zero-point detection results, the timestamps of adjacent zero points are obtained, and the time difference between each pair of adjacent zero points is calculated to generate a zero-point interval set.

[0091] It should be noted that, based on the zero-point detection sequence obtained in step S101, the timestamps of adjacent zero points are extracted, the time difference between each pair of consecutive zero points is calculated, and a set of zero-point intervals is generated.

[0092] In one specific implementation, in a power grid with a rated frequency of 50Hz, the standard half-cycle is 10ms. Assuming that the timestamps obtained from the zero-point detection sequence are in seconds: 0.0000, 0.0100, 0.0201, 0.0299, 0.0402, the zero-point interval set is calculated to be: 0.0100, 0.0101, 0.0098, 0.0103 seconds.

[0093] In step S1022, if the interval time in the zero-point interval set deviates from the preset standard period by more than a preset deviation threshold, the interval time is marked as a potential phase shift event.

[0094] It should be noted that the standard period (T_std) described in this invention is a preset value, calculated based on the rated frequency (f_std) of the power grid. The preset standard period is set based on the rated frequency of the power grid; for a 50Hz power grid, the standard half-cycle is 10ms. The preset deviation threshold must be set to ensure that false alarms are not triggered within the normal fluctuation range of the power grid frequency (e.g., within ±0.5Hz as specified in GB / T 15945-2008). Therefore, the deviation threshold should be greater than the period change corresponding to a frequency deviation of ±0.5Hz (approximately ±400μs).

[0095] Taking into account both monitoring accuracy and anti-interference capability, the deviation threshold (ΔT_threshold) is preferably set to 300μs to 600μs (i.e., 0.0003 seconds to 0.0006 seconds). This range can effectively filter out periodic fluctuations caused by normal frequency adjustment, while also sensitively detecting real phase shift events caused by load switching, faults, etc.

[0096] In step S1023, all the marked potential phase shift events are summarized to generate the offset event list.

[0097] It should be noted that all the marked potential phase shift events are summarized to generate a structured list of shift events.

[0098] In one specific implementation, all tagged potential phase shift events are aggregated in chronological order and stored in a unified structured data list or array structure to form a complete offset event list. Each event entry in this list contains the following fields: the timestamp of the event, the measured interval, and the absolute value of the deviation from the standard period, among other key data.

[0099] For example, in a programming implementation, an array or list object can be created, and each identified potential phase shift event can be added as an element to the collection in chronological order. Each event element contains the specific value of the aforementioned fields, thus constructing a structured list of shift events, providing a complete and ordered data foundation for subsequent frequency analysis and event classification.

[0100] In step S103, the frequency components corresponding to the events in the offset event list are analyzed to determine the phase angle change value and trend. Based on the trend, the phase offset events are classified to obtain the classified offset types, including:

[0101] Step S1031: Based on the offset event list, obtain the start time and offset amplitude data of each event, and calculate the frequency component corresponding to each event through Fourier transform to generate a frequency component set;

[0102] Step S1032: Based on the set of frequency components, compare the frequency spectra within the time window before and after the event, calculate the phase difference, and obtain a set of phase angle change values.

[0103] Step S1033: Extract the trend of change from the set of phase angle change values;

[0104] Step S1034: If the matching degree between the change trend and the preset typical load switching characteristic pattern exceeds the preset threshold, the corresponding event is classified as an offset caused by load switching, and a preliminary offset type set is obtained.

[0105] Step S1035: Based on the preliminary offset type set, link the start time and the offset magnitude data, and obtain the classified offset type through event tag linking operation.

[0106] In step S1031, the start time and offset amplitude data of each event are obtained according to the offset event list, and the frequency component corresponding to each event is calculated by Fourier transform to generate a frequency component set.

[0107] It should be noted that this step extracts the spectral characteristics of the offset events through frequency domain analysis. In practice, the start time and offset amplitude data of the events are first obtained from the offset event list. Then, voltage waveform data within a specific time window before and after the event are selected, and the time-domain signal is converted into a frequency-domain representation using a fast Fourier transform algorithm to extract the amplitude and phase information of the main frequency components.

[0108] In one specific implementation, taking a substation monitoring system as an example, for a phase shift event with a start time of 10.05 seconds and an offset amplitude of 0.003 seconds, voltage waveform data within a 0.1-second time window (9.95 seconds to 10.15 seconds) before and after the event are selected. Through 1024-point Fast Fourier Transform processing at a sampling rate of 10kHz, the frequency component data corresponding to the event are obtained: at the 50Hz fundamental frequency, the phase before the event is 178.2 degrees and the amplitude is 219.8V; after the event, the phase is 183.5 degrees and the amplitude is 218.3V. Simultaneously, the amplitude and phase information of the second harmonic (100Hz) and the third harmonic (150Hz) are extracted to form a complete set of frequency components.

[0109] In step S1032, based on the set of frequency components, the frequency spectrum within the time window before and after the event occurs is compared, and the phase difference is calculated to obtain a set of phase angle change values.

[0110] It should be noted that this step quantifies the phase angle change by comparing the spectral characteristics before and after the event. In practice, based on the data in the frequency component set, the phase difference of each frequency component before and after the event is calculated, with a focus on the phase change of the fundamental frequency, while also considering the influence of harmonic components.

[0111] In one specific implementation, taking the aforementioned offset event as an example, the phase change of the 50Hz fundamental frequency is calculated: the phase before the event is 178.2 degrees, and the phase after the event is 183.5 degrees, with a phase angle change of +5.3 degrees. Simultaneously, the phase change of the second harmonic is calculated as +2.1 degrees, and the phase change of the third harmonic is calculated as -1.8 degrees. These phase angle change values ​​are categorized and stored according to event number and frequency component, forming a set of phase angle change values.

[0112] In step S1033, the trend of change is extracted from the set of phase angle change values.

[0113] In one specific implementation, the phase angle changes of the 50Hz fundamental frequency are arranged chronologically based on a set of phase angle change values ​​from multiple offset events: Event 1 (10.05 seconds) +5.3 degrees, Event 2 (15.30 seconds) +7.8 degrees, and Event 3 (21.45 seconds) +9.2 degrees. Linear regression analysis reveals that the phase angle changes exhibit an increasing trend, with a rate of change of approximately 0.5 degrees / minute, and the change pattern is a positive step jump.

[0114] In step S1034, if the matching degree between the change trend and the preset typical load switching characteristic pattern exceeds a preset threshold, the corresponding event is classified as an offset caused by load switching, and a preliminary offset type set is obtained.

[0115] It should be noted that the typical load switching feature pattern library is constructed based on the accumulation and learning of historical data. In the initial stage of system deployment, a basic pattern library is pre-configured, consisting of known event data under simulation and typical operating conditions (such as high-power motor start-up and shutdown, transformer switching, capacitor bank switching, etc.). After the system is running, manually confirmed or verified, clearly categorized offset events and their corresponding frequency components and phase angle change trends are continuously and automatically stored in the historical database. Historical features are periodically mined using clustering algorithms (such as K-Means) to generate new feature patterns, which are used to dynamically expand and optimize the initial pattern library. The confidence threshold for pattern matching is set at 85%, based on the 95th percentile of the matching degree data of all verified correctly matched events during the initial training phase of the system (e.g., three consecutive months).

[0116] In one specific implementation, the changing trends of the aforementioned three events are matched with typical load switching characteristic patterns, achieving a matching degree of 92%, exceeding the 85% confidence threshold. The phase jump amplitude and duration highly match the load switching characteristics; therefore, these three events are classified as offsets caused by load switching, forming a preliminary offset type set containing information such as event number, classification result, and matching degree.

[0117] In step S1035, the starting time and the offset magnitude data are linked according to the preliminary offset type set, and the classified offset type is obtained through the event tag linking operation.

[0118] In one specific implementation, the relevant data of three load switching events are linked: Event 1 (start time 10.05 seconds, offset 0.003 seconds, category: load switching); Event 2 (start time 15.30 seconds, offset 0.005 seconds, category: load switching); Event 3 (start time 21.45 seconds, offset 0.007 seconds, category: load switching). Data linking is achieved by assigning a unique event identifier to each event and establishing an event feature index table. Specifically, the system creates a structured data table where each record contains fields such as event identifier, timestamp, offset, event type, and matching degree. Furthermore, the event identifier is used to establish associations with multi-source data such as original waveform data, frequency analysis results, and correlation parameters. This indexing and linking method based on unique identifiers forms a complete record of category offset types, providing multi-dimensional data support for subsequent correlation analysis and status assessment.

[0119] In step S104, based on the classified offset type, the associated state parameters are obtained, and a state association relationship is established between the offset type and the state parameters, including:

[0120] Step S1041: Extract event timestamps and classification tags according to the classified offset types, and align the event timestamps with current load data and new energy access data through time-series matching operations to obtain a preliminary time-series association set;

[0121] Step S1042: Calculate the correlation coefficient between the current load data and the new energy access data and the offset type, filter and retain parameters with correlation coefficients greater than a preset correlation threshold, and form a set of highly correlated parameters;

[0122] Step S1043: Cluster the parameters in the highly correlated parameter set according to their time-series characteristics to obtain a parameter group set;

[0123] Step S1044: Based on the matching relationship between the parameter grouping set and the offset type, perform an association analysis operation between the state parameters and the offset type to obtain the state association relationship.

[0124] In step S1041, the event timestamp and classification tag are extracted according to the classified offset type. The event timestamp is then time-aligned with the current load data and new energy access data through a time-series matching operation to obtain a preliminary time-series association set.

[0125] It should be noted that this step establishes the correlation between offset events and power grid operating parameters through time-series matching operations. In specific implementation, firstly, event timestamps and classification tags are extracted from the classified offset types. Then, a cubic spline interpolation algorithm is used to unify monitoring data with different sampling rates to the same time base, setting the time alignment tolerance threshold to 0.01 seconds to ensure the time synchronization accuracy of the data.

[0126] In one specific implementation, taking a substation monitoring system as an example, the timestamps of three load switching events were extracted from the classification results: 10.05 seconds, 15.30 seconds, and 21.45 seconds. Through time-series matching, the corresponding operating parameters for each time point were obtained: at 10.05 seconds, the current load was 152.3 amps, and the photovoltaic power station's connected power was 85.6 kW; at 15.30 seconds, the current load was 158.7 amps, and the photovoltaic power station's connected power was 82.1 kW; at 21.45 seconds, the current load was 165.2 amps, and the photovoltaic power station's connected power was 78.9 kW. These matching results were then organized according to time sequence to form a preliminary time-series association set containing multi-dimensional data such as timestamps, event types, and electrical parameters.

[0127] In step S1042, the correlation coefficient between the current load data, the new energy access data and the offset type is calculated, and parameters with correlation coefficients greater than a preset correlation threshold are filtered and retained to form a set of highly correlated parameters.

[0128] It should be noted that this step uses correlation coefficient analysis to screen operational parameters highly correlated with the offset event. The Pearson correlation coefficient is used, with a preset correlation threshold typically set to 0.5-0.7. This threshold is determined based on a statistical significance level of α=0.01 to ensure that the screened parameters have a statistically significant correlation. This significance level is verified through the following process: First, a null hypothesis of "the overall correlation coefficient is 0" is established. Then, the Pearson correlation coefficient r and its corresponding p-value are calculated based on the sample data. When the calculated p-value is less than 0.01, the null hypothesis can be rejected at a 99% confidence level, indicating that the correlation is statistically significant. This significance test ensures that the correlation of the screened parameters is not caused by random factors but is statistically significant.

[0129] In one specific implementation, based on 20 consecutive sampling points in the initial time-series correlation set, the correlation coefficient between the current load value and the load switching event was calculated to be 0.92, and the correlation coefficient between the renewable energy access power and the load switching event was -0.87. Since the absolute values ​​of the correlation coefficients for both parameters are greater than the preset threshold of 0.8, these two parameters are retained to form a highly correlated parameter set containing information such as parameter name, correlation coefficient, and significance level.

[0130] In step S1043, the parameters in the highly correlated parameter set are clustered and grouped according to their time-series characteristics to obtain a parameter group set.

[0131] It should be noted that this step groups highly correlated parameters according to their time-series characteristics through cluster analysis. The K-means clustering algorithm is used, with a cluster size of 2, and the optimal cluster size is determined based on the elbow rule. The feature vectors include time-series data of current load values ​​and renewable energy access power, and Euclidean distance is used as the similarity measure.

[0132] In one specific implementation, after iterative calculation, the parameters are divided into two distinct clusters: Cluster 1 contains samples of high current loads (150-170 amps) and low renewable energy access power (75-85 kW), and all load switching events are located in this cluster; Cluster 2 contains samples of low current loads (130-150 amps) and high renewable energy access power (85-95 kW), and does not contain any offset events. This ultimately forms a parameter grouping set with clearly distinguishable characteristics.

[0133] In step S1044, based on the matching relationship between the parameter group set and the offset type, an association analysis operation is performed between the state parameters and the offset type to obtain the state association relationship.

[0134] It should be noted that this step reveals the intrinsic relationship between state parameters and offset types by establishing a quantitative correlation model. A logistic regression algorithm is used to build the quantitative correlation model. The dataset must contain a sufficient number of classified offset event samples, with a total sample size no less than 10 times the number of independent variables in the model, and at least 100 samples for each offset type. Each sample includes the state parameters aligned at the time of the event and their corresponding offset type labels. After obtaining a historical dataset that meets the sample size requirements, the samples are divided into training and test sets in a 7:3 ratio. The SMOTE technique is used to address class imbalance in the training set, and L2 regularization is introduced to prevent overfitting. Using the previously selected highly correlated state parameters as independent variables and the offset type as the dependent variable, the model parameters are solved using maximum likelihood estimation to obtain the weight coefficients and probability calculation formulas for each parameter. After the model training is complete, the output includes the weight coefficients, probability calculation formulas, and decision boundary state correlation relationships, providing a reliable quantitative basis for real-time monitoring systems.

[0135] In one specific implementation, a logistic regression model is established based on the clustering results. Analysis shows that when the current load is greater than 150 amperes and the renewable energy access power is less than 85 kilowatts, the probability of a load switching event occurs reaches 89.7%. Validated on a test set, the model achieved an accuracy of 91.5%, a recall of 88.7%, and an F1 score of 90.1%. The final result includes state correlations containing weight coefficients, probability calculation formulas, and decision boundaries, providing a reliable quantitative basis for real-time monitoring systems.

[0136] In step S105, based on the state correlation, a highly correlated parameter group is selected, and when the phase offset amplitude of the highly correlated parameter group exceeds a preset warning threshold, a real-time response signal is triggered, generating a response signal sequence, including:

[0137] Step S1051: Obtain the phase offset amplitude data and timing features in the state association relationship;

[0138] Step S1052: Compare the phase offset amplitude data with a preset warning threshold. When the warning threshold is exceeded, a real-time response signal is generated to form a set of real-time response signals.

[0139] Step S1053: After sorting the signals in the real-time response signal set according to the timestamp, match and align them with the time sequence features to generate the response signal sequence.

[0140] In step S1051, the phase offset amplitude data and timing features in the state association relationship are obtained.

[0141] It should be noted that this step extracts key phase offset feature parameters from the established state association relationships. In practice, the state association relationship database is accessed through a database query interface to extract the stored phase offset amplitude data and its corresponding time series features, including time series attributes such as the time of offset occurrence, duration, and rate of change.

[0142] In one specific implementation, taking a power distribution network monitoring system as an example, detailed data for three phase offset events are extracted from the state correlation database: Event A has a phase offset amplitude of 0.05 degrees, an occurrence duration of 2.5 seconds, and a change rate of 0.02 degrees / second; Event B has a phase offset amplitude of 0.08 degrees, an occurrence duration of 3.2 seconds, and a change rate of 0.025 degrees / second; Event C has a phase offset amplitude of 0.12 degrees, an occurrence duration of 4.1 seconds, and a change rate of 0.029 degrees / second. Simultaneously, the timestamps and durations corresponding to each event are extracted as temporal features.

[0143] In step S1052, the phase offset amplitude data is compared with a preset warning threshold. When the warning threshold is exceeded, a real-time response signal is generated, forming a set of real-time response signals.

[0144] It should be noted that this step uses a threshold comparison mechanism to achieve real-time alarm for abnormal events. The preset warning thresholds are determined based on power grid safety operation standards and historical statistical data, and a multi-level threshold system is set up, including a warning threshold (0.05 degrees), an alarm threshold (0.08 degrees), and an emergency threshold (0.12 degrees), each corresponding to a different response level.

[0145] In one specific implementation, the extracted phase offset amplitude data is compared with multi-level warning thresholds: if the amplitude of event A is 0.05 degrees, it reaches the warning threshold, generating a yellow warning signal; if the amplitude of event B is 0.08 degrees, it reaches the alarm threshold, generating an orange alarm signal; if the amplitude of event C is 0.12 degrees, it exceeds the emergency threshold, generating a red emergency signal. These real-time response signals contain information such as event number, alarm level, and timestamp, forming a set of real-time response signals.

[0146] In step S1053, the signals in the real-time response signal set are sorted by timestamp and then matched and aligned with the time sequence features to generate the response signal sequence.

[0147] It should be noted that this step generates a normalized response signal sequence through temporal sorting and feature matching. A timestamp-based fast sorting algorithm is used to arrange the response signals temporally, and then a time window matching algorithm is used to associate and align the response signals with the corresponding temporal features.

[0148] In one specific implementation, the three response signals are sorted by timestamp: Event A (10:05:02), Event B (10:05:05), and Event C (10:05:08). They are then matched and aligned with time-series features: Event A matches a 2.5-second duration feature, Event B matches a 3.2-second duration feature, and Event C matches a 4.1-second duration feature. This ultimately generates a response signal sequence containing time-series information, alarm level, and duration characteristics, providing complete input for subsequent control operations.

[0149] In step S106, the priority of the control operation is adjusted according to the response signal sequence, and an operation suggestion queue is generated, including:

[0150] Step S1061: Obtain fluctuation characteristic information in the new energy access data based on the response signal sequence;

[0151] Step S1062: Extract fluctuation records whose amplitude exceeds a preset fluctuation amplitude threshold from the fluctuation feature information, and generate a candidate fluctuation event set;

[0152] Step S1063: The time identifier of the candidate fluctuation event set is matched with the start time of the offset event list. When the time deviation is less than the preset time matching threshold, the time matching is determined to be successful.

[0153] Step S1064: Based on the severity parameters of each event in the successfully matched set of related events, recalculate the priority weight of the control operation;

[0154] Step S1065: Sort the control operations according to the priority weights to generate the operation suggestion queue.

[0155] In step S1061, fluctuation characteristic information in the new energy access data is obtained based on the response signal sequence.

[0156] It should be noted that this step involves analyzing the output characteristics of new energy power generation equipment to extract fluctuation features related to voltage phase shift. Based on the time signature of the response signal sequence, the corresponding time period's new energy output data is obtained from the SCADA system, and wavelet packet decomposition algorithm is used to extract fluctuation feature parameters, including fluctuation amplitude, rate of change, and duration.

[0157] In one specific implementation, for a wind farm access node, the output power data of the wind turbines within a specific time window is obtained based on the event time window recorded in the response signal sequence. Fluctuation components in the 0.5-5Hz frequency band are extracted through three-layer wavelet packet decomposition. The calculated fluctuation amplitude of a single 2.5MW unit is 12.5% ​​of its rated power, with a rate of change of 85kW / s and a duration of 3.2 seconds. These characteristic parameters reflect the correlation between the fluctuation of renewable energy output and voltage phase shift.

[0158] In step S1062, fluctuation records with amplitudes exceeding a preset fluctuation amplitude threshold are extracted from the fluctuation feature information to generate a candidate fluctuation event set.

[0159] It should be noted that this step identifies significant renewable energy fluctuation events through a multi-level threshold screening mechanism. The preset fluctuation amplitude thresholds are set based on the technical characteristics of renewable energy power plants and the requirements for safe grid operation, combined with the rated capacity of specific power plants (e.g., 50MW) and the grid inertia constant, and comprehensively consider industry standards (e.g., GB / T 19963-2011). Ultimately, the first-level fluctuation amplitude threshold is set at 8% of the rated output, and the second-level threshold is set at 12%. This dual-threshold setting has been verified through real-time digital simulation, enabling tiered early warning of fluctuation events while balancing monitoring sensitivity and system anti-interference capabilities.

[0160] In one specific implementation, a first-level threshold is set to 8% of the rated output, and a second-level threshold is set to 12%. Three excessive fluctuation events were detected: Event 1 had a fluctuation amplitude of 9.2% and a duration of 2.8 seconds; Event 2 had a fluctuation amplitude of 13.5% and a duration of 3.5 seconds; Event 3 had a fluctuation amplitude of 15.8% and a duration of 4.2 seconds. These events were classified according to severity, generating a candidate fluctuation event set that includes timestamps, fluctuation characteristics, and hazard levels.

[0161] In step S1063, the time identifier of the candidate fluctuation event set is matched with the start time of the offset event list. When the time deviation is less than the preset time matching threshold, the time matching is determined to be successful.

[0162] It should be noted that this step employs a dynamic time warping algorithm for time series matching analysis to address the time series alignment issue between data at different sampling rates. The preset time matching threshold is determined based on the following technical requirements: First, the time scale of the power system transient process is considered (typical fault clearing time 80-120ms, oscillation period 0.1-2s), combined with the accuracy of the measurement system (synchronous phasor measurement unit (PMU) accuracy ±1% total error, data frame period 10-20ms), using the formula... Calculate the baseline value (where For the characteristic time of a transient process, (This is the minimum resolution of the measurement system). The final time matching threshold was determined to be 20-100ms, which covers more than 90% of the transient process timing characteristics and meets the measurement accuracy requirements.

[0163] In one specific implementation, the DTW algorithm is used to calculate the time series similarity between candidate fluctuation events and phase shift events. A matching threshold of 0.8 and a time tolerance of 0.1 seconds are set. The matching results show that the similarity between fluctuation event 1 and phase shift event A is 0.92, with a time deviation of 0.3 seconds; the similarity between event 2 and phase shift event B is 0.87, with a time deviation of 0.4 seconds; and the similarity between event 3 and phase shift event C is 0.94, with a time deviation of 0.2 seconds. Since the similarity of all matching pairs exceeds the threshold and the time deviation is less than the tolerance, the time series matching is considered successful.

[0164] In step S1064, the priority weight of the control operation is recalculated based on the severity parameters of each event in the successfully matched set of associated events.

[0165] In one specific implementation, the comprehensive weight of each event is calculated using the analytic hierarchy process (AHP) based on the severity parameters of the matched events (including offset (0.05-0.12 degrees), fluctuation amplitude (9.2-15.8%), and duration (2.8-4.2 seconds)). The severity parameters are obtained through multi-source data fusion, where the offset originates from the real-time monitoring value of the phase angle measurement unit (PMU), and the fluctuation amplitude is calculated according to the formula for the volatility rate of new energy power. The duration is determined by the difference between the start and end timestamps of the event. The specific process of the analytic hierarchy process includes: first, establishing a judgment matrix based on the degree of influence of offset, fluctuation amplitude, and duration on system stability; then, calculating the weights of each indicator using the eigenvalue method (in this example, the weights of offset, fluctuation amplitude, and duration are 0.4, 0.35, and 0.25, respectively); next, performing a consistency check to ensure that the consistency ratio (CR) of the judgment matrix is ​​less than 0.1; finally, normalizing the parameter values ​​of each event and weighting them with the indicator weights, the weights of events A, B, and C are 0.25, 0.35, and 0.40, respectively.

[0166] In step S1065, the control operations are sorted according to the priority weights to generate the operation suggestion queue.

[0167] In one implementation, based on the weight allocation results, control operations such as voltage adjustment (including adjusting transformer taps and switching capacitors and reactors) and power control (such as adjusting the active / reactive output of renewable energy power plants) are prioritized to generate an optimized operation queue for prioritizing high-weight events, thereby ensuring stable system operation. It should be noted that the optimized operation queue is a priority operation suggestion queue for grid control, i.e., the operation suggestion queue, which is uploaded to the grid dispatching system as a reference for control decisions, but the final execution right still rests with the grid dispatching system.

[0168] In step S107, the waveform sequence change after executing the operation suggestion queue is verified. If the interval after the change meets the preset standard period, it is marked as a stable recovery event, and a recovery event log is generated, including:

[0169] Step S1071: Acquire the real-time voltage waveform after the control operation is performed, and extract the waveform sequence to be analyzed;

[0170] Step S1072: Calculate the similarity between the waveform sequence and the preset reference waveform. When the similarity value exceeds the preset similarity threshold, mark it as a candidate stable recovery event.

[0171] Step S1073: Calculate the time interval between the candidate stable recovery event and the adjacent event to generate the actual periodic sequence;

[0172] Step S1074: Compare the actual periodic sequence with the standard periodic sequence. When the periodic deviation is less than the preset tolerance threshold, it is confirmed as a stable recovery event and a recovery event log is generated.

[0173] In step S1071, the real-time voltage waveform after the control operation is performed is acquired, and the waveform sequence to be analyzed is extracted.

[0174] It should be noted that this step acquires the regulated voltage waveform data using high-precision sampling equipment and extracts the waveform sequence within the key analysis period. Anti-aliasing filters and synchronous sampling techniques are employed to ensure the accuracy and timing consistency of waveform acquisition. According to the Nyquist sampling theorem, a sampling rate higher than 4kHz is required for accurate analysis of the 50Hz fundamental and 40th harmonic (2kHz) components. Simultaneously, referring to the requirements for synchronous phasor measurement in the IEEE C37.118.1 standard and considering engineering margins, a sampling rate of no less than 10kHz is ultimately determined to ensure that 200 data points are acquired per power frequency cycle, meeting the technical requirement of phase measurement accuracy better than ±0.01°.

[0175] In one specific implementation, a power quality monitoring device installed at the 10kV busbar of a substation acquires real-time voltage waveforms at a sampling rate of 12.8kHz after a voltage adjustment operation. Waveform data within a 2-second time window following the completion of the adjustment operation is extracted, containing 10 power frequency cycle waveform sequences with an amplitude range of -325V to +325V and 25,600 sampling points, providing a data foundation for subsequent stability analysis.

[0176] In step S1072, the similarity between the waveform sequence and the preset reference waveform is calculated. When the similarity value exceeds the preset similarity threshold, it is marked as a candidate stable recovery event.

[0177] It should be noted that this step uses a dynamic time warping algorithm to calculate the similarity between the actual waveform and the ideal reference waveform. This is achieved by non-linearly aligning the two time series on the time axis and then normalizing the calculated cumulative path distance. The reference waveform is a standard 50Hz sine wave. The similarity threshold is determined based on statistical analysis of historical stable operating data, typically set to 0.85. In historical data, the similarity of normally operating waveforms usually falls between 0.9 and 1.0, while the similarity of abnormal waveforms is generally below 0.7. 0.85 serves as an intermediate value, effectively distinguishing between normal and abnormal waveforms to ensure accurate identification.

[0178] In one specific implementation, the acquired waveform sequence is compared with a standard 50Hz reference waveform using DTW similarity calculation. The calculated similarity values ​​for three time periods are 0.92, 0.88, and 0.95, all exceeding the similarity threshold of 0.85. These time periods are marked as candidate stable recovery events, and their start time, duration, and similarity value are recorded to form a candidate event list.

[0179] In step S1073, the time interval between the candidate stable recovery event and the adjacent event is calculated to generate the actual periodic sequence.

[0180] It should be noted that this step calculates the periodic characteristics of candidate events using a zero-crossing detection algorithm. An improved zero-point interpolation method is employed to enhance the accuracy of period measurement, and the period tolerance threshold is set to ±0.5% of the standard period to meet the requirements of power grid frequency stability.

[0181] In one specific implementation, zero-crossing analysis is performed on the marked candidate stable recovery events to calculate the time interval between adjacent zero-crossings. The time intervals of 10 consecutive cycles were measured as follows: 20.01ms, 19.98ms, 20.03ms, 19.99ms, 20.02ms, 20.00ms, 19.97ms, 20.04ms, 19.98ms, and 20.01ms, generating an actual cycle sequence. All cycle values ​​are within ±0.5% of the 50Hz standard period of 20ms.

[0182] In step S1074, the actual periodic sequence is compared with the standard period. When the periodic deviation is less than a preset tolerance threshold, it is confirmed as a stable recovery event and a recovery event log is generated.

[0183] It should be noted that this step uses statistical process control methods to evaluate the stability of the periodic sequence. The preset tolerance threshold is determined based on the power grid frequency deviation standard and the accuracy of the measurement system, and the control limits are set using the 3σ principle to ensure the reliability of the judgment.

[0184] In one specific implementation, the statistical characteristics of the actual periodic sequence are calculated: average value 20.003 ms, standard deviation 0.023 ms. The deviation of all periodic data from the standard period of 20 ms is less than the tolerance threshold of 0.05 ms. This is confirmed as a stable recovery event, and a recovery event log is generated, containing the event number, recovery time, duration, periodic deviation statistics, and waveform quality, providing a basis for system performance evaluation.

[0185] In summary, this invention discloses a high-precision real-time monitoring method and system for voltage fluctuations based on an energy meter. The method includes: acquiring voltage waveform data and performing denoising and zero-point detection; calculating the interval between adjacent zero points and identifying phase shift events; analyzing frequency components to determine the phase change trend and classifying shift events; establishing state correlation relationships; triggering response signals; adjusting control priorities; verifying waveform stability and generating recovery logs. This invention achieves high-precision real-time monitoring and intelligent control of voltage phase shifts through wavelet transform denoising, cross-correlation zero-point detection, and Fourier transform frequency domain analysis, effectively improving the stability and power supply quality of the power system under renewable energy integration.

[0186] Reference Figure 2 The second embodiment of the present invention provides a high-precision real-time voltage fluctuation monitoring system based on an electricity meter, comprising:

[0187] The waveform acquisition and preprocessing module acquires voltage waveform data, performs noise reduction on the voltage waveform data, extracts the zero-point position, and determines the zero-point detection result.

[0188] The phase shift identification module calculates the interval time between adjacent zero points based on the zero point detection results, and identifies potential phase shift events based on the deviation between the interval time and a preset standard period, and generates a list of shift events.

[0189] The event classification and analysis module analyzes the frequency components corresponding to the events in the offset event list, determines the phase angle change value and trend, and classifies the phase offset events based on the trend to obtain the classified offset type.

[0190] The state association establishment module obtains the associated state parameters based on the classified offset type and establishes the state association relationship between the offset type and the state parameters.

[0191] The response signal generation module, based on the state correlation, filters out highly correlated parameter groups, and triggers a real-time response signal to generate a response signal sequence when the phase offset amplitude of the highly correlated parameter groups exceeds the warning threshold.

[0192] The operation queue optimization module adjusts the priority of the control operations based on the response signal sequence and generates an operation suggestion queue.

[0193] The stability verification feedback module verifies the waveform sequence changes after executing the operation suggestion queue. If the interval after the change meets the preset standard period, it is marked as a stable recovery event and a recovery event log is generated.

[0194] It should be noted that the high-precision voltage fluctuation real-time monitoring system based on an electricity meter provided in this embodiment of the invention is used to execute all the process steps of the high-precision voltage fluctuation real-time monitoring method based on an electricity meter in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0195] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a voltage fluctuation monitoring program. When the processor executes the computer program, it implements the steps in the various method embodiments described above, for example... Figure 1 The step S101 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module in the above system embodiments, such as the waveform acquisition and preprocessing module.

[0196] For example, the computer program can be divided into multiple modules, which are stored in the memory and executed by the processor to complete the present invention. The multiple modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the electronic device.

[0197] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0198] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0199] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0200] If the modules integrated into the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0201] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0202] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A high-precision real-time monitoring method for voltage fluctuations based on an electricity meter, characterized in that, include: Collect voltage waveform data, perform noise reduction processing on the voltage waveform data and extract the zero-point position to determine the zero-point detection result; Based on the zero-point detection results, the interval time between adjacent zero points is calculated, and based on the deviation between the interval time and the preset standard period, potential phase shift events are identified, and a list of shift events is generated. Analyze the frequency components corresponding to the events in the offset event list, determine the phase angle change value and trend, and classify the phase offset events based on the trend to obtain the classified offset type; Based on the classified offset type, obtain the associated state parameters and establish the state association relationship between the offset type and the state parameters; Based on the state correlation, highly correlated parameter groups are selected, and when the phase offset amplitude of the highly correlated parameter groups exceeds a preset warning threshold, a real-time response signal is triggered to generate a response signal sequence. Based on the response signal sequence, the priority of the control operation is adjusted, and an operation suggestion queue is generated; Verify the waveform sequence changes after executing the operation suggestion queue. If the interval after the change meets the preset standard period, mark it as a stable recovery event and generate a recovery event log. The process of analyzing the frequency components corresponding to events in the offset event list, determining the phase angle change value and trend, and classifying the phase offset events based on the trend to obtain the classified offset types includes: obtaining the start time and offset amplitude data of each event from the offset event list, and calculating the frequency components corresponding to each event through Fourier transform to generate a frequency component set; comparing the frequency spectrum within the time window before and after the event occurrence based on the frequency component set, and calculating the phase difference to obtain a phase angle change value set; extracting the trend from the phase angle change value set; if the matching degree of the trend with the preset typical load switching characteristic mode exceeds a preset threshold, then classifying the corresponding event as an offset caused by load switching to obtain a preliminary offset type set; linking the start time and offset amplitude data based on the preliminary offset type set, and obtaining the classified offset types through event tag linking operations; The step of obtaining associated state parameters based on the classified offset types and establishing state association relationships between offset types and state parameters includes: extracting event timestamps and classification tags based on the classified offset types; aligning the event timestamps with current load data and renewable energy access data through time-series matching operations to obtain a preliminary time-series association set; calculating the correlation coefficients between the current load data and renewable energy access data and the offset types; filtering and retaining parameters with correlation coefficients greater than a preset correlation threshold to form a high-correlation parameter set; clustering the parameters in the high-correlation parameter set according to time-series characteristics to obtain a parameter grouping set; and performing association analysis operations between state parameters and offset types based on the matching relationship between the parameter grouping set and the offset types to obtain the state association relationship. The verification process, which involves checking the waveform sequence changes after the operation suggestion queue is executed, and marking it as a stable recovery event if the interval after the change meets a preset standard period, and generating a recovery event log, includes: acquiring the real-time voltage waveform after the control operation is executed and extracting the waveform sequence to be analyzed; calculating the similarity between the waveform sequence and a preset reference waveform, and marking it as a candidate stable recovery event when the similarity value exceeds a preset similarity threshold; calculating the time interval between the candidate stable recovery event and adjacent events to generate an actual period sequence; comparing the actual period sequence with the standard period, and confirming it as a stable recovery event and generating a recovery event log when the period deviation is less than a preset tolerance threshold.

2. The high-precision real-time voltage fluctuation monitoring method based on an electricity meter according to claim 1, characterized in that, The process of acquiring voltage waveform data, denoising the voltage waveform data, extracting the zero-point position, and determining the zero-point detection result includes: The voltage waveform data is denoised to obtain a denoised waveform sequence; Zero-point positions are extracted from the denoised waveform sequence to obtain a set of zero-point positions; The similarity between the denoised waveform sequence and the pre-established reference sine wave is calculated, and the precise moment when the waveform crosses the zero axis is determined by combining the zero position set, so as to obtain the zero detection sequence as the zero detection result.

3. The high-precision real-time voltage fluctuation monitoring method based on an electricity meter according to claim 1, characterized in that, The step involves calculating the interval time between adjacent zero points based on the zero-point detection results, and identifying potential phase shift events based on the deviation between the interval time and a preset standard period, generating a list of shift events, including: Based on the zero-point detection results, the timestamps of adjacent zero points are obtained, and the time difference between each pair of adjacent zero points is calculated to generate a set of zero-point intervals; If the interval time in the zero-point interval set deviates from the preset standard period by more than a preset deviation threshold, then the interval time is marked as a potential phase shift event. Summarize all the marked potential phase offset events to generate a list of offset events.

4. The high-precision real-time voltage fluctuation monitoring method based on an electricity meter according to claim 1, characterized in that, Based on the state correlation, a highly correlated parameter group is selected, and when the phase shift amplitude of the highly correlated parameter group exceeds a preset warning threshold, a real-time response signal is triggered, generating a response signal sequence, including: Obtain the phase offset amplitude data and timing features from the state association relationship; The phase offset amplitude data is compared with a preset warning threshold. When the warning threshold is exceeded, a real-time response signal is generated, forming a set of real-time response signals. After sorting the signals in the real-time response signal set according to their timestamps, they are matched and aligned with the time sequence features to generate the response signal sequence.

5. The high-precision real-time voltage fluctuation monitoring method based on an electricity meter according to claim 1, characterized in that, The step of adjusting the priority of the control operation and generating an operation suggestion queue based on the response signal sequence includes: Based on the response signal sequence, obtain the fluctuation characteristic information in the new energy access data; Extract fluctuation records whose amplitude exceeds a preset fluctuation amplitude threshold from the fluctuation feature information to generate a candidate fluctuation event set; The time identifiers of the candidate fluctuation event set are matched with the start times of the offset event list. When the time deviation is less than a preset time matching threshold, the time matching is considered successful. Based on the severity parameters of each event in the successfully matched set of related events, the priority weight of the control operation is recalculated; The control operations are sorted according to the priority weights to generate the operation suggestion queue.

6. A high-precision real-time voltage fluctuation monitoring system based on an electricity meter, characterized in that, For implementing the method as described in any one of claims 1-5, comprising: The waveform acquisition and preprocessing module acquires voltage waveform data, performs noise reduction on the voltage waveform data, extracts the zero-point position, and determines the zero-point detection result. The phase shift identification module calculates the interval time between adjacent zero points based on the zero point detection results, and identifies potential phase shift events based on the deviation between the interval time and a preset standard period, and generates a list of shift events. The event classification and analysis module analyzes the frequency components corresponding to the events in the offset event list, determines the phase angle change value and trend, and classifies the phase offset events based on the trend to obtain the classified offset type. The state association establishment module obtains the associated state parameters based on the classified offset type and establishes the state association relationship between the offset type and the state parameters. The response signal generation module, based on the state correlation, filters out highly correlated parameter groups, and triggers a real-time response signal to generate a response signal sequence when the phase offset amplitude of the highly correlated parameter groups exceeds the warning threshold. The operation queue optimization module adjusts the priority of the control operations based on the response signal sequence and generates an operation suggestion queue. The stability verification feedback module verifies the waveform sequence changes after executing the operation suggestion queue. If the interval after the change meets the preset standard period, it is marked as a stable recovery event and a recovery event log is generated.

Images