New energy field station broadband oscillation risk redisk assessment method based on multivariate data

Abstract

The invention belongs to the technical field of power system risk assessment, and particularly relates to a new energy station broadband oscillation risk redisk assessment method based on multivariate data, which comprises the following steps: performing data acquisition and data preprocessing on a new energy station-power grid coupling system to obtain a historical data set; extracting oscillation characteristic indexes to form index vectors; carrying out oscillation frequency source positioning analysis, calculating a time-frequency spectrum energy function to obtain an oscillation energy spectrum, carrying out main frequency energy track extraction, and introducing a frequency-power cross analysis mechanism to realize accurate identification of an oscillation source device; constructing a source-network coupled transfer function, performing stability margin calculation, and judging the stable state of the coupled system; and based on the index vector and the stability margin, carrying out oscillation risk grade division, and combining a source positioning result to generate an oscillation redisk map. According to the invention, panoramic recovery and risk quantitative evaluation of the broadband oscillation problem of the new energy field station can be realized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of power system risk assessment, and particularly relates to a new energy station wide frequency oscillation risk review evaluation method based on multi-element data. BACKGROUND

[0002] With the rapid development of "double high" characteristic power systems, that is, high proportion of new energy access and high proportion of power electronic equipment application, the operation characteristics of the power grid are becoming increasingly complex. Under this background, the wide frequency oscillation problem has become a prominent problem affecting the safe and stable operation of the power system. During the grid-connected operation of the new energy station, the coupling between the large number of power electronic converters inside the station and the weak power grid often leads to oscillation phenomena with a wide frequency range and complex energy distribution. Such wide frequency oscillation not only covers low frequency mechanical-electrical modes, but also widely exists in the medium frequency and high frequency bands, with characteristics of multi-mode, multi-source, and multi-time and space propagation.

[0003] The current wide frequency oscillation problem presents three significant trends: first, the oscillation frequency band is widely distributed, and oscillation events from 2Hz to thousands of Hz frequently occur; second, the oscillation source is complex, involving new energy controllers, grid interfaces, power electronic devices, impedance coupling, and other multi-factor interactions; third, the propagation path is uncertain, and the oscillation signal may spread through multiple channels in the system, causing local or even systemic instability risks.

[0004] However, existing oscillation risk monitoring methods are mainly real-time detection, focusing on single-point early warning, and lack the ability to analyze systemic, implicit, and evolutionary risks. In the high proportion of new energy access scenario, the oscillation problem often shows strong concealment, complex triggering mechanism, and nonlinear development process, making it difficult for traditional monitoring methods to timely and comprehensively reveal its evolution path and essential mechanism. SUMMARY

[0005] In view of the above deficiencies in the prior art, the present application aims to provide a new energy station wide frequency oscillation risk review evaluation method based on multi-element data, which can realize panoramic review and risk quantitative evaluation of new energy station wide frequency oscillation problems.

[0006] To achieve the above purpose, the present application provides a new energy station wide frequency oscillation risk review evaluation method based on multi-element data, comprising the following steps: S1, for the coupled system of the new energy station and the power grid, data acquisition is performed, including station side data, grid side data and auxiliary data; S2, the collected data is preprocessed to obtain a historical data set; S3, based on the historical data set, oscillation characteristic indexes are extracted, including frequency domain indexes, time domain indexes and statistical indexes, to form an index vector; S4, carry out oscillation frequency source positioning analysis, carry out short-time Fourier transform on high-frequency voltage or high-frequency current signal in the field station side data to generate time-frequency spectrum, calculate time-frequency spectrum energy function, obtain oscillation energy spectrum, carry out main frequency energy track extraction, and introduce frequency-power cross analysis mechanism, realize accurate identification of oscillation source device; S5, based on historical data set, construct source-net coupling transfer function, carry out stability margin calculation, and judge the stable state of the coupling system; S6, based on index vector and stability margin, carry out oscillation risk level division, and generate oscillation review atlas combined with the result of source positioning.

[0007] As a preferred scheme of the application, the data collected in S1 specifically includes: The field station side data includes: High-frequency voltage / current signal: the instantaneous value signal of voltage / current of each measuring point is collected by using voltage transformer and current transformer; Device control parameter: the control parameter of the power electronic device in the new energy station is collected, and the power electronic device includes inverter, converter, static var generator, energy storage converter and DC-DC converter; Power change record: the real-time change of active power and reactive power of the new energy station to the power grid is continuously recorded by the power measuring device of each measuring point; Device state quantity information: the running state information of the power electronic device in the new energy station is collected, including whether it is in the commissioning state, whether it has fault or protection action, and running mode switching condition; The grid-connected side data includes: Public connection point impedance spectrum scanning result: the impedance at the public connection point of the new energy station and the power grid is scanned in wide frequency domain by using impedance measuring device or modeling simulation method, and the impedance amplitude and phase characteristics under different frequencies are obtained; Power flow data: the power flow data of the public connection point and the related line of the power grid are collected, including the power flow distribution of active power, reactive power, voltage and current; Voltage stability parameter: the fluctuation range of voltage amplitude, voltage change rate and voltage stability margin index are collected; Auxiliary data includes: Dispatching instruction record: the dispatching instruction issued by the power grid to the new energy station is recorded, including power regulation instruction and running mode switching instruction; Environmental meteorological data: the environmental meteorological data of the area where the new energy station is located are collected, including wind speed, light intensity and temperature.

[0008] As a preferred scheme of the application, the preprocessing process in S2 includes: Time synchronization, time stamp synchronization is carried out to all acquisition channels, re-sampling processing is carried out by adopting a unified sampling frequency, for high-frequency voltage / current signals, a band-pass filter is adopted to remove noise; An event-triggered working condition slicing mechanism is introduced to schedule instruction mutation, device operation mode switching and wind speed jump as slicing boundaries, and the collected time sequence data is divided into multiple stable operation segments. After data slicing, structure normalization processing is carried out, the dimension and unit of different physical quantities are unified, a standard input data matrix is generated, and a historical data set is obtained.

[0009] As a preferred scheme of the application, in the S3, the frequency domain index includes minimum impedance margin, phase margin, gain margin and harmonic magnification factor, wherein the minimum impedance margin is expressed as: ; In the formula, is the current frequency, and j is the imaginary unit; is the frequency response of the equivalent impedance of the field station side; is the frequency response of the equivalent impedance of the grid-connected side; and min represents the minimum value. Based on the common connection point impedance spectrum scanning result, a Bode diagram is drawn, the source-grid impedance amplitude crossover frequency is found, the difference between the phase angle and 180° at the frequency is calculated as the phase margin, the phase crossover frequency is found, and the difference between 1 and the impedance amplitude ratio at the frequency is calculated as the gain margin. The harmonic magnification factor HMF is defined as: ; In the formula, max represents the maximum value. The time domain index includes main frequency energy proportion and oscillation duration, wherein the main frequency energy proportion is defined as the proportion of signal energy in the dominant frequency band to the total energy: ; In the formula, is the main frequency; is the frequency band width; represents the frequency domain amplitude obtained by performing short-time Fourier transform on the high-frequency voltage or high-frequency current signal; represents the energy spectrum density of the signal at the current frequency . An energy threshold is set, the time when the signal energy first exceeds the energy threshold is identified as the oscillation starting time , the time when the signal energy last time is lower than the energy threshold is identified as the oscillation ending time , and the oscillation duration is defined as: ; The statistical indicators include spectral entropy and waveform distortion rate, wherein the spectral entropy H is defined as: ; In the formula, i and z represent the index of the frequency component, is the energy proportion probability of the i-th frequency component; represents the signal energy on the i-th frequency component; represents the signal energy on the z-th frequency component.

[0010] As a preferred scheme of the present application, in the S4, the process of obtaining the oscillation energy spectrum and extracting the main frequency energy track is: S4.1, performing short-time Fourier transform on the high-frequency voltage or high-frequency current signal to obtain a time-frequency spectrum ; S4.2, calculating a time-frequency spectrum energy function , converting the complex-valued signal in the time-frequency domain into an energy distribution, The value of the energy distribution indicates the energy intensity of the corresponding time t and the current frequency , taking t as the horizontal axis, as the vertical axis, mapping the value of the energy distribution to color or gray scale to form an oscillation energy spectrum, and the peak track of the oscillation energy spectrum indicates the excitation and decay process of the frequency component with time; S4.3, extracting the main frequency energy track from the oscillation energy spectrum to determine the oscillation starting time, the main frequency energy proportion, the energy growth stage, and the energy decay stage.

[0011] As a preferred scheme of the present application, in the S4, the process of introducing a frequency-power cross analysis mechanism to realize accurate identification of the oscillation source device is: S4.4, jointly analyzing the power change rate of each measuring point and the main frequency component intensity change , wherein P represents the power, represents the main frequency; using the Pearson correlation coefficient to quantify the linear correlation degree r between and , setting a correlation threshold, when is greater than the correlation threshold, it is preliminarily determined that the power electronic equipment associated with P and is related to the oscillation, and it is classified as a suspected oscillation source equipment; S4.5, based on the load response behavior before and after the power data acquisition, combining the device control parameters, the device state quantity information and the load response behavior before and after the acquisition, excluding the power electronic equipment which is only a response source, and accurately positioning the oscillation source equipment.

[0012] ​As a preferred scheme of the present application, in the S4.5, if the following conditions are met at the same time, it is determined that the power electronic device is only a response source, and is excluded: Load response time sequence: the power change of the post-load lags behind the pre-load, and the power change trend of the pre-load is consistent with the power change trend of the power electronic device; Device control parameters: the control strategy is a passive following type, and the control parameters are within the normal range; Device state quantity: the device is in a normal operation and full-power stable operation state.

[0013] As a preferred scheme of the present application, in the S5, the transfer function of the source-grid coupling is constructed, the stability margin is calculated, and the process of determining the stable state of the coupling system is as follows: S5.1, the impedance relationship at the interface of the new energy station and the power grid is simplified as: ; In the formula, is the transfer function of the source-grid coupling; S5.2, if in a certain frequency band and , the coupling system has a stability threat, wherein arg represents the argument of a complex number; S5.3, the stability margin is calculated, the stability margin includes the phase margin, the gain margin and the harmonic magnification factor, when the HMF is greater than the set empirical threshold, the coupling system has a stability threat; S5.4, based on and curves, a Nyquist diagram is drawn, and the relative position of the trajectory of the transfer function and the point of -1+j0 is observed, if the trajectory encloses the point, the coupling system has a stability threat.

[0014] As a preferred scheme of the present application, in the S6, the oscillation risk level is divided into three categories: low risk, medium risk and high risk: High risk: simultaneously satisfying the phase margin less than 10°, HMF greater than 0.9, less than 1.2, greater than 70%, H greater than 0.8, more than 3 seconds; Medium risk: at least one of the following conditions exists, but not all of them are met: the phase margin is less than 10°, the HMF is greater than 0.9, less than 1.2, greater than 70%, H greater than 0.8, more than 3 seconds; Low risk: none of the following conditions exists: the phase margin is less than 10°, the HMF is greater than 0.9, less than 1.2, more than 70%, H more than 0.8, more than 3 seconds.

[0015] As a preferred scheme of the application, in the S6, the oscillation complex atlas comprises: a main frequency evolution trajectory diagram, which shows the trend of the main frequency frequency changing over time based on a frequency domain index; a frequency domain amplitude-phase margin curve diagram, which is based on a Bode diagram and marks a crossover point and a margin value; a time domain oscillation waveform diagram, which is based on a high-frequency voltage or high-frequency current signal and superimposes oscillation start / end time; a device state and control response superposition diagram, which is based on device control parameters and device state quantity information and superimposes an oscillation source positioning result to show the correlation between power electronic device state changes and oscillation; a risk level spatial distribution diagram, which marks the risk levels of various power electronic devices in the area where the new energy station is located.

[0016] The application has the beneficial effects that: The application constructs a full-process, multi-dimensional systematic evaluation system, breaks through the limitations of traditional real-time detection, integrates high-frequency voltage and current, impedance scanning, device state, weather and other multi-source historical data, forms a unified analysis basis through standardized preprocessing, and then constructs a complete index system from the frequency domain (such as impedance margin, phase margin), the time domain (such as main frequency energy proportion, oscillation duration), and the statistical domain (such as spectral entropy, distortion rate), which can realize the comprehensive quantification of oscillation characteristics, solve the fragmentation and localization problems of traditional methods, and the core index has clear physical meaning and clear safety boundary, which is convenient for engineering landing and standardized promotion.

[0017] The application uses time-frequency methods (short-time Fourier transform STFT), combines power changes and device states, can accurately locate the oscillation source start time, propagation path and suspected devices, restores the whole chain of oscillation evolution, and through the source-network impedance model and Bode diagram, Nyquist diagram and other tools to invert the stability margin, can identify the risk of the coupled system approaching the stability boundary in advance. On this basis, through fuzzy logic and threshold mapping, the risk level is divided, and a visual complex atlas including the main frequency trajectory, device response, etc. can be generated, which can directly support operation management, parameter setting and other scenes, significantly improve the grid control and decision-making efficiency, and can adapt to different types of new energy stations and access modes, and has strong adaptability. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 is a flow principle diagram of the application. DETAILED DESCRIPTION

[0019] The embodiments of the present invention will be further described below with reference to the accompanying drawings: Example 1: As Figure 1 As shown, the method for reviewing and assessing the broadband oscillation risk of new energy power plants based on multivariate data includes the following steps: S1. For the coupled system of new energy power plant and power grid, data is collected, including data from the power plant side, data from the grid connection side, and auxiliary data; S2. Preprocess the collected data to obtain the historical dataset; S3. Extract oscillation characteristic indicators based on historical datasets, including frequency domain indicators, time domain indicators, and statistical indicators, to form an indicator vector; S4. Perform oscillation frequency source localization analysis, perform short-time Fourier transform on the high-frequency voltage or high-frequency current signal in the station side data to generate time spectrum, calculate the time spectrum energy function, obtain oscillation energy spectrum, extract the main frequency energy trajectory, and introduce a frequency-power cross-analysis mechanism to achieve accurate identification of oscillation source device; S5. Based on historical datasets, construct the source-network coupling transfer function, calculate the stability margin, and determine the stable state of the coupled system. S6. Based on the indicator vector and stability margin, classify the oscillation risk level and generate an oscillation review map by combining the source location results.

[0020] In S1, the collected data specifically includes: Station-side data includes: High-frequency voltage / current signals: Using voltage transformers and current transformers, instantaneous voltage / current signals at each measuring point are acquired; Equipment control parameters: Collect control parameters of power electronic equipment inside the new energy power station. Power electronic equipment includes inverters, converters, static var generators, energy storage converters, and DC-DC converters. Power change recording: Through power measurement devices at each measuring point, the real-time changes in active and reactive power transmitted from new energy power plants to the grid are continuously recorded; Equipment status information: Collect the operating status information of power electronic equipment inside the new energy power station, including whether it is in operation, whether a fault or protection action has occurred, and the status of operation mode switching; Grid-connected data includes: Impedance spectrum scan results at the point of common coupling: Using impedance measurement equipment or modeling and simulation methods, the impedance at the point of common coupling between the new energy power station and the power grid is scanned in a wide frequency domain to obtain the impedance amplitude and phase characteristics at different frequencies. Power flow data: Collect power flow data at the point of common coupling and related power grid lines, including the distribution of active power, reactive power, voltage, and current. Voltage stability parameters: the fluctuation range of voltage amplitude, voltage rate of change, voltage stability margin index; Auxiliary data is used to assist in identifying disturbance trigger background and oscillation development boundary, including: Dispatching instruction record: record the dispatching instructions issued by the power grid to the new energy station, including power regulation instructions, operation mode switching instructions; Environmental meteorological data: collect environmental meteorological data in the area where the new energy station is located, including wind speed, light intensity, temperature.

[0021] In S2, due to the above multi-source data with different sampling frequencies, large differences in physical quantity dimensions, and non-uniform units, format conversion and standardization processing are needed, and the preprocessing process includes: Time synchronization, synchronize the time stamp of all collection channels, use uniform sampling frequency for resampling processing, and use band-pass filter (20Hz~5kHz range) for noise removal for high-frequency voltage / current signals; Introduce the working condition slicing mechanism based on event triggering, take the mutation of dispatching instructions, the switching of device operation mode, and the jump of wind speed as the slicing boundary, divide the collected time series data (various dynamic data collected continuously in time sequence and related to the operation of new energy stations and grid-connected power grids) into multiple stable running segments; After data slicing, structure normalization processing is performed to unify the dimensions and units of different physical quantities, generate standard input data matrix, and obtain historical data set, which can be expressed as: ; In the formula, is the historical data set at time t; is the high-frequency voltage signal at time t; is the high-frequency current signal at time t; is the impedance frequency response at time t; is the dispatching instruction sequence at time t; is the device state quantity information at time t; is the environmental meteorological data at time t.

[0022] In S3, the frequency domain index includes minimum impedance margin, phase margin, gain margin and harmonic amplification factor, wherein the minimum impedance margin is expressed as: ; In the formula, is the current frequency (frequency range that may be involved in wide frequency oscillation), and j is the imaginary unit; is the frequency response of the equivalent impedance on the station side (reflecting the impedance characteristics of the entire source side from the point of view of the new energy station grid connection point); The frequency response of the equivalent impedance on the grid side (reflecting the impedance characteristics of the grid system from the grid point of the new energy station); min represents the minimum value; subscript represents all possible , the value of is calculated, and the minimum value is selected as ; When and tend to be equal or the modulus is close, the coupling system stability margin decreases, and wideband oscillation caused by control interaction is easily excited.

[0023] Based on the common connection point impedance frequency spectrum scanning result, a Bode diagram (a kind of diagram method of frequency response) is drawn, the source-grid impedance amplitude crossover frequency is found, the difference between the phase angle at this frequency and 180° is calculated as the phase margin; the phase crossover frequency is found, and the difference between 1 and the impedance amplitude ratio at this frequency is calculated as the gain margin; The phase margin is the phase angle distance at the source-grid impedance amplitude crossover frequency, reflecting the phase stability of the coupling system to small signal disturbance; the gain margin measures the margin of the coupling system when the open loop gain reaches unity gain, both of which are intuitively represented as frequency boundaries of system stability in the Bode diagram.

[0024] The harmonic magnification factor HMF is defined as: ; In the formula, max represents the maximum value; subscript represents all possible , the value of is calculated, and the maximum value is selected as HMF; When the coupling system has high-order control loop resonance, nonlinear saturation or filter parameter mismatch, HMF can significantly increase, becoming an important precursor of oscillation risk.

[0025] The time domain indicators include the dominant frequency energy ratio and the oscillation duration, wherein the dominant frequency energy ratio (Energy Ratio of Dominant Frequency) is defined as the proportion of signal energy in the dominant frequency band to the total energy: ; In the formula, is the dominant frequency (the frequency of dominant oscillation); is the frequency band width; represents the frequency domain amplitude obtained by short-time Fourier transform on the high-frequency voltage or high-frequency current signal; represents the signal at the current frequency Energy spectral density at; A higher value indicates stronger oscillation concentration; conversely, a lower value may indicate broadband distribution or multimodal coupled oscillations. The oscillation duration is determined by the signal energy threshold detection window, reflecting the time span of oscillation behavior in a certain frequency band, which is of great guiding significance in determining whether intervention is necessary.

[0026] Set an energy threshold and identify the moment when the signal energy first exceeds the energy threshold as the oscillation start moment. The moment when the signal energy last falls below the energy threshold is identified as the end of the oscillation. Oscillation duration Defined as: ; Statistical metrics include spectral entropy and waveform distortion rate (for high-frequency voltage / current), where spectral entropy H is defined as: ; In the formula, i and z represent the indices of the frequency components. The probability of the energy proportion of the i-th frequency component; This represents the signal energy at the i-th frequency component; This represents the signal energy at the z-th frequency component.

[0027] H reflects the degree of concentration of the frequency distribution. The larger the value, the more discrete the frequency distribution of the coupled system, the more modes it has, and the greater the difficulty of analysis and control. Combining frequency skewness, waveform distortion rate (THD, an existing indicator), and other indicators can further identify complex phenomena such as multi-frequency oscillations and controller coupling.

[0028] All indicators constitute an indicator vector within each operating condition segment, serving as input support for subsequent oscillation source location, risk level assessment, and retrospective chart construction.

[0029] In S4, the process of obtaining the oscillation energy spectrum and extracting the main frequency energy trajectory is as follows: S4.1 Perform a short-time Fourier transform on the high-frequency voltage or high-frequency current signal to obtain the time spectrum. ; S4.2 Calculation of the time-frequency energy function This converts complex-valued signals in the time-frequency domain into energy distributions. The value indicates the current frequency at the corresponding time t. The strength of energy, plotted on the horizontal axis as t. Using the vertical axis, The values ​​are mapped to colors or grayscale to form an oscillating energy spectrum, whose peak trajectory indicates the excitation and decay process of frequency components over time. S4.3, extract the main frequency energy trajectory from the oscillation energy spectrum, determine the oscillation starting time, main frequency energy proportion, energy growth stage, and energy decay stage.

[0030] The introduction of frequency-power cross analysis mechanism realizes the accurate identification of the oscillation source device process is: S4.4, the power change rate of each measuring point and the main frequency component intensity change Joint analysis, where P represents power, The main frequency is represented by Pearson correlation coefficient to quantify The linear correlation degree r of , set the correlation threshold, when Greater than the correlation threshold, preliminary determine the power electronic equipment associated with P and Participate in the oscillation, and it is classified as a suspected oscillation source device; S4.5, based on the power data acquisition before and after the load response behavior, combined with the device control parameters, device state quantity information and load response behavior before and after, exclude the power electronic equipment only as a response source, so as to accurately locate the oscillation source equipment.

[0031] If the following conditions are met at the same time, it is determined that the power electronic equipment is only a response source, which is excluded: Load response timing: the power change of the post-load lags behind the pre-load, and the power change trend of the pre-load is consistent with the power change trend of the power electronic equipment; Device control parameters: control strategy is passive following type, and control parameters are in normal range; Device state quantity: the device is in normal operation, full power stable operation state.

[0032] In addition, combined with the scale-time representation in wavelet analysis, transient oscillation and wide frequency sweep phenomenon can be located. Especially in the presence of multi-frequency coupling or controller switching, wavelet packet decomposition can show clearer modal overlap and amplitude mutation, effectively supporting the source identification and propagation trend restoration in complex oscillation scenarios.

[0033] In S5, the transfer function of source-grid coupling is constructed, the stability margin is calculated, and the process of judging the stability state of the coupled system is: S5.1, the impedance relationship at the interface of new energy station and power grid is simplified as: ; In the formula, The transfer function of source-grid coupling is S5.2, if Satisfy And If , then the coupled system is under a stability threat, where arg represents the argument of the complex number; S5.3 Perform stability margin calculation. The stability margin includes phase margin, gain margin and harmonic amplification factor. When the HMF is greater than the set empirical threshold (0.9), the coupled system is under stability threat. S5.4, based on and Plot the curve, draw a Nyquist plot (a frequency response diagram of a linear control system), and observe the transfer function. The relative position of the trajectory and the point -1+j0 (a key point on the complex plane, which can also be expressed as the point (-1,0)) indicates that the coupled system is under a stability threat if the trajectory encloses the point (indicating that stability indicators such as phase margin do not meet the requirements).

[0034] In S6, the oscillation risk level is divided into three categories: low risk, medium risk, and high risk. High risk: Simultaneously meeting the requirements of phase margin less than 10° and HMF greater than 0.9. Less than 1.2 Greater than 70%, H greater than 0.8 More than 3 seconds; Medium risk: Phase margin less than 10°, HMF greater than 0.9, Less than 1.2 Greater than 70%, H greater than 0.8 At least one of the conditions that takes longer than 3 seconds is met, but not all of them are satisfied; Low risk: No phase margin less than 10°, HMF greater than 0.9, Less than 1.2 Greater than 70%, H greater than 0.8 Any of the following conditions that take longer than 3 seconds.

[0035] The oscillation risk level assessment mechanism enables automatic determination and visual review of the oscillation level of new energy power plants under different scenarios. By comprehensively considering frequency domain indicators, time domain indicators, and statistical indicators, it classifies and identifies oscillations based on tiered thresholds, quantifying the oscillation risk level of the power plant under specific operating conditions. For example, in a certain scenario, if the spectrum concentration is extremely high and the main frequency band matches the open-loop frequency of the equipment controller, and the frequency source location indicates that the oscillation originates in critical equipment, then it can be comprehensively assessed as a high-risk operating condition, requiring timely attention or mitigation measures.

[0036] The oscillation recap chart includes: The main frequency evolution trajectory diagram, based on frequency domain indicators, shows the trend of the main frequency changing over time; Frequency domain amplitude-phase margin curve: based on Bode plot, with crossover points and margin values ​​marked; Time-domain oscillation waveform chart: based on high-frequency voltage or high-frequency current signals, and superimposed with oscillation start / end time; Device state and control response superposition chart: based on device control parameters and device state quantity information, and superimposed with oscillation source positioning results, showing the correlation between power electronic device state changes and oscillation; Risk level spatial distribution chart: marking the risk level of each power electronic device in the area where the new energy station is located.

[0037] Under multi-point measurement conditions, the oscillation review chart can also reflect the oscillation propagation path. For example, by the order of frequency component excitation between measurement points, an oscillation propagation time sequence chain can be constructed, and spatial mapping can be performed in combination with device positions, and finally the full-chain reproduction of the oscillation event “from where, how to diffuse, when to end” can be realized. These review information has high reference value for subsequent control strategy modification, device parameter optimization and dispatching mode adjustment.

[0038] The method of the embodiment can be applied to a new energy station, a regional control center or a power grid operation management platform, to realize real-time data archiving and periodic review evaluation, and to enhance the foresight and scientificity of power system oscillation prevention and control.

[0039] A typical broadband oscillation signal is simulated, and based on the method of the embodiment, the broadband oscillation offline evaluation table generated is shown in Table 1. It can be seen that the method of the embodiment successfully identifies the main oscillation components, evaluates the risk of each frequency band, and points out the potential optimization direction. This set of process and method can be applied to the analysis and diagnosis of complex oscillation problems in actual engineering systems.

[0040] Table 1: Broadband oscillation offline evaluation table

[0041] Embodiment 2: On the basis of embodiment 1, a structured auxiliary decision evaluation report is further generated, including oscillation event occurrence time, dominant frequency interval, index overrun information, suspected source device location, stability margin change trend, oscillation risk level determination conclusion and optimization suggestions for the station or control department.

[0042] The auxiliary decision evaluation report supports a variety of engineering scene applications such as grid connection evaluation, operation and maintenance governance, control parameter setting and control strategy formulation, and has strong practical guidance value.

[0043] Embodiment 3: A new energy station broadband oscillation risk review evaluation device based on multi-element data, comprising: one or more processors; a memory for storing one or more computer programs; Wherein the one or more programs, when executed by one or more processors, cause the one or more processors to perform the method of embodiment 1 or embodiment 2.

[0044] Embodiment 4: A computer-readable storage medium having stored thereon executable instructions that, when executed by a processor, cause the processor to perform the method of embodiment 1 or embodiment 2.

Claims

1. A method for wide frequency oscillation risk review and evaluation of a new energy station based on multi-element data, characterized in that The method comprises the following steps: S1, for the coupling system of new energy station-grid, data acquisition is carried out, including station side data, grid side data and auxiliary data; S2, the collected data is preprocessed to obtain a historical data set; S3, based on the historical data set, the oscillation characteristic index is extracted, including frequency domain index, time domain index and statistical index, and an index vector is formed; S4, the oscillation frequency source positioning analysis is carried out, the high-frequency voltage or high-frequency current signal in the station side data is subjected to short-time Fourier transform to generate a time-frequency spectrum, the time-frequency spectrum energy function is calculated, the oscillation energy spectrum is obtained, the main frequency energy track is extracted, and the frequency-power cross analysis mechanism is introduced to realize accurate identification of the oscillation source device; S5, based on the historical data set, the transfer function of the source-grid coupling is constructed, the stability margin is calculated, and the stability state of the coupling system is judged; S6, based on the index vector and the stability margin, the oscillation risk level is divided, and the oscillation review atlas is generated in combination with the source positioning result.

2. The method of claim 1, wherein the method further comprises: The data collected in S1 specifically includes: The station side data includes: High-frequency voltage / current signal: the instantaneous value signal of voltage / current of each measuring point is collected by using a voltage transformer and a current transformer; Device control parameter: the control parameter of the power electronic device in the new energy station is collected, and the power electronic device includes an inverter, a converter, a static var generator, a energy storage converter and a DC-DC converter; Power change record: the real-time change of active power and reactive power of the new energy station to the grid is continuously recorded by the power measuring device of each measuring point; Device state quantity information: the running state information of the power electronic device in the new energy station is collected, including whether it is in the operation state, whether it has a fault or protection action, and the running mode switching condition; The grid side data includes: Common connection point impedance spectrum scanning result: the impedance at the common connection point of the new energy station and the grid is scanned in a wide frequency domain by using an impedance measuring device or a modeling simulation method, and the impedance amplitude and phase characteristics at different frequencies are obtained; Power flow data: the power flow data of the common connection point and the related lines of the grid are collected, including the power flow distribution of active power, reactive power, voltage and current; Voltage stability parameter: the fluctuation range of voltage amplitude, voltage change rate and voltage stability margin index are collected; The auxiliary data includes: Dispatching instruction record: the dispatching instruction issued by the grid to the new energy station is recorded, including the power regulation instruction and the running mode switching instruction; Environmental meteorological data: the environmental meteorological data of the area where the new energy station is located are collected, including wind speed, light intensity and temperature.

3. The method of claim 2, wherein the method further comprises: In S2, the preprocessing process includes: Time synchronization: all collected channels are synchronized by time stamp, resampling is carried out by using a unified sampling frequency, and for high-frequency voltage / current signal, a band-pass filter is used for noise removal; An event-triggered working condition slicing mechanism is introduced, the dispatching instruction mutation, the device running mode switching and the wind speed jump are taken as the slicing boundary, and the collected time series data is divided into multiple stable running segments; After data slicing, structure normalization processing is performed to unify the dimensions and units of different physical quantities, and a standard input data matrix is generated to obtain a historical data set.

4. The method of claim 2, wherein the method further comprises: In the S3, the frequency domain indicators include a minimum impedance margin, a phase margin, a gain margin, and a harmonic amplification factor, wherein the minimum impedance margin is represented as: ； wherein is the current frequency, j is the imaginary unit; is the frequency response of the equivalent impedance on the site side; is the frequency response of the equivalent impedance on the grid side; min denotes the minimum. Based on the public connection point impedance frequency spectrum scanning result, a Bode diagram is drawn, the source-grid impedance amplitude crossover frequency is found, the difference between the phase angle and 180° at this frequency is calculated as the phase margin, the phase crossover frequency is found, and the difference between 1 and the impedance amplitude ratio at this frequency is calculated as the gain margin; The harmonic magnification factor HMF is defined as: ； In the formula, max represents taking the maximum value; The time domain indicators include a main frequency energy proportion and an oscillation duration. The main frequency energy proportion is defined as a proportion of signal energy in a dominant frequency band to total energy. defined as a proportion of signal energy in a dominant frequency band to total energy. ； wherein is the main frequency frequency; is the band width; denotes the frequency domain amplitude obtained by short-time Fourier transformation of the high-frequency voltage or high-frequency current signal; denotes the energy spectral density of the signal at the current frequency f. An energy threshold is set, and the moment when the signal energy first exceeds the energy threshold is identified as the oscillation start moment , and the moment when the signal energy last falls below the energy threshold is identified as the oscillation end moment , and the oscillation duration is defined as: ； The statistical indicators include spectral entropy and waveform distortion rate, wherein the spectral entropy H is defined as: ； In the formula, i, z represent the index of the frequency component, The energy proportion probability of the i-th frequency component; The signal energy on the i-th frequency component is represented by Ei. The signal energy on the z-th frequency component is represented by Ez.

5. The method of claim 2, wherein the method further comprises: In the S4, the process of obtaining the oscillation energy spectrum and extracting the main frequency energy trajectory is: S4.1, short-time Fourier transform of the high-frequency voltage or high-frequency current signal, resulting in a time-frequency spectrum ; S4.2, calculate time-frequency spectrum energy function , convert the complex-valued signal in time-frequency domain to energy distribution, , the value size represents the energy intensity of the corresponding time t, current frequency , with t as the horizontal axis, , the vertical axis, the value of , map the value of to color or grayscale to form an oscillation energy spectrum, and the peak value trajectory indicates the excitation and decay process of the frequency component over time; S4.3, extracting the main frequency energy trajectory from the oscillation energy spectrum, determining the oscillation starting time, main frequency energy proportion, energy growth stage and energy decay stage.

6. The method of claim 5, wherein the method further comprises: In the S4, the process of introducing a frequency-power cross analysis mechanism to realize accurate identification of the oscillation source device is: S4.4, the power change rate of each measuring point with the main frequency component intensity change joint analysis, wherein P represents power, represent the main frequency; using Pearson correlation coefficient to quantify with linear correlation degree r, set the correlation threshold, when greater than the correlation threshold, preliminary determination of the power electronic equipment associated with P and related to the oscillation, which is classified as a suspected oscillation source equipment; S4.5, based on the load response behavior before and after the power data are acquired, combining the device control parameters, device state quantity information and load response behavior before and after, excluding the power electronic device which is only a response source, and accurately positioning the oscillation source device.

7. The method of claim 6, wherein the method further comprises: In the S4.5, if the following conditions are met at the same time, it is determined that the power electronic device is only a response source and is excluded: Load response timing: the post-load power change lags behind the pre-load, and the power change trend of the pre-load is consistent with the power change trend of the power electronic device; Device control parameters: the control strategy is passive following type, and the control parameters are within the normal range; Device state quantity: the device is in normal operation, full power stable operation state.

8. The method of claim 4, wherein the method further comprises: In the S5, the transfer function of the source-grid coupling is constructed, the stability margin is calculated, and the stability state of the coupling system is determined: S5.1, the impedance relationship at the interface between the new energy station and the power grid is simplified as: ； wherein is the transfer function of the source-net coupling; S5.2, if Under certain frequency bands and arg (S5.1) < arg (S5.2), then the coupled system is at risk of instability, where arg denotes the argument of a complex number. S5.3, stability margin calculation, the stability margin includes phase margin, gain margin and harmonic magnification factor, when HMF is greater than the set empirical threshold, the coupling system has a stability threat; S5.4, based on and Plotting the Nyquist plot, the transfer function is observed The relative position of the trajectory to the -1+j0 point, if the trajectory encloses this point, the coupled system is threatened by stability.

9. The method for reviewing and assessing the broadband oscillation risk of new energy power plants based on multivariate data as described in claim 4, characterized in that, In the S6, the oscillation risk level is divided into low risk, medium risk and high risk: High risk: both phase margin less than 10°, HMF greater than 0.9, less than 1.2, greater than 70%, H greater than 0.8, greater than 3 seconds; Medium risk: at least one of the following conditions is met, but not all: phase margin less than 10°, HMF greater than 0.9, less than 1.2, more than 70%, H more than 0.8, more than 3 seconds. Low risk: None of the phase margin less than 10°, HMF greater than 0.9, less than 1.2, greater than 70%, H greater than 0.8, any of greater than 3 seconds.

10. The method of claim 9, wherein the method further comprises: In the S6, the oscillation risk review atlas includes: Main frequency evolution trajectory graph: based on frequency domain indicators, showing the trend of main frequency frequency change over time; Frequency domain amplitude-phase margin curve: based on Bode diagram, marking the crossover point and margin value; Time domain oscillation waveform graph: based on high frequency voltage or high frequency current signal, and superimposed with oscillation starting / ending time; Device state and control response superposition graph: based on device control parameters and device state quantity information, and superimposed with oscillation source positioning results, showing the correlation between power electronic device state change and oscillation; Risk level spatial distribution graph: marking the risk level of each power electronic device in the area where the new energy station is located.

Images