Scheduling fault analysis and identification method for digital smart power grid

Abstract

The invention relates to the technical field of data processing, in particular to a dispatching fault analysis and identification method for a digital smart power grid. The method comprises the steps of firstly collecting electric signals in an intelligent power grid to obtain initial electric signals; then, constructing an initial signal oscillogram, screening extreme points containing harmonic signals as target signals, and grouping the target signals to obtain target signal groups; obtaining an abnormal value of the target signal group according to the frequency characteristic and the fluctuation characteristic of the target signal; screening according to the abnormal value to obtain an invalid target signal group, and removing the invalid target signal group in the initial electric signal to obtain an effective electric signal; and finally, obtaining an EMD decomposition result of the effective electric signal, and carrying out fault identification on the intelligent power grid according to the EMD decomposition result. According to the method, the interference of harmonic signals on the EMD decomposition algorithm in the fault identification process is reduced, the integrity of fault information is reserved, and the accuracy of the EMD decomposition algorithm in the fault identification process is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method for analyzing and identifying dispatching faults in a digital smart grid. Background Art

[0002] Grid dispatching refers to the process of unified management and control of power generation, transmission, distribution, and power consumption in a power system to ensure the safe, stable, and economic operation of the power system. The main purpose of grid dispatching is to balance supply and demand, optimize resource allocation, improve power quality, and ensure reliable power supply in the power system. The application of smart grid technology can improve the work efficiency and quality of power supply enterprises, reduce the burden on staff, and promote the construction and development of power dispatching automation systems.

[0003] During the process of grid dispatching, there will be dispatching fault problems. Grid dispatching faults often obtain fault information by analyzing and processing the detected grid signals. Since EMD can automatically decompose signals according to the actual characteristics of the signals without presetting signal models, and can reveal the local characteristics of the signals and obtain the time-frequency representation of the signals, it can quickly and accurately identify various faults. Therefore, traditional methods generally use EMD to analyze and process fault problems during grid dispatching. EMD decomposes grid signals into multiple IMF components, and fault information is usually stored in high-frequency IMF components. However, in fact, due to the existence of interference, there will be harmonics in the grid signals, and the harmonics will also be stored in the high-frequency IMF components, resulting in harmonic interference when analyzing fault information, and reducing the accuracy and efficiency of EMD decomposition for fault analysis and identification.

[0004] To solve the above problems, existing technologies generally adopt methods of not processing or directly removing harmonics. However, not processing will result in harmonic interference when EMD analyzes fault information, and directly removing all harmonics will also remove some fault signals that overlap with harmonic signals, resulting in the loss of key data information in the original signal, which also affects the accuracy and precision of the final fault signal identification.

[0005] Therefore, how to remove unnecessary harmonic interference and improve the accuracy of the EMD decomposition algorithm during the process of fault identification has become an urgent problem to be solved. Summary of the Invention

[0006] In view of this, embodiments of the present invention provide a method for analyzing and identifying dispatching faults in a digital smart grid to solve the problem of how to remove unnecessary harmonic interference and improve the accuracy of the EMD decomposition algorithm during the process of fault identification.

[0007] Embodiments of the present invention provide a method for analyzing and identifying dispatching faults in a digital smart grid, and the method includes the following steps:

[0008] Collect the electrical signals in the smart grid to obtain the initial electrical signals;

[0009] Construct the initial signal waveform diagram of the initial electrical signals, where the abscissa of the initial signal waveform diagram is time and the ordinate is frequency. Screen the extreme points containing harmonic signals as target signals in the initial signal waveform diagram, and group the target signals in the initial signal waveform diagram according to the distribution characteristics of the target signals in the initial signal waveform diagram to obtain at least two target signal groups;

[0010] For any target signal group, obtain the first anomaly coefficient of the target signal group according to the frequency characteristics of the target signals in the target signal group, obtain the second anomaly coefficient of the target signal group according to the fluctuation characteristics of the target signals in the target signal group, and obtain the anomaly value of the target signal group according to the first anomaly coefficient and the second anomaly coefficient of the target signal group;

[0011] Obtain the anomaly values of each target signal group, screen out the invalid target signal groups from all target signal groups according to the anomaly values of each target signal group, remove the invalid target signal groups from the initial electrical signals to obtain the effective electrical signals, obtain the EMD decomposition result of the effective electrical signals, and perform fault identification on the smart grid according to the EMD decomposition result.

[0012] Preferably, the step of grouping the target signals in the initial signal waveform diagram according to the distribution characteristics of the target signals in the initial signal waveform diagram to obtain at least two target signal groups includes:

[0013] Form all the target signals within a continuous time into a target signal group according to the time of the target signals in the initial signal waveform diagram to obtain at least two target signal groups.

[0014] Preferably, the step of obtaining the first anomaly coefficient of the target signal group according to the frequency characteristics of the target signals in the target signal group includes:

[0015] Respectively obtain the remainders obtained by dividing the frequency of each target signal in the target signal group by the fundamental frequency, correspondingly obtain the remainder mean value, obtain the first addition result of the remainder mean value and a preset value, and perform normalization processing on the reciprocal of the first addition result to obtain the first anomaly coefficient of the target signal group.

[0016] Preferably, the step of obtaining the second anomaly coefficient of the target signal group according to the fluctuation characteristics of the target signals in the target signal group includes:

[0017] Obtain the peak signals and trough signals in the target signal group, sort them separately in chronological order, and obtain the peak numbers of each peak signal and the trough numbers of each trough signal in the target signal group;

[0018] Obtain the phases of each peak signal and each trough signal in the target signal group. According to the difference between the phase of the peak signal in the target signal group and the theoretical peak phase, and the difference between the phase of the trough signal in the target signal group and the theoretical trough phase, obtain the lateral overall stability index of the target signal group;

[0019] According to the amplitude difference between the peak signal and the trough signal in the target signal group, obtain the longitudinal overall stability index of the target signal group;

[0020] Construct the target curve of the target signal group, divide the target curve into at least two sub-curves, obtain the curvature of each sub-curve, and obtain the smoothness index of the target curve according to the curvature of each sub-curve in the target curve;

[0021] According to the smoothness index, the lateral overall stability index and the longitudinal overall stability index, obtain the second anomaly coefficient of the target signal group.

[0022] Preferably, the obtaining the lateral overall stability index of the target signal group according to the difference between the phase of the peak signal in the target signal group and the theoretical peak phase, and the difference between the phase of the trough signal in the target signal group and the theoretical trough phase includes:

[0023] For any peak signal in the target signal group, obtain the difference between the phase of the peak signal and the theoretical peak phase to get the first difference, calculate the ratio of the first difference to the circumferential angle to get the first ratio, obtain the difference between the peak number of the peak signal and the constant 1 to get the second difference, calculate the difference between the first ratio and the second difference to get the phase difference of the peak signal, and calculate the phase differences of each peak signal in the target signal group respectively to correspondingly obtain the average value of the phase differences of the peak signals;

[0024] For any trough signal in the target signal group, obtain the difference between the phase of the trough signal and the theoretical trough phase to get the third difference, calculate the ratio of the third difference to the circumferential angle to get the second ratio, obtain the difference between the trough number of the trough signal and the constant 1 to get the fourth difference, calculate the difference between the second ratio and the fourth difference to get the phase difference of the trough signal, and calculate the phase differences of each trough signal in the target signal group respectively to correspondingly obtain the average value of the phase differences of the trough signals;

[0025] Obtain the sum of the mean phase difference of the peak signals, the mean phase difference of the valley signals, and a preset value to obtain a second addition result, and obtain the reciprocal of the second addition result as the lateral overall stability index of the target signal group.

[0026] Preferably, obtaining the longitudinal overall stability index of the target signal group according to the amplitude difference between the peak signal and the valley signal in the target signal group includes:

[0027] Respectively obtain the frequency difference between each peak signal and the valley signal with the same serial number in the target signal group to correspondingly obtain a frequency difference sequence, respectively obtain the absolute value of the difference between every two adjacent frequency differences in the frequency difference sequence to correspondingly obtain the mean value of the absolute value of the difference, obtain the sum of the mean value of the absolute value of the difference and a preset value to obtain a third addition result, and obtain the reciprocal of the third addition result as the longitudinal overall stability index of the target signal group.

[0028] Preferably, obtaining the smoothness index of the target curve according to the curvature of each sub-curve in the target curve includes:

[0029] According to the curvature of each sub-curve in the target curve, obtain the mean curvature of the target curve, and obtain the reciprocal of the mean curvature as the smoothness index of the target curve.

[0030] Preferably, obtaining the second abnormality coefficient of the target signal group according to the smoothness index, the lateral overall stability index, and the longitudinal overall stability index includes:

[0031] Obtain the mean value of the lateral overall stability index and the longitudinal overall stability index to obtain a first mean value, and obtain the first mean value and the mean value of the smoothness index as the second abnormality coefficient of the target signal group.

[0032] Preferably, obtaining the abnormal value of the target signal group according to the first abnormality coefficient and the second abnormality coefficient of the target signal group includes:

[0033] Obtain the mean value of the first abnormality coefficient and the second abnormality coefficient of the target signal group as the abnormal value of the target signal group.

[0034] Preferably, screening out invalid target signal groups from all target signal groups according to the abnormal value of each target signal group includes:

[0035] Set an abnormal value threshold. If the abnormal value of any target signal group is greater than or equal to the abnormal value threshold, then determine that the target signal group is an invalid target signal group.

[0036] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows:

[0037] The present invention collects electrical signals in a smart grid to obtain initial electrical signals; constructs an initial signal waveform diagram of the initial electrical signals, where the abscissa of the initial signal waveform diagram is time and the ordinate is frequency, and filters out extreme points containing harmonic signals as target signals in the initial signal waveform diagram. According to the distribution characteristics of the target signals in the initial signal waveform diagram, the target signals in the initial signal waveform diagram are grouped to obtain at least two target signal groups; for any target signal group, according to the frequency characteristics of the target signals in the target signal group, a first anomaly coefficient of the target signal group is obtained, and according to the fluctuation characteristics of the target signals in the target signal group, a second anomaly coefficient of the target signal group is obtained. According to the first anomaly coefficient and the second anomaly coefficient of the target signal group, an anomaly value of the target signal group is obtained; the anomaly values of each target signal group are obtained, and according to the anomaly values of each target signal group, invalid target signal groups are screened out among all target signal groups, and the invalid target signal groups in the initial electrical signals are removed to obtain effective electrical signals. An EMD decomposition result of the effective electrical signals is obtained, and a fault identification of the smart grid is performed according to the EMD decomposition result. The present invention first filters out target signals containing harmonic signals, then groups the target signals, and by calculating the anomaly values of the target signal groups, determines whether the target signal groups contain fault information, and screens out and removes invalid target signal groups (i.e., target signal groups that do not contain fault information), which not only reduces the interference of harmonic signals on the subsequent EMD decomposition algorithm in the fault identification process, but also retains the integrity of the fault information, thereby improving the accuracy of the EMD decomposition algorithm in the fault identification process. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0039] Figure 1 is a flowchart of a method for dispatching fault analysis and identification of a digital smart grid provided in Embodiment 1 of the present invention;

[0040] Figure 2 is an initial signal waveform diagram provided in Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] Embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, but should not be understood as limiting the present disclosure.

[0042] It should be noted that the terms "first", "second", etc. in the specification of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0043] In order to illustrate the technical solution of the present invention, specific embodiments are provided below for illustration.

[0044] See also Figure 1 , is a method flow chart of a method for analyzing and identifying scheduling faults in a digital smart grid provided in the first embodiment of the present invention, such as Figure 1 As shown, the method may include:

[0045] Step S101, collecting electrical signals in the smart grid to obtain initial electrical signals.

[0046] Collect the electrical signals of the digital smart grid within 24 hours at a frequency of once per second to obtain the initial electrical signals. There is no restriction here and it can be set according to the specific implementation scenario.

[0047] Step S102, constructing an initial signal waveform diagram of the initial electrical signal, wherein the horizontal axis of the initial signal waveform diagram is time and the vertical axis is frequency, selecting extreme points containing harmonic signals in the initial signal waveform diagram as target signals, and grouping the target signals in the initial signal waveform diagram according to the distribution characteristics of the target signals in the initial signal waveform diagram to obtain at least two target signal groups.

[0048] Due to the presence of harmonic interference in the initial electrical signal, the accuracy of fault analysis of smart grids using the EMD decomposition algorithm is reduced. Therefore, before fault analysis, it is necessary to remove signals containing unnecessary harmonics (i.e., signals that do not contain fault information) to reduce the interference of harmonics on the subsequent EMD decomposition algorithm in the fault identification process. Therefore, after obtaining the initial electrical signal, it is necessary to filter out the target signal containing harmonic signals.

[0049] According to the obtained initial electrical signal, the initial signal waveform is constructed with the time T of the initial electrical signal as the horizontal coordinate and the frequency Hz of the initial electrical signal as the vertical coordinate, as shown inFigure 2 as shown

[0050] Since normal harmonics usually only contain components that are integer multiples of the fundamental frequency, that is, the nth harmonic (n is an integer). For example, when the fundamental frequency is 50 Hz, normal harmonics may contain harmonic components such as 100 Hz (2nd harmonic), 150 Hz (3rd harmonic), etc. According to the regulation, the fundamental frequency is 50 HZ. The derivative method is used to obtain each extreme point in the initial signal waveform diagram, and the extreme points containing harmonic signals (i.e., the extreme points with frequency values greater than 50 HZ) in the initial signal waveform diagram are screened as the target signals. Among them, the derivative method belongs to the prior art and will not be elaborated here.

[0051] After obtaining the target signals, for the convenience of subsequent analysis and calculation, the target signals in the initial signal waveform diagram are grouped. The specific grouping method is as follows: According to the time of the target signals in the initial signal waveform diagram, all the target signals within a continuous time are grouped into a target signal group, and at least two target signal groups are obtained.

[0052] So far, the target signal groups containing harmonic signals in the initial electrical signal are obtained.

[0053] Step S103, for any target signal group, according to the frequency characteristics of the target signals in the target signal group, obtain the first abnormal coefficient of the target signal group. According to the fluctuation characteristics of the target signals in the target signal group, obtain the second abnormal coefficient of the target signal group. According to the first abnormal coefficient and the second abnormal coefficient of the target signal group, obtain the abnormal value of the target signal group.

[0054] After obtaining the target signal groups containing harmonic signals, regarding the necessity of the target signals, compared with the target signals with high necessity (i.e., the signals containing faults), the target signals with low necessity (i.e., the signals not containing faults) belong to interference signals. Therefore, the target signals with low necessity (i.e., the signals not containing fault information) are abnormal data compared with the target signals with high necessity (i.e., the signals containing fault information). Thus, the target signals with low necessity (i.e., the signals not containing fault information) can be eliminated by calculating the abnormal values of each target signal group and according to the abnormal values.

[0055] Since the frequency of normal harmonics is an integer multiple of the fundamental frequency, and the harmonic signal containing a fault has fault interference, and the frequency of the fault signal has irregular and unruly characteristics, it will cause distortion of the harmonic signal containing the fault, making the frequency of the harmonic signal containing the fault not an integer multiple of the fundamental frequency; moreover, the amplitude and phase of the normal harmonic signal are relatively stable and have a certain regularity. In an ideal situation, the amplitude and phase of each segment of the harmonic signal within a certain range may remain relatively consistent at different time points, while the overall stability and regularity of the harmonic signal containing the fault will be reduced due to fault interference; at the same time, the harmonic signal containing the fault may have obvious fluctuations on the waveform curve, and the waveform curve is unstable. Therefore, for any target signal group, the outlier of the target signal group is calculated according to the frequency characteristics and fluctuation characteristics of the target signals in the target signal group.

[0056] Among them, the method for calculating the outlier of the target signal group according to the frequency characteristics and fluctuation characteristics of the target signals in the target signal group is as follows:

[0057] (1) According to the frequency characteristics of the target signals in the target signal group, obtain the first outlier coefficient of the target signal group.

[0058] Since the frequency of normal harmonics is an integer multiple of the fundamental frequency, but in practice, there will be various interferences, and the situation that is the same as the theory cannot be obtained, making the frequency of the normal harmonic signal not an integer multiple of the fundamental frequency, that is, there will be a remainder when the frequency of the normal harmonic signal is divided by the fundamental frequency, but the remainder will be small, while the remainder when the frequency of the harmonic signal containing the fault is divided by the fundamental frequency will be large. Therefore, the first outlier coefficient of the target signal group can be obtained by calculating the remainder when the frequency of the target signals in the target signal group is divided by the fundamental frequency.

[0059] Specifically, obtain the remainder when the frequency of each target signal in the target signal group is divided by the fundamental frequency respectively, correspondingly obtain the mean value of the remainders, obtain the first addition result of the mean value of the remainders and a preset value, and perform normalization processing on the reciprocal of the first addition result to obtain the first outlier coefficient of the target signal group.

[0060] In an embodiment, obtain the remainder when the frequency of each target signal in the target signal group is divided by the fundamental frequency, and calculate the first outlier coefficient of the target signal group:

[0061]

[0062] Among them, α is the first outlier coefficient of the target signal group; N is the number of target signals in the target signal group; c iis the remainder obtained by dividing the frequency of the \(i\)-th target signal in the target signal group by the fundamental frequency; \(i\) is the serial number of the target signal in the target signal group; \(b\) is a preset value; \(norm()\) is a normalization function.

[0063] It should be noted that the preset value \(b\) is a non-zero constant, which is used to prevent calculation errors in the formula when the denominator takes a value of 0. It is set to 0.1 according to historical experience. There is no limit here and it can be set according to specific implementation scenarios; the smaller the remainder obtained by dividing the frequency of the target signal in the target signal group by the fundamental frequency, the larger the first anomaly coefficient of the target signal group. Since target signals with low necessity (i.e., signals that do not contain fault information) are abnormal data compared to target signals with high necessity, therefore, the larger the first anomaly coefficient of the target signal group, the smaller the possibility that the target signal group contains fault information, and the lower the necessity for the existence of the target signal.

[0064] (2) Obtain the second anomaly coefficient of the target signal group according to the fluctuation characteristics of the target signals in the target signal group.

[0065] Since the amplitude and phase of harmonic signals are relatively stable and have a certain regularity, under ideal conditions, the amplitude and phase of the same harmonic signal may remain relatively consistent at different time points, while the amplitude and phase of harmonic signals with faults will show large fluctuations and uncertainties. At the same time, the harmonic waveform curve without faults will be relatively smooth with few local fluctuations, while the harmonic waveform curve with faults will be more fluctuating. Therefore, the second anomaly coefficient of the target signal group can be obtained according to the fluctuation characteristics of the target signals in the target signal group.

[0066] The specific method for obtaining the second anomaly coefficient of the target signal group is as follows:

[0067] a Obtain the peak signals and valley signals in the target signal group, sort them separately in chronological order, and obtain the peak numbers of each peak signal and the valley numbers of each valley signal in the target signal group.

[0068] According to the frequencies of the target signals in the target signal group, the target signals are divided into positive-class target signals and negative-class target signals. All the positive-class target signals within a continuous time period are grouped into a positive-class target signal subgroup, and the one with the maximum frequency value in the positive-class target signal subgroup is obtained as the peak signal (a single positive-class target signal is also a positive-class target signal subgroup, and this positive-class target signal is the peak signal); all the negative-class target signals within a continuous time period are grouped into a negative-class target signal subgroup, and the one with the maximum frequency value in the negative-class target signal subgroup is obtained as the trough signal (a single negative-class target signal is also a negative-class target signal subgroup, and this negative-class target signal is the trough signal); all the peak signals and trough signals in the target signal group are obtained, and the peak signals and trough signals are sorted separately in chronological order to obtain the peak numbers of each peak signal and the trough numbers of each trough signal in the target signal group.

[0069] b Obtain the phases of each peak signal and each trough signal in the target signal group, and obtain the lateral overall stability index of the target signal group according to the difference between the phase of the peak signal in the target signal group and the theoretical peak phase, and the difference between the phase of the trough signal in the target signal group and the theoretical trough phase.

[0070] Specifically, for any peak signal in the target signal group, obtain the difference between the phase of the peak signal and the theoretical peak phase to get a first difference, calculate the ratio of the first difference to the circumferential angle to get a first ratio, obtain the difference between the peak number of the peak signal and the constant 1 to get a second difference, calculate the difference between the first ratio and the second difference to get the phase difference of the peak signal, and calculate the phase differences of each peak signal in the target signal group respectively to correspondingly obtain the average value of the phase differences of the peak signals;

[0071] For any trough signal in the target signal group, obtain the difference between the phase of the trough signal and the theoretical trough phase to get a third difference, calculate the ratio of the third difference to the circumferential angle to get a second ratio, obtain the difference between the trough number of the trough signal and the constant 1 to get a fourth difference, calculate the difference between the second ratio and the fourth difference to get the phase difference of the trough signal, and calculate the phase differences of each trough signal in the target signal group respectively to correspondingly obtain the average value of the phase differences of the trough signals;

[0072] Obtain the sum of the average value of the phase differences of the peak signals, the average value of the phase differences of the trough signals and a preset value to get a second addition result, and obtain the reciprocal of the second addition result as the lateral overall stability index of the target signal group.

[0073] In one embodiment, the actual phase of each peak signal and each valley signal is obtained by using the inverse tangent method of phase. The inverse tangent method of phase belongs to the prior art and will not be elaborated here. Since the fundamental wave is a sine wave and the harmonic wave is also a sine wave, when the phase is (90° + 2ωπ) (ω = 0, 1, 2, 3, 4,...), it is a peak, and when the phase is (270° + 2ωπ), it is a valley. Therefore, the theoretical peak phase of the peak signal is recorded as 90°, and the theoretical valley phase of the valley signal is recorded as 270°. Calculate the horizontal overall stability index of the target signal group:

[0074]

[0075] Wherein, is the horizontal overall stability index of the target signal group; U j is the actual phase of the j-th peak signal in the target signal group; j is the serial number of the peak signal and the valley signal in the target signal group; n1 is the number of peak signals in the target signal group; 2π is the angle of a circle; D j is the actual phase of the j-th valley signal in the target signal group; n2 is the number of valley signals in the target signal group; b is a preset value.

[0076] It should be noted that the preset value b is a non-zero constant, which is used to prevent the formula calculation from going wrong when the denominator takes a value of 0. It is set to 0.1 according to historical experience and is not limited here. It can be set according to the specific implementation scenario; the smaller the difference between the actual phase of the peak signal in the target signal group and the theoretical peak phase, and the difference between the actual phase of the valley signal in the target signal group and the theoretical valley phase, the larger the horizontal overall stability index of the target signal group. Since the target signal with low necessity (i.e., the signal that does not contain fault information) belongs to abnormal data compared with the target signal with high necessity, therefore, the larger the horizontal overall stability index of the target signal group, the smaller the possibility that the target signal group contains fault information, and the lower the necessity of the existence of the target signal.

[0077] c Obtain the vertical overall stability index of the target signal group according to the amplitude difference between the peak signal and the valley signal in the target signal group.

[0078] Specifically, respectively obtain the frequency difference between each peak signal and the valley signal with the same serial number in the target signal group, correspondingly obtain a frequency difference sequence, respectively obtain the absolute value of the difference between every two adjacent frequency differences in the frequency difference sequence, correspondingly obtain the mean value of the absolute value of the difference, obtain the sum of the mean value of the absolute value of the difference and the preset value, obtain the reciprocal of the sum result as the vertical overall stability index of the target signal group.

[0079] In one embodiment, calculate the vertical overall stability index of the target signal group:

[0080]

[0081] Among them, B is the longitudinal overall stability index of the target signal group; WU j is the frequency of the j-th peak signal in the target signal group; WD j is the frequency of the j-th trough signal in the target signal group; WU j-1 is the frequency of the (j - 1)-th peak signal in the target signal group; WD j-1 is the frequency of the (j - 1)-th trough signal in the target signal group; j is the serial number of the peak signal and the trough signal in the target signal group; n1 is the number of peak signals in the target signal group; n2 is the number of trough signals in the target signal group; b is a preset value; || is the absolute value symbol; min is the minimum value function.

[0082] It should be noted that the preset value b is a non-zero constant, which is used to prevent calculation errors in the formula caused by the denominator taking the value of 0. It is set to 0.1 according to historical experience, and there is no restriction here. It can be set according to specific implementation scenarios; |(WU j - WD j ) - (WU j-1 - WD j-1 )| is the absolute value of the difference between the difference of the j-th peak signal and the trough signal and the difference of the (j - 1)-th peak signal and the trough signal in the target signal group. The smaller |(WU j - WD j ) - (WU j-1 - WD j-1 )| is, the greater the longitudinal overall stability index of the target signal group. Since the target signals with low necessity (i.e., signals that do not contain fault information) are abnormal data compared to the target signals with high necessity, therefore, the greater the longitudinal overall stability index of the target signal group, the smaller the possibility that the target signal group contains fault information, and the lower the necessity for the existence of the target signal.

[0083] Construct the target curve of the target signal group, divide the target curve into at least two sub-curves, obtain the curvature of each sub-curve, and obtain the smoothness index of the target curve according to the curvature of each sub-curve in the target curve.

[0084] Specifically, according to the curvature of each sub-curve in the target curve, obtain the average curvature of the target curve, and obtain the reciprocal of the average curvature as the smoothness index of the target curve.

[0085] In one embodiment, according to the frequencies of the target signals in the target signal group, a target curve of the target signal group is constructed. The target signals in the target signal group are grouped by three target signals (if the number of target signals in the target signal group is not a multiple of 3, the last group with less than three target signals is supplemented with target signals at adjacent times). The target curve is divided into multiple sub-curves, which is not limited here and can be set according to specific implementation scenarios. The curvature of each sub-curve is obtained. Curvature belongs to the prior art and will not be elaborated here. Calculate the smoothness index of the target curve:

[0086]

[0087] where δ is the smoothness index of the target curve; ρ k is the curvature of the k-th sub-curve into which the target curve is divided; k is the serial number of the sub-curves into which the target curve is divided; N is the number of target signals in the target signal group.

[0088] It should be noted that the smaller the curvature of each sub-curve into which the target curve is divided, the smaller the degree of bending of each sub-curve, and the larger the smoothness index of the target curve. Since the target signals with low necessity (i.e., signals that do not contain fault information) are abnormal data compared to the target signals with high necessity, therefore, the larger the smoothness index of the target curve, the smaller the possibility that the target signal group contains fault information, and the lower the necessity for the existence of the target signal.

[0089] e Obtain the second anomaly coefficient of the target signal group according to the smoothness index, the lateral overall stability index, and the longitudinal overall stability index.

[0090] Specifically, obtain the mean value of the lateral overall stability index and the longitudinal overall stability index to get the first mean value, and obtain the mean value of the first mean value and the smoothness index as the second anomaly coefficient of the target signal group.

[0091] In one embodiment, calculate the second anomaly coefficient of the target signal group:

[0092]

[0093] where β is the second anomaly coefficient of the target signal group; is the lateral overall stability index of the target signal group; B is the longitudinal overall stability index of the target signal group; δ is the smoothness index of the target curve.

[0094] It should be noted that the greater the horizontal overall stability index of the target signal group, the greater the second anomaly coefficient of the target signal group; the greater the vertical overall stability index of the target signal group, the greater the second anomaly coefficient of the target signal group; the greater the smoothness index of the target curve, the greater the second anomaly coefficient of the target signal group; since the target signals with low necessity (i.e., signals that do not contain fault information) are abnormal data compared to the target signals with high necessity, therefore, the greater the second anomaly coefficient of the target signal group, the smaller the possibility that the target signal group contains fault information, and the lower the necessity for the existence of the target signal.

[0095] (3) Obtain the anomaly value of the target signal group according to the first anomaly coefficient and the second anomaly coefficient of the target signal group.

[0096] Specifically, obtain the mean value of the first anomaly coefficient and the second anomaly coefficient of the target signal group as the anomaly value of the target signal group.

[0097] In one embodiment, calculate the anomaly value of the target signal group:

[0098]

[0099] Where γ is the anomaly value of the target signal group; α is the first anomaly coefficient of the target signal group; β is the second anomaly coefficient of the target signal group.

[0100] It should be noted that the greater the first anomaly coefficient of the target signal group, the smaller the possibility that the target signal group contains fault information, and the greater the anomaly value of the target signal group; the greater the second anomaly coefficient of the target signal group, the smaller the possibility that the target signal group contains fault information, and the greater the anomaly value of the target signal group; since the target signals with low necessity (i.e., signals that do not contain fault information) are abnormal data compared to the target signals with high necessity, therefore, the greater the anomaly value of the target signal group, the smaller the possibility that the target signal group contains fault information, and the lower the necessity for the existence of the target signal.

[0101] Thus, the anomaly value of the target signal group is obtained.

[0102] Step S104, obtain the anomaly value of each target signal group, screen out the invalid target signal groups from all the target signal groups according to the anomaly value of each target signal group, remove the invalid target signal groups from the initial electrical signal to obtain the effective electrical signal, obtain the EMD decomposition result of the effective electrical signal, and perform fault identification on the smart grid according to the EMD decomposition result.

[0103] After obtaining the outliers of the target signal groups, according to the above method for obtaining the outliers of the target signal groups, obtain the outliers of each target signal group in the initial electrical signal. Since the smaller the outlier of the target signal group, the greater the possibility that the target signal group contains fault information. Therefore, set the outlier threshold to 0.6. The outlier threshold is obtained through calculation experiments on historical signals and is not limited here. It can be set according to specific implementation scenarios. If the outlier of any target signal group is less than 0.6, it is determined that the target signal group contains fault information. If the outlier of any target signal group is greater than or equal to 0.6, it is determined that the target signal group does not contain fault information, that is, the target signal group is an invalid target signal group. Remove the invalid target signal groups from the initial electrical signal to obtain the effective electrical signal.

[0104] Since the invalid target signal groups (i.e., the target signal groups that do not contain fault information) are removed, unnecessary harmonic interference is reduced. At this time, the obtained effective electrical signal retains the integrity of the fault information. Therefore, analyzing and processing the effective electrical signal improves the accuracy of the subsequent EMD decomposition algorithm in the process of fault identification.

[0105] After obtaining the effective electrical signal, use the EMD decomposition algorithm to decompose the effective electrical signal to obtain multiple intrinsic mode functions (IMFs) and a residue. Since the EMD decomposition algorithm belongs to the prior art, the following is a brief description of the steps: (1) Identify all local maximum and minimum points of the effective electrical signal; (2) Construct the upper and lower envelope lines of the signal by interpolating the extreme points, calculate the average value of the upper and lower envelope lines to obtain the first mean curve of the effective electrical signal; (3) Subtract the first mean curve from the effective electrical signal to obtain the first IMF component; (4) If the first IMF component does not meet the conditions of the IMF (that is, within the entire data range, the number of extreme points and zero-crossing points is equal or differs by at most 1, and at any moment, the average value of the upper and lower envelope lines is zero), then use it as a new signal and repeat the above steps until an IMF component that meets the conditions is extracted; After each IMF component that meets the conditions is extracted, subtract it from the effective electrical signal, and the remaining part is used as a new signal to continue the decomposition; When the residue becomes a monotonic sum function or meets the preset termination conditions, the decomposition process ends.

[0106] After the decomposition of the effective electrical signal is completed, multiple IMF components are obtained. Each IMF component represents an inherent vibration mode in the signal, and they usually contain different frequency components of the signal. Analyze the statistical characteristics of each IMF component, such as amplitude, frequency, and energy, etc. Extract the features related to faults from the high-frequency IMF components, and use the extracted features to effectively analyze and identify faults in the smart grid.

[0107] In summary, in this embodiment, electrical signals in the smart grid are collected to obtain initial electrical signals; an initial signal waveform diagram of the initial electrical signals is constructed, where the abscissa of the initial signal waveform diagram is time and the ordinate is frequency. Extreme points containing harmonic signals are screened out as target signals in the initial signal waveform diagram. According to the distribution characteristics of the target signals in the initial signal waveform diagram, the target signals in the initial signal waveform diagram are grouped to obtain at least two target signal groups; for any target signal group, according to the frequency characteristics of the target signals in the target signal group, a first anomaly coefficient of the target signal group is obtained, according to the fluctuation characteristics of the target signals in the target signal group, a second anomaly coefficient of the target signal group is obtained, and according to the first anomaly coefficient and the second anomaly coefficient of the target signal group, an anomaly value of the target signal group is obtained; the anomaly values of each target signal group are obtained, and according to the anomaly values of each target signal group, invalid target signal groups are screened out from all target signal groups, the invalid target signal groups in the initial electrical signals are removed to obtain effective electrical signals, the EMD decomposition result of the effective electrical signals is obtained, and the smart grid is fault-identified according to the EMD decomposition result. In this embodiment, target signals containing harmonic signals are first screened out, then the target signals are grouped, and by calculating the anomaly values of the target signal groups, it is judged whether the target signal groups contain fault information, and the invalid target signal groups (i.e., the target signal groups that do not contain fault information) are screened out and removed, which not only reduces the interference of harmonic signals on the subsequent EMD decomposition algorithm in the fault identification process, but also retains the integrity of the fault information, thereby improving the accuracy of the EMD decomposition algorithm in the fault identification process.

[0108] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for analyzing and identifying dispatching faults in a digital smart grid, characterized in that: The method comprises: Collecting electrical signals in the smart grid to obtain initial electrical signals; Constructing an initial signal waveform diagram of the initial electrical signal, wherein the abscissa of the initial signal waveform diagram is time and the ordinate is frequency, selecting extreme value points containing harmonic signals in the initial signal waveform diagram as target signals, and grouping the target signals in the initial signal waveform diagram according to the distribution characteristics of the target signals in the initial signal waveform diagram to obtain at least two target signal groups; For any target signal group, according to the frequency characteristics of the target signals in the target signal group, a first abnormal coefficient of the target signal group is obtained, according to the fluctuation characteristics of the target signals in the target signal group, a second abnormal coefficient of the target signal group is obtained, and according to the first abnormal coefficient and the second abnormal coefficient of the target signal group, an abnormal value of the target signal group is obtained; Obtain an abnormal value of each target signal group, screen all target signal groups to obtain an invalid target signal group based on the abnormal value of each target signal group, remove the invalid target signal group in the initial electrical signal to obtain a valid electrical signal, obtain an EMD decomposition result of the valid electrical signal, and perform fault identification on the smart grid based on the EMD decomposition result.

2. A method for analyzing and identifying dispatching faults for a digital smart grid according to claim 1, characterized in that: The target signals in the initial signal waveform diagram are grouped according to the distribution characteristics of the target signals in the initial signal waveform diagram to obtain at least two target signal groups, including: According to the time of the target signal in the initial signal waveform diagram, all target signals in a continuous time are grouped into a target signal group to obtain at least two target signal groups.

3. A method for analyzing and identifying dispatching faults for a digital smart grid according to claim 1, characterized in that: The step of obtaining a first abnormal coefficient of the target signal group according to the frequency characteristics of the target signal in the target signal group includes: Respectively obtain the remainders of the frequency of each target signal in the target signal group divided by the fundamental frequency, obtain the corresponding remainder mean, obtain the first addition result of the remainder mean and a preset value, normalize the inverse of the first addition result, and obtain the first abnormal coefficient of the target signal group.

4. A method for analyzing and identifying dispatching faults for a digital smart grid according to claim 1, characterized in that: The step of obtaining a second abnormal coefficient of the target signal group according to the fluctuation characteristics of the target signal in the target signal group includes: Acquire the peak signals and trough signals in the target signal group, and sort them in chronological order to obtain the peak number of each peak signal and the trough number of each trough signal in the target signal group; Obtaining the phase of each crest signal and each trough signal in the target signal group, and obtaining the lateral overall stability index of the target signal group according to the difference between the phase of the crest signal in the target signal group and the theoretical crest phase, and the difference between the phase of the trough signal in the target signal group and the theoretical trough phase; Obtaining a longitudinal overall stability index of the target signal group according to an amplitude difference between a peak signal and a trough signal in the target signal group; Constructing a target curve of the target signal group, dividing the target curve into at least two sub-curves, obtaining the curvature of each sub-curve, and obtaining a smoothness index of the target curve according to the curvature of each sub-curve in the target curve; A second abnormality coefficient of the target signal group is obtained according to the smoothness index, the lateral overall stability index and the longitudinal overall stability index.

5. A method for analyzing and identifying dispatching faults for a digital smart grid according to claim 4, characterized in that: The step of obtaining the lateral overall stability index of the target signal group according to the difference between the phase of the crest signal in the target signal group and the theoretical crest phase, and the difference between the phase of the trough signal in the target signal group and the theoretical trough phase, comprises: For any peak signal in the target signal group, obtain the difference between the phase of the peak signal and the theoretical peak phase to obtain a first difference, calculate the ratio of the first difference to the circular angle to obtain a first ratio, obtain the difference between the peak number of the peak signal and a constant 1 to obtain a second difference, calculate the difference between the first ratio and the second difference to obtain the phase difference of the peak signal, respectively calculate the phase difference of each peak signal in the target signal group, and obtain the corresponding average phase difference of the peak signals; For any trough signal in the target signal group, obtain the difference between the phase of the trough signal and the theoretical trough phase to obtain a third difference, calculate the ratio of the third difference to the circular angle to obtain a second ratio, obtain the difference between the trough number of the trough signal and a constant 1 to obtain a fourth difference, calculate the difference between the second ratio and the fourth difference to obtain the phase difference of the trough signal, respectively calculate the phase difference of each trough signal in the target signal group, and obtain the corresponding average phase difference of the trough signals; Obtain the sum of the average phase difference value of the peak signal, the average phase difference value of the trough signal and a preset value to obtain a second addition result, and obtain the inverse of the second addition result as the lateral overall stability index of the target signal group.

6. A method for analyzing and identifying dispatching faults for a digital smart grid according to claim 4, characterized in that: The step of obtaining the longitudinal overall stability index of the target signal group according to the amplitude difference between the peak signal and the trough signal in the target signal group includes: Obtain the frequency difference between each peak signal and the trough signal with the same sequence number in the target signal group respectively, and obtain a corresponding frequency difference sequence; obtain the absolute value of the difference between every two adjacent frequency differences in the frequency difference sequence respectively, and obtain the mean of the absolute values ​​of the differences respectively; obtain the sum of the mean of the absolute values ​​of the differences and a preset value to obtain a third addition result; obtain the inverse of the third addition result as the longitudinal overall stability indicator of the target signal group.

7. A method for analyzing and identifying dispatching faults for a digital smart grid according to claim 4, characterized in that: The step of obtaining the smoothness index of the target curve according to the curvature of each sub-curve in the target curve includes: According to the curvature of each sub-curve in the target curve, the mean curvature of the target curve is obtained, and the inverse of the mean curvature is obtained as a smoothness index of the target curve.

8. A method for analyzing and identifying dispatching faults for a digital smart grid according to claim 4, characterized in that: The step of obtaining a second abnormality coefficient of the target signal group according to the smoothness index, the lateral overall stability index, and the longitudinal overall stability index comprises: The mean of the lateral overall stability index and the longitudinal overall stability index is obtained to obtain a first mean, and the mean of the first mean and the smoothness index is obtained as a second abnormal coefficient of the target signal group.

9. A method for analyzing and identifying dispatching faults for a digital smart grid according to claim 1, characterized in that: The obtaining, according to the first abnormal coefficient and the second abnormal coefficient of the target signal group, the abnormal value of the target signal group comprises: The average of the first abnormal coefficient and the second abnormal coefficient of the target signal group is obtained as the abnormal value of the target signal group.

10. A method for analyzing and identifying dispatching faults for a digital smart grid according to claim 1, characterized in that: The step of screening all target signal groups to obtain invalid target signal groups according to the abnormal value of each target signal group includes: An abnormal value threshold is set, and if the abnormal value of any target signal group is greater than or equal to the abnormal value threshold, the target signal group is determined to be an invalid target signal group.

Images