Wind-solar grid-connected voltage disturbance evaluation method, system, equipment and medium

Abstract

The invention belongs to the technical field of voltage disturbance monitoring, and particularly relates to a wind-solar grid-connected voltage disturbance assessment method, system and device and a medium, and the method comprises the steps: firstly carrying out the voltage collection and phase calibration of a voltage channel and a photovoltaic inverter branch of a wind turbine generator, so as to generate a voltage disturbance time sequence structure set; performing partition statistics on amplitude energy on the structure set by using short-time Fourier transform, identifying an energy abrupt change interval, extracting an amplitude fluctuation sequence, and generating a disturbance energy characteristic matrix; generating a disturbance level probability distribution sequence based on the disturbance energy characteristic matrix; based on the probability distribution sequence, strong disturbance and weak disturbance partitions are determined, a disturbance stage mapping structure is generated, finally, based on the disturbance stage mapping structure, stage duration and grade rate are compared, a trend extension section is identified, an energy fluctuation increasing rate is calculated, a change direction is determined, and a disturbance trend evolution result is generated. According to the invention, through dynamic connection and information recursion of each step, accurate evaluation of voltage disturbance is realized.

Description

Technical Field

[0001] This invention belongs to the field of voltage disturbance monitoring technology, specifically relating to a method, system, equipment, and medium for assessing voltage disturbances in wind and solar grid connections. Background Technology

[0002] Wind and solar grid-connected voltage disturbance assessment refers to the detection, analysis, and quantitative evaluation of voltage fluctuations caused by wind and solar power generation systems during grid-connected operation, in order to accurately determine the impact of new energy grid connection on power quality and voltage stability. Through real-time monitoring of the grid connection point voltage and extraction of disturbance characteristics, disturbance phenomena such as voltage fluctuations, voltage sags, voltage flicker, and harmonic distortion are identified, and quantifiable voltage disturbance level indicators are generated, providing quantifiable assessment basis for grid operation.

[0003] Existing voltage disturbance monitoring technologies generally employ single-frequency domain calculation and static threshold determination logic during the analysis process. The processing flow lacks dynamic connection and information recursion mechanisms between algorithms. Frequency and amplitude are calculated independently only through a single sampling, resulting in discrete signal characteristics over time, failing to form a continuous record of disturbance evolution. The energy statistics stage relies heavily on global average integration or fixed-interval accumulation, leading to a lack of time-response characteristics in energy distribution and difficulty in reflecting transient fluctuations. When flicker, sags, and harmonic superposition coexist in the voltage signal, the interactive effects of multiple disturbances cannot be distinguished through single-point spectrum inversion, resulting in delays or errors in the results. State determination still relies on fixed threshold comparisons, neglecting the composite relationship between rate and duration, thus leading to over-response or judgment lag during level switching. The trend identification section lacks directional aggregation and energy gradient calculation processes, relying solely on changes in the average amplitude to determine the trend direction, making it difficult to quantitatively describe the duration and intensity of disturbances. When superimposed disturbances occur from renewable energy grid-connected signals, the system cannot effectively analyze multi-stage energy dynamics and trend extension paths, thus limiting the ability to accurately assess the development patterns of voltage disturbances. Summary of the Invention

[0004] The purpose of this invention is to address the aforementioned problems in the existing technology by providing a method, system, device, and medium for assessing voltage disturbances in wind and solar grid connection, which quantitatively describes the persistence and intensity of disturbances through continuous disturbance evolution and achieves accurate voltage disturbance assessment.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] In a first aspect, the present invention provides a method for assessing voltage disturbances in wind and solar grid-connected systems, the method comprising:

[0007] S1. Perform voltage acquisition and phase calibration on the voltage channel of the wind turbine and the photovoltaic inverter branch to generate a voltage disturbance time sequence structure set.

[0008] S2. Based on the voltage disturbance time series structure set, the amplitude energy is statistically analyzed by partition, the energy ratio and its variation amplitude of different frequency bands are calculated, the energy mutation interval is identified based on the variation amplitude and the amplitude fluctuation sequence of the energy mutation interval is extracted to generate the disturbance energy feature matrix.

[0009] S3. Based on the disturbance energy feature matrix, track the relationship between disturbance amplitude rate and energy fluctuation, determine the level state transition conditions, and combine the disturbance duration to correct the level weights and generate a disturbance level probability distribution sequence.

[0010] S4. Based on the probability distribution sequence of disturbance levels, sort out the trajectory of level changes, analyze the corresponding direction of the output of the reactive power compensation unit and the level sequence, determine the strong disturbance and weak disturbance partitions, and generate the disturbance stage mapping structure.

[0011] S5. Based on the perturbation stage mapping structure, compare the stage duration with the level rate, identify the trend extension segment, calculate the energy fluctuation increase rate and determine the direction of change, and generate the perturbation trend evolution result.

[0012] The voltage perturbation timing structure set includes a voltage amplitude sequence, a phase offset angle sequence, and a frequency offset sequence;

[0013] The disturbance energy feature matrix includes energy proportion coefficient, entropy change gradient, and frequency fluctuation number parameter;

[0014] The disturbance level probability distribution sequence includes a disturbance level label, a state transition weight, and a level probability vector;

[0015] The disturbance stage mapping structure includes stage number, level interval boundary and reactive power compensation state index;

[0016] The disturbance trend evolution result includes a time index, a disturbance trend intensity value, and a direction indicator.

[0017] S1 includes:

[0018] S11. Based on the voltage channel of the wind turbine and the photovoltaic inverter branch, record the amplitude of each channel point by point and align them with a unified timestamp. Calculate the difference of the channel amplitude at the same time, calculate the frequency offset according to the adjacent sample points and enter it into the table according to the time sequence. Organize it into the channel and time index corresponding to the row and column to generate a voltage offset parameter set.

[0019] S12. Based on the voltage frequency offset parameter set, register the peak and valley positions on the time axis, calculate the phase difference according to the peak and valley positions and calculate the angle change rate of the continuous interval, record the amplitude ratio and offset direction in pairs, and generate a voltage disturbance time sequence structure set after being aggregated by time period.

[0020] S2 includes:

[0021] S21. Based on the voltage disturbance time sequence structure set, short-time Fourier transform is used. The amplitude signal of each time period is divided into fixed intervals. The amplitude of each interval is squared and accumulated and the result is recorded. Then, the energy ratio is calculated according to the frequency dimension and arranged in chronological order to form a distribution record table to generate an energy distribution sequence set.

[0022] S22. Based on the energy distribution sequence set, calculate the change amplitude of energy ratio in adjacent frequency bands, label the frequency bands whose amplitude changes exceed the set range as abrupt segments, extract their time index and amplitude fluctuation information and archive them in order to generate an energy abrupt feature set.

[0023] S23. Based on the energy mutation feature set, each energy change value is converted into a logarithmic value and its weighted total is calculated. Combined with the time axis record, an entropy value sequence is formed. Then, the entropy change rate and the number of frequency fluctuations are cross-matched and organized into matrix entries to establish a perturbation energy feature matrix.

[0024] S3 includes:

[0025] S31. First, based on the perturbation energy feature matrix, Kalman filtering is used to calculate the amplitude difference between adjacent time points and extract the rate change direction. At the same time, the corresponding energy fluctuation interval is recorded. The rate change direction, energy amplitude and time index are combined into a unified recording sequence to generate a perturbation change parameter set.

[0026] S32. Based on the set of disturbance change parameters, compare the relationship between the rate change direction and energy fluctuation over a continuous time period. When the two directions reverse or the amplitude deviates beyond the threshold, mark the level state change, record each state interval and the number of changes, and generate a disturbance state mapping table.

[0027] S33. Based on the disturbance state mapping table, calculate the duration of each level state and its proportion, compare and correct the time weight with the frequency of state changes, arrange the probabilities of each level in time order to form a continuous sequence, and generate the disturbance level probability distribution sequence.

[0028] S4 includes:

[0029] S41. Based on the probability distribution sequence of disturbance levels, arrange the level labels in order of time step number, calculate the sign of the difference between adjacent labels to mark the upward and downward directions, count the start and end and duration of continuous same direction, merge adjacent micro-amplitude fluctuation segments, and generate a disturbance level sequence mapping table.

[0030] S42. Based on the disturbance level sequence mapping table, the output change rate of the reactive power compensation unit is compared with the level direction segment by segment, the same-direction and opposite-direction intervals are marked, the length of each interval and the level span are counted, the strong disturbance and weak disturbance segments are defined and numbered according to the span threshold, and the disturbance stage mapping structure is established.

[0031] S5 includes:

[0032] S51. Based on the disturbance stage mapping structure, compare the difference between the duration and grade rate of adjacent stages, screen out continuous segments with the same direction and mark them for extension, record the stage number and the start and end time of the segment, and generate a trend extension segment index table.

[0033] S52. Based on the trend extension segment index table, extract the energy fluctuation peak-valley difference sequence of each extension segment and calculate the increment rate over time. Accumulate the values ​​segment by segment to obtain the intensity quantity. Combine the level direction to give the composite symbol. Output the trend strength and direction in time order to generate the disturbance trend evolution result.

[0034] Secondly, the present invention provides a wind and solar grid-connected voltage disturbance assessment system, which includes a disturbance signal synchronization module, an energy entropy construction module, a disturbance probability evolution module, a level mapping judgment module, and a trend prediction inference module.

[0035] The disturbance signal synchronization module is used to collect voltage and perform phase calibration on the voltage channel of the wind turbine and the photovoltaic inverter branch to generate a voltage disturbance timing structure set.

[0036] The energy entropy construction module is used to calculate the energy ratio and its variation amplitude of different frequency bands based on the voltage disturbance time series structure set, using short-time Fourier transform partitioned statistical amplitude energy, identifying energy mutation intervals based on the variation amplitude and extracting the amplitude fluctuation sequence of the energy mutation intervals to generate a disturbance energy feature matrix.

[0037] The disturbance probability evolution module is used to track the relationship between disturbance amplitude rate and energy fluctuation based on the disturbance energy feature matrix using Kalman filtering, determine the level state transition conditions, and modify the level weights in combination with the disturbance duration to generate a disturbance level probability distribution sequence.

[0038] The level mapping judgment module is used to sort out the level change trajectory based on the disturbance level probability distribution sequence, analyze the corresponding direction of the reactive power compensation unit output and the level sequence, determine the strong disturbance and weak disturbance partitions, and generate a disturbance stage mapping structure.

[0039] The trend prediction and inference module is used to compare the duration of the stage with the rate of change based on the perturbation stage mapping structure, identify the trend extension segment, calculate the energy fluctuation increase rate and determine the direction of change, and generate the perturbation trend evolution result.

[0040] The voltage perturbation timing structure set includes a voltage amplitude sequence, a phase offset angle sequence, and a frequency offset sequence;

[0041] The disturbance energy feature matrix includes energy proportion coefficient, entropy change gradient, and frequency fluctuation number parameter;

[0042] The disturbance level probability distribution sequence includes a disturbance level label, a state transition weight, and a level probability vector;

[0043] The disturbance stage mapping structure includes stage number, level interval boundary and reactive power compensation state index;

[0044] The disturbance trend evolution result includes a time index, a disturbance trend intensity value, and a direction indicator.

[0045] The disturbance signal synchronization module includes a voltage acquisition submodule and a phase calibration submodule;

[0046] The voltage acquisition submodule is used to perform voltage acquisition according to the following steps and generate a voltage offset parameter set: based on the wind turbine voltage channel and the photovoltaic inverter branch, record the amplitude of each channel point by point and align them with a unified timestamp, calculate the difference of the channel amplitude at the same time, calculate the frequency offset according to adjacent sample points and enter it into the table according to the time sequence, organize it into the channel and time index corresponding to the row and column, and generate a voltage offset parameter set.

[0047] The phase calibration submodule is used to perform phase calibration according to the following steps to generate a voltage disturbance time series structure set: based on the voltage frequency offset parameter set, register the peak and valley positions on the time axis, calculate the phase difference according to the peak and valley positions and calculate the angle change rate of the continuous interval, record the amplitude ratio and offset direction in pairs, and after being aggregated by time period, generate a voltage disturbance time series structure set.

[0048] The energy entropy construction module includes an energy partition calculation submodule, an energy mutation identification submodule, and an entropy matrix generation submodule.

[0049] The energy partitioning calculation submodule is used to perform energy partitioning calculation according to the following steps to generate an energy distribution sequence set: based on the voltage disturbance time sequence structure set, using short-time Fourier transform, the amplitude signal of each time period is divided into fixed intervals, the amplitude of each interval is squared and accumulated and the result is recorded, then the energy ratio is calculated according to the frequency dimension, and the distribution record table is formed according to the time sequence to generate an energy distribution sequence set.

[0050] The energy mutation identification submodule is used to perform energy mutation identification calculations according to the following steps to generate an energy mutation feature set: based on the energy distribution sequence set, calculate the change amplitude of the energy ratio in adjacent frequency bands, label the frequency bands whose amplitude changes exceed the set range as mutation segments, extract their time index and amplitude fluctuation information and archive them in order to generate an energy mutation feature set;

[0051] The entropy matrix generation submodule is used to perform entropy matrix generation calculation according to the following steps to establish a disturbance energy feature matrix: based on the energy mutation feature set, each energy change value is converted into a logarithmic value and its weighted total is calculated. Combined with the time axis record, an entropy value sequence is formed. Then, the entropy change rate and the number of frequency fluctuations are cross-matched and organized into matrix entries to establish the disturbance energy feature matrix.

[0052] The perturbation probability evolution module includes a perturbation rate tracking submodule, a state determination submodule, and a level update submodule.

[0053] The disturbance rate tracking submodule is used to perform disturbance rate tracking calculations according to the following steps to generate a disturbance change parameter set: First, based on the disturbance energy feature matrix, Kalman filtering is used to calculate the amplitude difference between adjacent time points and extract the rate change direction. At the same time, the corresponding energy fluctuation interval is recorded. The rate change direction, energy amplitude and time index are combined into a unified recording sequence to generate a disturbance change parameter set.

[0054] The state determination submodule is used to perform state determination according to the following steps to generate a disturbance state mapping table: based on the disturbance change parameter set, the relationship between the rate change direction and energy fluctuation over a continuous time period is compared. When the two directions reverse or the amplitude deviates beyond the threshold, the level state change is marked, each state interval and the number of changes are recorded, and a disturbance state mapping table is generated.

[0055] The level update submodule is used to perform level updates according to the following steps to generate a disturbance level probability distribution sequence: based on the disturbance state mapping table, calculate the duration of each level state and its proportion, compare and correct the time weight with the state change frequency, arrange the probabilities of each level in time order to form a continuous sequence, and generate a disturbance level probability distribution sequence.

[0056] The level mapping judgment module includes a level trajectory parsing submodule and a stage partition recognition submodule;

[0057] The grade trajectory analysis is used to perform grade trajectory analysis according to the following steps to generate a disturbance grade sequence mapping table: based on the disturbance grade probability distribution sequence, the grade labels are arranged sequentially according to the time step number, the sign of the difference between adjacent labels is calculated to mark the upward and downward directions, the start and end and duration of continuous same direction are counted, and adjacent micro-amplitude fluctuation segments are merged to generate a disturbance grade sequence mapping table.

[0058] The stage partition identification submodule is used to perform stage partition identification according to the following steps to establish a disturbance stage mapping structure: based on the disturbance level sequence mapping table, the output change rate of the reactive power compensation unit is compared with the level direction segment by segment, the same-direction and opposite-direction intervals are marked, the length of each interval and the level span are counted, strong disturbance and weak disturbance segments are defined and numbered according to the span threshold, and a disturbance stage mapping structure is established.

[0059] The trend prediction and inference module includes a trend extension identification submodule and a trend strength calculation submodule;

[0060] The trend extension identification submodule is used to perform trend extension identification according to the following steps to generate a trend extension segment index table: based on the disturbance stage mapping structure, compare the difference between the duration and grade rate of adjacent stages, screen out continuous segments with the same direction and mark them for extension, record the stage number and the start and end time of the segment, and generate a trend extension segment index table.

[0061] The trend intensity calculation submodule is used to perform trend intensity calculation according to the following steps to generate the disturbance trend evolution result: based on the trend extension segment index table, extract the energy fluctuation peak-valley difference sequence of each extension segment and calculate the increment rate over time, accumulate segment by segment to obtain the intensity, combine the level direction to give the composite symbol, output the trend strength and direction in time order, and generate the disturbance trend evolution result.

[0062] Thirdly, the present invention provides a wind and solar grid-connected voltage disturbance assessment device, the wind and solar grid-connected voltage disturbance assessment device including a memory and a processor; the memory is used to store computer program code and transmit the computer program code to the processor; the processor is used to execute the aforementioned wind and solar grid-connected voltage disturbance assessment method according to the instructions in the computer program code.

[0063] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the aforementioned wind-solar grid-connected voltage disturbance assessment method.

[0064] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0065] 1. In the wind-solar grid-connected voltage disturbance assessment method of this invention, voltage acquisition and phase calibration are first performed on the voltage channel of the wind turbine and the photovoltaic inverter branch to generate a voltage disturbance time series structure set that is continuous in the time dimension, providing data support for subsequent disturbance evolution; then, based on the voltage disturbance time series structure set, short-time Fourier transform is used to partition and statistically analyze amplitude energy, calculate the energy ratio and its variation amplitude of different frequency bands, identify energy mutation intervals based on the variation amplitude, and extract the amplitude fluctuation sequence of the energy mutation intervals to generate a disturbance energy feature matrix to reflect the transient characteristics of fluctuations; subsequently, based on the disturbance energy feature matrix, Kalman filtering is used to track the relationship between disturbance amplitude rate and energy fluctuation, determine the level state transition conditions, and combine the disturbance... The system adjusts the level weight based on the duration of disturbances to generate a probability distribution sequence of disturbance levels. The composite relationship between rate and duration is considered during state determination to ensure timely response during level switching. Based on this probability distribution sequence, the system traces the level change trajectory, analyzes the correspondence between the reactive power compensation unit output and the level sequence, identifies strong and weak disturbance zones, and generates a disturbance stage mapping structure. Based on this structure, the stage duration and level rate are compared to identify trend extension segments, calculate the energy fluctuation increase rate, determine the direction of change, and generate disturbance trend evolution results. By identifying trend extension segments, the system can effectively analyze multi-stage energy dynamics and trend extension paths, ultimately achieving an accurate assessment of the voltage disturbance development pattern. Attached Figure Description

[0066] Figure 1 This is a flowchart of the wind and solar grid-connected voltage disturbance assessment method described in this invention.

[0067] Figure 2 This is a schematic diagram of the wind and solar grid-connected voltage disturbance assessment system described in this invention.

[0068] Figure 3 This is a schematic diagram of the wind and solar grid-connected voltage disturbance assessment device described in this invention. Detailed Implementation

[0069] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings.

[0070] Example 1:

[0071] See Figure 1 A method for assessing voltage disturbances in wind and solar grid connection is performed in the following steps:

[0072] S1. Perform voltage acquisition and phase calibration on the voltage channel of the wind turbine and the photovoltaic inverter branch to generate a voltage disturbance time sequence structure set.

[0073] The voltage perturbation timing structure set includes: a voltage amplitude sequence, a phase offset angle sequence, and a frequency offset sequence. S1 specifically includes the following steps:

[0074] S11. Voltage Acquisition: Based on the wind turbine voltage channel and the photovoltaic inverter branch, the voltage amplitude of the wind turbine voltage channel and the voltage amplitude of the output side of the photovoltaic inverter branch are recorded point by point and aligned with a unified timestamp. The difference of the channel amplitude at the same time is calculated, the frequency offset is calculated according to the adjacent sample points and entered into the table according to the time sequence, and the data is organized into a channel and time index corresponding to the row and column to generate a voltage offset parameter set.

[0075] For example, for the voltage channel of the wind turbine and the photovoltaic inverter branch, the Fast Fourier Transform (FFT) algorithm is used to segment and sample the channel signal. The number of sampling segments is set to 512, and the sampling frequency is set to 1000 Hz. Windowing operation is performed on each channel signal, with the window type being Hanning window. Frequency domain conversion is performed using signal processing instructions. The amplitude sequence is extracted from each sampling segment and aligned with a unified timestamp to generate a voltage amplitude sequence. The difference operation is performed on the channel amplitude between adjacent sampling segments at the same time point to obtain the amplitude difference between adjacent sampling segments, thereby generating voltage frequency offset data. The calculated results are arranged in a time sequence, with the channel number defined as the row index and the time number defined as the column index. The voltage frequency offset matrix is ​​generated using matrix storage commands. The frequency offset values ​​are entered in time order to construct a two-dimensional correspondence table between channels and time, generating a voltage frequency offset parameter set.

[0076] S12. Phase calibration: Based on the voltage frequency offset parameter set, register the peak and valley positions on the time axis, calculate the phase difference according to the peak and valley positions and calculate the angle change rate of the continuous interval, record the amplitude ratio and offset direction in pairs, and generate a voltage disturbance time sequence structure set after being aggregated by time period.

[0077] For example, for the voltage frequency offset parameter set, the Hilbert Transform is used to perform signal phase envelope processing, and the input signal is analytically transformed to extract the instantaneous phase value. Then, the sample positions of peak and valley values ​​are registered on the time axis. The instantaneous phase obtained by the Hilbert Transform is calculated after peak and valley position registration, and the angle change rate is calculated for adjacent samples. The angle change rate is organized into a phase offset angle sequence, and the amplitude sequence is kept separately. The phase offset direction is used as a derived quantity of the phase difference sign. The three items of voltage amplitude sequence, phase offset angle sequence and frequency offset sequence are formed according to the time index, which serves as the voltage disturbance time series structure set.

[0078] The registration of peak and valley positions on the time axis refers to identifying the sampling points where local peaks and valleys are located by detecting continuous amplitude variation trends, recording the time interval between adjacent extreme points as periodic segments, and using the first maximum value as a time reference to interpolate and align the peak and valley sequences of all channels on the same time axis. The order of peak occurrence within the same periodic segment is compared according to channel number, and the amplitude ratio sequence and phase shift sequence are calculated, with both types of sequences marked in a time index table. Specifically, the amplitude values ​​of each channel are extracted at the same peak and valley position, the ratio of the wind power channel amplitude to the photovoltaic channel amplitude is calculated to form an amplitude ratio sequence, and the phase difference is calculated based on the instantaneous phase sequence obtained by Hilbert transform. The sign of the phase difference determines the phase shift direction value, which is recorded as a sign value. The amplitude ratio sequence and the phase shift direction are independently derived features.

[0079] S2. Based on the voltage disturbance time series structure set, the amplitude energy is statistically analyzed by partitioning using short-time Fourier transform, the energy ratio and its variation amplitude of different frequency bands are calculated, the energy mutation interval is identified based on the variation amplitude, and the amplitude fluctuation sequence of the energy mutation interval is extracted to generate the disturbance energy feature matrix.

[0080] The disturbance energy feature matrix includes an energy proportion coefficient, entropy change gradient, and frequency fluctuation number parameter. S2 specifically includes the following steps:

[0081] S21. Energy partitioning calculation: Based on the voltage disturbance time sequence structure set, the short-time Fourier transform is used to divide the amplitude signal of each time period into fixed intervals. The amplitude of each interval is squared and accumulated and the results are recorded. Then, the energy ratio is calculated according to the frequency dimension and arranged in chronological order to form a distribution record table, generating an energy distribution sequence set.

[0082] The Short Time Fourier Transform (STFT) algorithm involves dividing the voltage disturbance time-series structure set into several time frames of fixed length. Within each time frame, an amplitude sequence is selected and a window function is applied for point-by-point weighting. Subsequently, discrete spectrum decomposition is performed on the weighted sequence to extract the amplitude and phase of the complex spectrum and calculate the square of the amplitude to obtain the power spectral density. Then, frequency coordinates are converted according to the sampling rate, and the frequency index is mapped to a preset frequency band interval. The power of each frequency band within the same frame is summed to obtain the energy distribution. This process is repeated for all time frames, and the data is organized in chronological order to form a time and frequency index table with corresponding rows and columns. Symmetrical padding and truncation are performed on the edge frame data, and the frequency band energy sequences of each time frame are summarized.

[0083] For example, based on the voltage perturbation time-series structure set, the Short-Time Fourier Transform (STFT) algorithm is used for time-frequency partitioning calculation. The input signal is framed with a frame length of 256 points, a frame shift of 128 points, and a rectangular window type. Spectral decomposition is performed on each frame with a frequency resolution of 0.5 Hz and a time resolution of 0.001 seconds. The amplitude of each frequency in each frame's spectrum is squared and accumulated. Numerical integration is used to record the accumulated results in a two-dimensional array. The intervals divided into time periods are indexed and stored sequentially. Then, the energy ratio of the frequency band energy obtained from the STFT is calculated. This calculation includes obtaining the cumulative energy ratio of each frequency band according to the frequency index, unifying the ratio to a numerical format ranging from zero to one, arranging the results in chronological order to generate a distribution record table, forming an energy sequence matrix arranged by frequency and time index, and generating an energy distribution sequence set. The energy distribution sequence set includes a time index and an energy ratio vector obtained by normalizing the frequency band energy of each time frame.

[0084] The formula for the short-time Fourier transform is as follows:

[0085] ;

[0086] In the above formula, For the first Voltage disturbance timing sequence in the first The time period, the first Weighted short-time spectral values ​​at each frequency index; For the first time collected Voltage disturbance timing sequence in the first Within the time period, the first Discrete amplitude signal at each sampling point; For window functions; In order to target the The sequence weighting coefficients set for the overall perturbation strength of the voltage perturbation timing sequence; In order to target the The time period reliability coefficient; In order to target the Frequency sensitivity coefficients of each frequency index; The step size between the starting sampling points of adjacent time periods; Indicates the sampling length for a single time period; In order to target the Window length correction for voltage disturbance timing; The frequency index ranges from 0 to... ; The time period index increments sequentially from the first segment. The index of local sampling points within the time period is from 0 to ; The imaginary unit; Pi is a constant.

[0087] The calculation process for S21 is as follows: First, the multiple voltage disturbance time series collected are numbered sequentially by time as follows: The time sequence is divided into segments, and the sampling length and sampling period of each segment are determined. Then, a fixed step size is used. Determine the first by translating along the time axis The starting sampling position of each time period, and a truncated section of length from the starting position. The continuous sampling points constitute the analysis window, and the amplitude signal of each sampling point within the window is denoted as . and window function The time segments are multiplied point by point to form a weighted time segment, which is used to suppress boundary effects and improve time-frequency resolution. Then, the perturbation intensity of the entire time series is calculated, and the sequence weight coefficient is obtained based on its ratio to the maximum energy of all samples. This is used to control the contribution ratio of different voltage disturbance samples to the overall energy calculation, and then the ratio of the instantaneous energy of the current time period to the average energy of the entire period is calculated to determine the weighting coefficient of the time period. It highlights the characteristic response during sudden disturbances, and indexes each frequency based on the system's preset power grid frequency and harmonic frequency characteristic ranges. Assigning frequency sensitivity coefficient This enhances the spectral energy response at characteristic frequencies, and then the window length correction is determined based on the bandwidth estimation of this time series. This is used to correct the balance between frequency resolution and time resolution, and to convert the signal... With window function Sequence weight coefficients Time period weighting coefficient Frequency sensitivity coefficient Multiply by each term in sequence, then multiply by the complex exponent. ,right From 0 to Perform a discrete summation operation to obtain the weighted short-time spectrum value. For all time periods and frequency index Repeat the above calculation process to obtain the complete time-frequency spectrum matrix. Finally, calculate the energy ratio of each interval in the frequency dimension to form an energy distribution sequence set, which is used to quantitatively evaluate the voltage disturbance characteristics of wind and solar grid connection.

[0088] S22. Energy mutation identification: Based on the energy distribution sequence set, compare the differences in energy ratios of each frequency band, scan the energy change amplitude of adjacent intervals, mark the segments whose amplitude changes exceed the set range as mutation segments, extract their time index and amplitude fluctuation information and archive them in order to generate an energy mutation feature set.

[0089] For example, based on the energy distribution sequence set, a sliding interval differential detection algorithm is used to identify energy mutations. The detection window width is set to five time units, the step interval is one time unit, and a differential scanning operation is performed on the energy ratio of adjacent intervals. The differential threshold is set to 0.1. Time periods with differences exceeding the threshold are marked, and the index of the marked time period and the corresponding amplitude sequence are written into a temporary cache table. Each marked segment is sorted by time and an index table is built. The mutation segment number, time index and energy change value are archived. The storage format adopts a matrix row and column mapping method, with rows corresponding to mutation segment numbers and columns corresponding to energy change values, generating an energy mutation feature set.

[0090] S23. Entropy Matrix Generation: Based on the energy mutation feature set, each energy change value is converted into a logarithmic value and its weighted total is calculated. Combined with the time axis record, an entropy value sequence is formed. Then, the entropy change rate and the number of frequency fluctuations are cross-matched and organized into matrix entries to establish the perturbation energy feature matrix.

[0091] For example, based on the energy mutation feature set, the Shannon Entropy algorithm is used to calculate energy entropy values. A logarithmic transformation is performed on each energy change value, with the base set to two. The transformation result is multiplied by the weight of the energy change value and then weighted and summed. A weighted summation command is executed to generate an entropy value sequence. A time series table is established by combining time index records. The entropy value sequence is differentially processed according to the time step to obtain the entropy change rate. The entropy change rate and the number of frequency fluctuations are cross-matched. This cross-matching refers to mapping the entropy change rate sequence to the cumulative number of frequency fluctuations obtained according to the time index based on the time index. The entropy change rate value and the number of frequency fluctuations at the same time point are combined into two fields in the same matrix row, forming a two-dimensional structure with the time index as the row and the combined value of the entropy change rate and the number of frequency fluctuations as the column. The matching step is set to one time unit. Matrix entries are generated and written into the two-dimensional matrix structure, with rows corresponding to the time index and columns corresponding to the combined value of the entropy change rate and the number of frequency fluctuations. Finally, a perturbation energy feature matrix is ​​generated.

[0092] S3. Based on the disturbance energy feature matrix, Kalman filtering is used to track the relationship between disturbance amplitude rate and energy fluctuation, determine the level state transition conditions, and combine the disturbance duration to correct the level weights, thereby generating a disturbance level probability distribution sequence.

[0093] The disturbance level probability distribution sequence includes a disturbance level label, state transition weights, and a level probability vector. S3 specifically includes the following steps:

[0094] S31. Disturbance Rate Tracking: First, based on the disturbance energy feature matrix, Kalman filtering is used to calculate the amplitude difference between adjacent time points and extract the direction of change. At the same time, the corresponding energy fluctuation interval is recorded. The rate, energy amplitude and time index are combined into a unified recording sequence to generate a set of disturbance change parameters.

[0095] For example, based on the disturbance energy feature matrix, the Kalman filter algorithm is used to calculate the disturbance rate tracking. An initialization operation is performed on the Kalman filter to give it three terms: amplitude, rate, and acceleration. The initial state vector is set to include these three terms. Specifically, during the initialization operation, the voltage amplitude sequence is used as the basic amplitude quantity. The initial amplitude difference is formed by the amplitude difference between adjacent time indices. The set initial rate and zero acceleration form the rate and acceleration terms of the initial state vector. At the same time index, the energy fluctuation interval number is read from the disturbance energy feature matrix. The initial amplitude difference, initial rate, and energy fluctuation interval are combined into the three basic quantities of the state vector. The initial covariance matrix is ​​set to 0.001 for the main diagonal elements, the process noise covariance matrix is ​​set to 0.01, and the observation noise covariance matrix is ​​set to 0.05. The state estimate is updated using the prediction command. State prediction and observation update operations are performed at each time step. In the prediction stage, the amplitude difference is calculated using a first-order linear recursion based on the state transition matrix. In the update stage, the measurement residuals are weighted and corrected. The amplitude difference, prediction rate, and energy fluctuation range at each time point are recorded synchronously. The direction of change is recorded in the direction array according to the positive and negative change indicators. The rate, energy amplitude, and time index are serialized and combined to establish a structured record table containing three types of index fields, generating a set of disturbance change parameters.

[0096] Kalman filtering refers to using the voltage amplitude sequence as the amplitude source and the energy field in the disturbance energy feature matrix as the energy source. The two are combined under the same time index to form the initial values ​​of the state vector and establish the corresponding error covariance matrix. Then, based on the time step length, a state transition matrix and an observation matrix are constructed to establish a mapping relationship between rate change and energy observation. Next, state prediction is performed at each time step. The current rate and energy amplitude are inferred from the estimated state at the previous time step, and the covariance is predicted. Then, the amplitude difference between adjacent time points and the energy fluctuation value are combined to form the observation vector. The observation residual and its covariance are calculated. Then, the Kalman gain is calculated using the residual and the state vector and covariance matrix are corrected. Finally, the updated rate component and change direction are extracted and paired with the time index and energy fluctuation interval and stored in a record table to form a disturbance change parameter set.

[0097] The formula for calculating Kalman filtering is as follows:

[0098] ;

[0099] In the above formula, This indicates the time index of the wind and solar grid-connected voltage disturbance assessment system. The posterior state estimation vector at the location; Indicates based on time index The prior state estimation vector obtained from the state transition model; Indicates an index for energy ranges The set interval confidence coefficient; Indicates time index The disturbance activity coefficient at the location; Indicates the first The stability coefficient of the perturbation sequence; Indicates time index The Kalman gain matrix at the location; Indicates time index Energy range The observations extracted from the perturbation energy feature matrix; Indicates time index Baseline correction amount at; Represents the observation matrix; Indicates a time index; Indicates time index The selected energy range index; Indicates the perturbation sequence number.

[0100] The calculation process of S31 is as follows: First, extract the current time index from the disturbance energy feature matrix. Index of the energy ranges with the most significant energy changes The energy difference is calculated to obtain the observable. The reliability coefficient of the energy range is determined based on the reliability of the range identification. The weights of observations on state updates are adjusted, and then the time activity coefficient is calculated using the moving variance. A larger value is chosen when there are frequent disturbances to improve the response speed, and a lower value is chosen when the signal is stable to maintain smooth filtering. At the same time, a sequence stability coefficient is set for different signal sources. The signal reliability of different grid-connected sources, such as wind power and photovoltaic nodes, is differentiated, and the baseline correction is calculated using the median and mean energy values ​​during the low-disturbance phase. This is used to eliminate the effects of sensor drift, and then the prior state is calculated based on the state of the previous time step and the state transition model. And calculate the Kalman gain. Minimize the estimation error and construct the observation residuals Multiply by Kalman gain in sequence Sequence stability coefficient Time activity coefficient Energy range credibility coefficient Adding the prior state to the posterior state yields the posterior state. And index the rate value, energy amplitude, and energy range. and time index A set of disturbance variation parameters is formed for voltage disturbance trend analysis and level assessment.

[0101] S32. State determination: Based on the set of disturbance change parameters, compare the relationship between the rate direction and energy fluctuation over a continuous time period. When the two directions reverse or the amplitude deviates beyond the threshold, mark the level of state change, record each state interval and the number of changes, and generate a disturbance state mapping table.

[0102] For example, based on the perturbation change parameter set, the Interval Logical Determination Algorithm is used to perform a comparison operation on the relationship between the rate direction and energy fluctuation over a continuous time period. The direction determination flag is set to one or negative one, and the amplitude deviation threshold is set to 0.2. The input sequence is traversed in chronological order, and a logical XOR calculation is performed on the rate direction of adjacent intervals. When the result is true, the direction is marked as reversed. At the same time, an absolute difference comparison operation is performed on the deviation of the energy amplitude of the current interval. When the result exceeds the threshold, the state change is marked. The state number, start and end time index and number of changes are recorded. All state identifiers are written into the state index table and arranged in chronological order to form a matrix structure, generating a perturbation state mapping table.

[0103] S33, Level Update: Based on the disturbance state mapping table, calculate the duration of each level state and its proportion, compare and correct the time weight with the frequency of state changes, arrange the probabilities of each level in time order to form a continuous sequence, and generate the disturbance level probability distribution sequence.

[0104] For example, based on the perturbation state mapping table, the TimeWeightedProbabilityUpdatingAlgorithm is used to update the level state. For different level states, the combination characteristics of the rate change direction and energy fluctuation amplitude of the perturbation change parameters are determined. Records with the rate direction remaining consistent in a continuous time period and the energy fluctuation amplitude being in the same range are grouped into the same state segment. Based on the preset energy amplitude level division threshold, the energy fluctuation amplitude of the state segment is mapped to the corresponding level label, so that the state segment has a clear level. The time weight factor is set to 0.5, and the state change frequency weight factor is set to 0.3. The weights are normalized. When correcting the level probability sequence using time weight and change frequency, the duration of each level state is used as the basic quantity of time weight, and the number of changes of each level state in the disturbance state mapping table is used as the basic quantity of change frequency. The two are linearly combined according to the preset weight factor to form the correction coefficient used to correct the initial probability of the level. The correction coefficient is then multiplied by the initial probability value of the level to obtain the corrected level probability value. The corrected level probability values ​​corresponding to each time step are concatenated into a continuous time series according to the time index order, and a corresponding matrix index is established. The level number, probability value and time field are arranged in order to generate the disturbance level probability distribution sequence.

[0105] S4. Based on the probability distribution sequence of disturbance levels, sort out the trajectory of level changes, analyze the corresponding direction of the output of the reactive power compensation unit and the level sequence, determine the strong disturbance and weak disturbance partitions, and generate the disturbance stage mapping structure.

[0106] The disturbance stage mapping structure includes stage number, level interval boundary, and reactive power compensation state index. S4 specifically includes the following steps:

[0107] S41. Level Trajectory Analysis: Based on the probability distribution sequence of disturbance levels, the level labels are arranged sequentially according to the time step number. The sign of the difference between adjacent labels is calculated to mark the upward and downward directions. The start and end times and durations of continuous same directions are counted. Adjacent micro-amplitude fluctuation segments are merged to generate a disturbance level sequence mapping table.

[0108] For example, based on the probability distribution sequence of disturbance levels, the Symbolic Differential Trajectory Analysis Algorithm is used to parse the level trajectory. The level labels of the input sequence are arranged sequentially according to the time step number. A difference operation is performed on adjacent level labels, with the difference step size set to one. The difference result is used to take the sign function value for direction labeling. A direction label of one represents upward and a negative one represents downward. A direction sequence array is formed for all direction labels. An interval statistical operation is performed on consecutive segments with the same sign. The start time index and end time index of each segment are recorded. The segment length is calculated and written into the segment index table. An interval merging instruction is used to merge adjacent segments with a time interval of less than three steps. The merged segments are renumbered and stored in the index matrix. In the matrix, the rows correspond to the direction labels, and the columns correspond to the time range and duration of the segments. The finally arranged direction index, segment length and level label are serialized and stored to generate a disturbance level sequence mapping table.

[0109] S42. Stage Partition Identification: The stage partition identification submodule: Based on the disturbance level sequence mapping table, the output change rate of the reactive power compensation unit is compared with the level direction segment by segment, the same-direction and opposite-direction intervals are marked, the length of each interval and the level span are counted, the strong disturbance and weak disturbance segments are defined and numbered according to the span threshold, and the disturbance stage mapping structure is established.

[0110] For example, based on the disturbance level sequence mapping table, the Segmental Comparison Determination Algorithm is used for stage partition identification. The input reactive power compensation unit output rate of change sequence and level direction sequence are matched segment by segment. The time step width is set to one, and the direction comparison threshold is 0.1. The direction difference between adjacent segments is multiplied and judged. When the result is positive, the same direction interval is marked, and when the result is negative, the opposite direction interval is marked. Direction marking is recorded for all intervals. The time length and level span of each interval are counted. The length values ​​are summed to form a length vector. The absolute difference of the level span is calculated to form a span array. The span threshold is set to two. Intervals with spans greater than the threshold are defined as strong disturbances, and intervals with spans less than or equal to the threshold are defined as weak disturbances. All intervals are indexed and registered in numerical order. The interval number, direction type, interval length and level span are stored in a matrix table structure, and finally the disturbance stage mapping structure is generated.

[0111] By classifying the nature of disturbance stages into strong and weak disturbance partitions, subsequent trend extension identification can use the disturbance stage as the input basis. This clarifies the disturbance intensity category of each stage in directional similarity determination, candidate segment selection, and extension segment continuity verification. This ensures that the trend identification process has background information on disturbance intensity when judging directional consistency and rate differences, thereby avoiding misidentification of level changes caused by brief and weak fluctuations as trend extensions. At the same time, it ensures that only segments in a clear disturbance stage participate in trend extension segment generation.

[0112] S5. Based on the perturbation stage mapping structure, compare the stage duration with the level rate, identify the trend extension segment, calculate the energy fluctuation increase rate and determine the direction of change, and generate the perturbation trend evolution result.

[0113] The disturbance trend evolution result includes the trend extension segment number, energy increase rate sequence, and level change direction. S5 specifically includes the following steps:

[0114] S51. Trend Extension Identification: Based on the disturbance stage mapping structure, compare the difference between the duration and grade rate of adjacent stages, screen out continuous segments with the same direction and mark them for extension, record the stage number and the start and end time of the segment, and generate a trend extension segment index table.

[0115] For example, for the perturbation stage mapping structure, the Directional Similarity Clustering Algorithm is used for trend extension identification. The input stage matrix is ​​sorted by time index, the time step interval is set to one, and the directional similarity threshold is 0.9. The difference between the durations of adjacent stages is calculated, and the absolute difference between the grade rate difference is calculated. The results of the two are input into the directional similarity judgment instruction. The continuity is judged by multiplying the directional identifiers. When the directions match and the rate difference is less than 0.1, it is recorded as an extension candidate area. The time continuity verification operation is performed on the candidate area. If the time interval is less than three steps, it is merged into the same extension segment. The merged segments are renumbered and the stage number and start and end time index are recorded. The start point, end point, direction mark and the stage number of each extension segment are written into the index matrix and arranged in chronological order to form a two-dimensional table structure, generating a trend extension segment index table.

[0116] S52. Trend Intensity Calculation: Based on the trend extension segment index table, extract the energy fluctuation peak-valley difference sequence of each extension segment and calculate the increment rate over time. Accumulate the values ​​segment by segment to obtain the intensity quantity. Combine the level direction to give the composite symbol. Output the trend strength and direction in time order to generate the disturbance trend evolution result.

[0117] For example, based on the trend extension segment index table, the FluctuationEnergy Gradient Integration Algorithm is used to calculate the trend intensity. The peak and trough sequences of energy fluctuations are extracted from the input extension segment data. The peak-to-trough difference is calculated sequentially within each extension segment, with a time step of 0.01 seconds. An increment rate is calculated for the difference between each adjacent pair of points, and the increment rate is stored in an intensity accumulation cache array. A segmented accumulation operation is performed on the cache array, with an accumulation interval of five time steps. The resulting accumulation is assigned to the corresponding extension segment number. A sign operation is performed on the extension segment direction marker, assigning a value of one to positive segments and a value of negative one to negative segments. The intensity and sign are combined into a composite value. The trend strength and direction are output sequentially for all extension segments. A matrix field containing three items—time index, intensity value, and direction marker—is established to generate the perturbation trend evolution result.

[0118] Example 2:

[0119] See Figure 2 A wind and solar grid-connected voltage disturbance assessment system includes a disturbance signal synchronization module, an energy entropy construction module, a disturbance probability evolution module, a level mapping judgment module, and a trend prediction inference module.

[0120] The disturbance signal synchronization module is used to acquire voltage and calibrate phase of the wind turbine voltage channel and photovoltaic inverter branch to generate a voltage disturbance time sequence structure set; the voltage disturbance time sequence structure set includes voltage amplitude sequence, phase offset angle sequence and frequency offset sequence.

[0121] Specifically, the disturbance signal synchronization module includes a voltage acquisition submodule and a phase calibration submodule. The voltage acquisition submodule is used to perform voltage acquisition according to the following steps to generate a voltage offset parameter set: based on the wind turbine voltage channel and the photovoltaic inverter branch, the amplitude of each channel is recorded point by point and aligned with a unified timestamp. The difference between the channel amplitudes at the same time is calculated, the frequency offset is calculated according to adjacent samples and entered into a table according to the time sequence, and the data is organized into a row and column corresponding channel and time index to generate a voltage offset parameter set. The phase calibration submodule is used to perform phase calibration according to the following steps to generate a voltage disturbance time sequence structure set: based on the voltage frequency offset parameter set, the peak and valley positions are registered on the time axis, the phase difference is calculated according to the peak and valley positions and the angle change rate of the continuous interval is calculated, the amplitude ratio and the offset direction are recorded in pairs, and the data is aggregated according to the time period to generate a voltage disturbance time sequence structure set.

[0122] The energy entropy construction module is used to calculate the energy ratio and its variation amplitude of different frequency bands based on the voltage disturbance time series structure set, using short-time Fourier transform partitioned statistical amplitude energy, identifying energy mutation intervals based on the variation amplitude and extracting the amplitude fluctuation sequence of the energy mutation intervals to generate a disturbance energy feature matrix; the disturbance energy feature matrix includes energy ratio coefficient, entropy change gradient, and frequency fluctuation number parameter.

[0123] Specifically, the energy entropy construction module includes an energy partitioning calculation submodule, an energy mutation identification submodule, and an entropy matrix generation submodule. The energy partitioning calculation submodule performs energy partitioning calculations according to the following steps to generate an energy distribution sequence set: based on a voltage perturbation time-series structure set, using short-time Fourier transform, the amplitude signals of each time period are divided into fixed intervals; the amplitude of each interval is squared, accumulated, and the results are recorded; then, the energy proportion is calculated according to the frequency dimension, and the results are arranged chronologically to form a distribution record table to generate an energy distribution sequence set. The energy mutation identification submodule performs energy mutation identification calculations according to the following steps to generate an energy distribution sequence set. Energy mutation feature set: Based on the energy distribution sequence set, the change amplitude of energy ratio in adjacent frequency bands is calculated. Frequency bands with amplitude changes exceeding the set range are marked as mutation segments. Their time index and amplitude fluctuation information are extracted and archived in sequence to generate an energy mutation feature set. The entropy matrix generation submodule is used to perform entropy matrix generation calculation according to the following steps to establish a disturbance energy feature matrix: Based on the energy mutation feature set, each energy change value is converted into a logarithmic value and its weighted total is calculated. Combined with the time axis record, an entropy value sequence is formed. Then, the entropy change rate and the number of frequency fluctuations are cross-matched and organized into matrix entries to establish a disturbance energy feature matrix.

[0124] The disturbance probability evolution module is used to track the relationship between disturbance amplitude rate and energy fluctuation based on the disturbance energy feature matrix using Kalman filtering, determine the level state transition conditions, and modify the level weights in combination with the disturbance duration to generate a disturbance level probability distribution sequence; the disturbance level probability distribution sequence includes a disturbance level label, state transition weights, and a level probability vector.

[0125] Specifically, the perturbation probability evolution module includes a perturbation rate tracking submodule, a state determination submodule, and a level update submodule. The perturbation rate tracking submodule performs perturbation rate tracking calculations according to the following steps to generate a perturbation change parameter set: first, based on the perturbation energy feature matrix, it uses Kalman filtering to calculate the amplitude difference between adjacent time points and extract the rate change direction, while simultaneously recording the corresponding energy fluctuation interval. The rate change direction, energy amplitude, and time index are combined into a unified recording sequence to generate the perturbation change parameter set. The state determination submodule performs state determination according to the following steps to generate a perturbation state mapping table. Based on the set of disturbance change parameters, the relationship between the rate change direction and energy fluctuation over a continuous time period is compared. When the two directions reverse or the amplitude deviates beyond the threshold, the level state change is marked, and the intervals and number of changes of each state are recorded to generate a disturbance state mapping table. The level update submodule is used to perform level updates according to the following steps to generate a disturbance level probability distribution sequence: Based on the disturbance state mapping table, the duration of each level state is calculated and its proportion is determined. The time weight is compared and corrected with the frequency of state changes. The probabilities of each level are arranged in chronological order to form a continuous sequence, generating a disturbance level probability distribution sequence.

[0126] The level mapping judgment module is used to sort out the level change trajectory based on the disturbance level probability distribution sequence, analyze the corresponding direction of the reactive power compensation unit output and the level sequence, determine the strong disturbance and weak disturbance partitions, and generate a disturbance stage mapping structure; the disturbance stage mapping structure includes stage number, level interval boundary and reactive power compensation state index.

[0127] Specifically, the level mapping judgment module includes a level trajectory analysis submodule and a stage partition identification submodule. The level trajectory analysis is used to perform level trajectory analysis according to the following steps to generate a disturbance level sequence mapping table: based on the disturbance level probability distribution sequence, the level labels are arranged sequentially according to the time step number, the sign of the difference between adjacent labels is calculated to mark the upward and downward directions, the start and end times and durations of continuous same directions are counted, and adjacent micro-amplitude fluctuation segments are merged to generate a disturbance level sequence mapping table. The stage partition identification submodule is used to perform stage partition identification according to the following steps to establish a disturbance stage mapping structure: based on the disturbance level sequence mapping table, the output change rate of the reactive power compensation unit is compared with the level direction segment by segment, the same-direction and opposite-direction intervals are marked, the length of each interval and the level span are counted, and strong disturbance and weak disturbance segments are defined and numbered according to the span threshold to establish a disturbance stage mapping structure.

[0128] The trend prediction and inference module is used to compare the duration of the stage with the level rate based on the disturbance stage mapping structure, identify the trend extension segment, calculate the energy fluctuation increase rate and determine the direction of change, and generate the disturbance trend evolution result; the disturbance trend evolution result includes a time index, a disturbance trend intensity value and a direction indicator.

[0129] Specifically, the trend prediction and inference module includes a trend extension identification submodule and a trend intensity calculation submodule. The trend extension identification submodule is used to perform trend extension identification according to the following steps to generate a trend extension segment index table: based on the disturbance stage mapping structure, compare the duration and grade rate difference of adjacent stages, screen out continuous segments with the same direction and mark them for extension, record the stage number and the start and end time of the segment, and generate a trend extension segment index table. The trend intensity calculation submodule is used to perform trend intensity calculation according to the following steps to generate the disturbance trend evolution result: based on the trend extension segment index table, extract the energy fluctuation peak-valley difference sequence of each extension segment and calculate the increment rate over time, accumulate segment by segment to obtain the intensity, combine it with the grade direction to give a composite symbol, output the trend strength and direction in chronological order, and generate the disturbance trend evolution result.

[0130] Example 3:

[0131] See Figure 3 A wind and solar grid-connected voltage disturbance assessment device includes a memory and a processor; the memory is used to store computer program code and transmit the computer program code to the processor; the processor is used to execute the wind and solar grid-connected voltage disturbance assessment method described in Embodiment 1 according to the instructions in the computer program code.

[0132] Example 4:

[0133] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the wind-solar grid-connected voltage disturbance assessment method described in Example 1.

[0134] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program goods. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0135] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0136] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0137] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0138] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for assessing voltage disturbances in wind and solar grid connection, characterized in that: The method for assessing voltage disturbances in wind and solar grid connection includes: S1. Perform voltage acquisition and phase calibration on the voltage channel of the wind turbine and the photovoltaic inverter branch to generate a voltage disturbance time sequence structure set. S2. Based on the voltage disturbance time series structure set, the amplitude energy is statistically analyzed by partition, the energy ratio and its variation amplitude of different frequency bands are calculated, the energy mutation interval is identified based on the variation amplitude and the amplitude fluctuation sequence of the energy mutation interval is extracted to generate the disturbance energy feature matrix. S3. Based on the disturbance energy feature matrix, track the relationship between disturbance amplitude rate and energy fluctuation, determine the level state transition conditions, and combine the disturbance duration to correct the level weights and generate a disturbance level probability distribution sequence. S4. Based on the probability distribution sequence of disturbance levels, sort out the trajectory of level changes, analyze the corresponding direction of the output of the reactive power compensation unit and the level sequence, determine the strong disturbance and weak disturbance partitions, and generate the disturbance stage mapping structure. S5. Based on the perturbation stage mapping structure, compare the stage duration with the level rate, identify the trend extension segment, calculate the energy fluctuation increase rate and determine the direction of change, and generate the perturbation trend evolution result.

2. The method for evaluating voltage disturbances in wind and solar grid connection according to claim 1, characterized in that: The voltage perturbation timing structure set includes a voltage amplitude sequence, a phase offset angle sequence, and a frequency offset sequence; The disturbance energy feature matrix includes energy proportion coefficient, entropy change gradient, and fluctuation frequency parameter; The disturbance level probability distribution sequence includes a disturbance level label, a state transition weight, and a level probability vector; The disturbance stage mapping structure includes stage number, level interval boundary and reactive power compensation state index; The disturbance trend evolution result includes a time index, a disturbance trend intensity value, and a direction indicator.

3. The method for evaluating voltage disturbances in wind and solar grid connection according to claim 1, characterized in that: S1 includes: S11. Based on the voltage channel of the wind turbine and the photovoltaic inverter branch, record the amplitude of each channel point by point and align them with a unified timestamp. Calculate the difference of the channel amplitude at the same time, calculate the frequency offset according to the adjacent sample points and enter it into the table according to the time sequence. Organize it into the channel and time index corresponding to the row and column to generate a voltage offset parameter set. S12. Based on the voltage frequency offset parameter set, register the peak and valley positions on the time axis, calculate the phase difference according to the peak and valley positions and calculate the angle change rate of the continuous interval, record the amplitude ratio and offset direction in pairs, and generate a voltage disturbance time sequence structure set after being aggregated by time period.

4. The method for evaluating voltage disturbances in wind and solar grid connection according to claim 1, characterized in that: S2 includes: S21. Based on the voltage disturbance time sequence structure set, short-time Fourier transform is used. The amplitude signal of each time period is divided into fixed intervals. The amplitude of each interval is squared and accumulated and the result is recorded. Then, the energy ratio is calculated according to the frequency dimension and arranged in chronological order to form a distribution record table to generate an energy distribution sequence set. S22. Based on the energy distribution sequence set, calculate the change amplitude of energy ratio in adjacent frequency bands, label the frequency bands whose amplitude changes exceed the set range as abrupt segments, extract their time index and amplitude fluctuation information and archive them in order to generate an energy abrupt feature set. S23. Based on the energy mutation feature set, each energy change value is converted into a logarithmic value and its weighted total is calculated. Combined with the time axis record, an entropy value sequence is formed. Then, the entropy change rate and the number of frequency fluctuations are cross-matched and organized into matrix entries to establish a perturbation energy feature matrix.

5. The method for evaluating voltage disturbances in wind and solar grid connection according to claim 1, characterized in that: S3 includes: S31. First, based on the perturbation energy feature matrix, Kalman filtering is used to calculate the amplitude difference between adjacent time points and extract the rate change direction. At the same time, the corresponding energy fluctuation interval is recorded. The rate change direction, energy amplitude and time index are combined into a unified recording sequence to generate a perturbation change parameter set. S32. Based on the set of disturbance change parameters, compare the relationship between the rate change direction and energy fluctuation over a continuous time period. When the two directions reverse or the amplitude deviates beyond the threshold, mark the level state change, record each state interval and the number of changes, and generate a disturbance state mapping table. S33. Based on the disturbance state mapping table, calculate the duration of each level state and its proportion, compare and correct the time weight with the frequency of state changes, arrange the probabilities of each level in time order to form a continuous sequence, and generate the disturbance level probability distribution sequence.

6. The method for evaluating voltage disturbances in wind and solar grid connection according to claim 1, characterized in that: S4 includes: S41. Based on the probability distribution sequence of disturbance levels, arrange the level labels in order of time step number, calculate the sign of the difference between adjacent labels to mark the upward and downward directions, count the start and end and duration of continuous same direction, merge adjacent micro-amplitude fluctuation segments, and generate a disturbance level sequence mapping table. S42. Based on the disturbance level sequence mapping table, the output change rate of the reactive power compensation unit is compared with the level direction segment by segment, the same-direction and opposite-direction intervals are marked, the length of each interval and the level span are counted, the strong disturbance and weak disturbance segments are defined and numbered according to the span threshold, and the disturbance stage mapping structure is established.

7. The method for evaluating voltage disturbances in wind and solar grid connection according to claim 1, characterized in that: S5 includes: S51. Based on the disturbance stage mapping structure, compare the difference between the duration and grade rate of adjacent stages, screen out continuous segments with the same direction and mark them for extension, record the stage number and the start and end time of the segment, and generate a trend extension segment index table. S52. Based on the trend extension segment index table, extract the energy fluctuation peak-valley difference sequence of each extension segment and calculate the increment rate over time. Accumulate the values ​​segment by segment to obtain the intensity quantity. Combine the level direction to give the composite symbol. Output the trend strength and direction in time order to generate the disturbance trend evolution result.

8. A wind-solar grid-connected voltage disturbance assessment system, characterized in that: The wind and solar grid-connected voltage disturbance assessment system includes a disturbance signal synchronization module, an energy entropy construction module, a disturbance probability evolution module, a level mapping judgment module, and a trend prediction inference module. The disturbance signal synchronization module is used to collect voltage and perform phase calibration on the voltage channel of the wind turbine and the photovoltaic inverter branch to generate a voltage disturbance timing structure set. The energy entropy construction module is used to calculate the energy ratio and its variation amplitude of different frequency bands based on the voltage disturbance time series structure set, using short-time Fourier transform partitioned statistical amplitude energy, identifying energy mutation intervals based on the variation amplitude and extracting the amplitude fluctuation sequence of the energy mutation intervals to generate a disturbance energy feature matrix. The disturbance probability evolution module is used to track the relationship between disturbance amplitude rate and energy fluctuation based on the disturbance energy feature matrix using Kalman filtering, determine the level state transition conditions, and modify the level weights in combination with the disturbance duration to generate a disturbance level probability distribution sequence. The level mapping judgment module is used to sort out the level change trajectory based on the disturbance level probability distribution sequence, analyze the corresponding direction of the reactive power compensation unit output and the level sequence, determine the strong disturbance and weak disturbance partitions, and generate a disturbance stage mapping structure. The trend prediction and inference module is used to compare the duration of the stage with the rate of change based on the perturbation stage mapping structure, identify the trend extension segment, calculate the energy fluctuation increase rate and determine the direction of change, and generate the perturbation trend evolution result.

9. A wind-solar grid-connected voltage disturbance assessment device, characterized in that: The wind and solar grid-connected voltage disturbance assessment device includes a memory and a processor; the memory is used to store computer program code and transmit the computer program code to the processor; the processor is used to execute the wind and solar grid-connected voltage disturbance assessment method as described in claim 1 according to the instructions in the computer program code.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the wind-solar grid-connected voltage disturbance assessment method as described in claim 1.

Images