Method, device, computer device, and storage medium for identifying voltage sags

Abstract

The present application relates to a method, device, computer equipment, storage medium and computer program product for identifying voltage sags. The method includes: if a voltage sag signal sent by a power system is received, performing short-time Fourier processing on a plurality of three-phase voltage data during the sag time period to obtain a plurality of processed target voltage data; wherein, each sag moment during the sag time period corresponds to a target voltage data. Feature extraction is performed on the target voltage data respectively corresponding to each sag moment to obtain a basic feature for characterizing the voltage sag and a curvature feature for characterizing the curvature change of the voltage sag; wherein, the basic feature includes at least one of three-phase unbalance degree, sag depth and sag duration. Through a sag classification model, voltage sag identification processing is performed on the basic feature and the curvature feature to obtain the identification type of the voltage sag. In this way, the accuracy of voltage sag identification is greatly improved.

Description

Technical Field

[0001] This application relates to the technical field of voltage sag identification, and particularly to a method, device, computer device, storage medium, and computer program product for identifying voltage sags. Background Art

[0002] With the development of voltage sag identification technology, the identification method of manually extracting features is often used to identify the categories of voltage sags. Among them, the identification method of manually extracting features is to obtain features such as the root mean square value and related extension amounts of voltage sags, and then process the obtained features by combining the identification method of manual experience.

[0003] However, if the identification method of manually extracting features is adopted, the extracted feature quantities are often affected by subjective factors and are difficult to accurately reflect the process of voltage sags, resulting in inaccurate voltage sag classification and identification results, and there is a problem of low accuracy in identifying voltage sags. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide a method, device, computer device, computer-readable storage medium, and computer program product for identifying voltage sags.

[0005] In a first aspect, this application provides a method for identifying voltage sags. The method includes:

[0006] If a voltage sag signal sent by the power system is received, perform short-time Fourier processing on multiple three-phase voltage data within the sag time period to obtain multiple processed target voltage data; wherein, each sag moment within the sag time period corresponds to a target voltage data;

[0007] Extract features from the target voltage data corresponding to each sag moment respectively to obtain a basic feature for characterizing the voltage sag and a curvature feature for characterizing the curvature change of the voltage sag; wherein, the basic feature includes at least one of three-phase unbalance degree, sag depth, and sag duration;

[0008] Through a sag classification model, perform voltage sag identification processing on the basic feature and the curvature feature to obtain the identification type of the voltage sag.

[0009] In a second aspect, this application also provides a device for identifying voltage sags. The device includes:

[0010] A processing module, configured to, if a voltage sag signal sent by the power system is received, perform short-time Fourier processing on multiple three-phase voltage data within the sag time period to obtain multiple processed target voltage data; wherein, each sag moment within the sag time period corresponds to a target voltage data;

[0011] An extraction module, configured to perform feature extraction on target voltage data respectively corresponding to each sag moment, to obtain a basic feature for characterizing voltage sag and a curvature feature for characterizing the curvature change of voltage sag; wherein, the basic feature includes at least one of three-phase unbalance degree, sag depth, and sag duration;

[0012] An identification module, configured to perform voltage sag identification processing on the basic feature and the curvature feature through a sag classification model, to obtain an identification type of voltage sag.

[0013] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0014] If a voltage sag signal sent by a power system is received, perform short-time Fourier processing on multiple three-phase voltage data within the sag time period, to obtain multiple processed target voltage data; wherein, each sag moment within the sag time period corresponds to a target voltage data;

[0015] Perform feature extraction on target voltage data respectively corresponding to each sag moment, to obtain a basic feature for characterizing voltage sag and a curvature feature for characterizing the curvature change of voltage sag; wherein, the basic feature includes at least one of three-phase unbalance degree, sag depth, and sag duration;

[0016] Perform voltage sag identification processing on the basic feature and the curvature feature through a sag classification model, to obtain an identification type of voltage sag. [[ID=...]]

[0017] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, and when the computer program is executed by a processor, the following steps are implemented:

[0018] If a voltage sag signal sent by a power system is received, perform short-time Fourier processing on multiple three-phase voltage data within the sag time period, to obtain multiple processed target voltage data; wherein, each sag moment within the sag time period corresponds to a target voltage data;

[0019] Perform feature extraction on target voltage data respectively corresponding to each sag moment, to obtain a basic feature for characterizing voltage sag and a curvature feature for characterizing the curvature change of voltage sag; wherein, the basic feature includes at least one of three-phase unbalance degree, sag depth, and sag duration;

[0020] Through the sag classification model, voltage sag identification processing is performed on the basic features and the curvature features to obtain the identification type of the voltage sag.

[0021] In a fifth aspect, the present application also provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0022] If a voltage sag signal sent by the power system is received, short-time Fourier processing is performed on multiple three-phase voltage data within the sag time period to obtain multiple processed target voltage data; wherein, each sag moment within the sag time period corresponds to a target voltage data;

[0023] Feature extraction is performed on the target voltage data respectively corresponding to each sag moment to obtain basic features for characterizing the voltage sag and curvature features for characterizing the curvature change of the voltage sag; wherein, the basic features include at least one of three-phase unbalance degree, sag depth, and sag duration;

[0024] Through the sag classification model, voltage sag identification processing is performed on the basic features and the curvature features to obtain the identification type of the voltage sag.

[0025] For the above voltage sag identification method, device, computer device, storage medium, and computer program product, if a voltage sag signal sent by the power system is received, short-time Fourier processing is performed on multiple three-phase voltage data within the sag time period to obtain multiple processed target voltage data; wherein, each sag moment within the sag time period corresponds to a target voltage data. Feature extraction is performed on the target voltage data respectively corresponding to each sag moment to obtain basic features for characterizing the voltage sag and curvature features for characterizing the curvature change of the voltage sag; wherein, the basic features include at least one of three-phase unbalance degree, sag depth, and sag duration. In this way, on the basis of the basic features, combined with the curvature features for characterizing the curvature change of the voltage sag, the detailed information of the voltage sag can be further expanded, thereby greatly improving the effectiveness of feature extraction. Through the sag classification model, voltage sag identification processing is performed on the basic features and the curvature features to obtain the identification type of the voltage sag. In this way, through the sag classification model, processing the basic features and the curvature features can comprehensively and accurately reflect the change process of the voltage sag, thereby accurately determining the identification type of the voltage sag and greatly improving the accuracy of voltage sag identification. Description of the Drawings

[0026] Figure 1 It is an application environment diagram of the voltage sag identification method in an embodiment;

[0027] Figure 2Schematic flowchart of the voltage sag identification method in an embodiment;

[0028] Figure 3 Schematic diagram of determining the updated short-time Fourier step in an embodiment;

[0029] Figure 4 Curve graph of the effective voltage value changing with time in an embodiment;

[0030] Figure 5 Schematic flowchart of the steps for determining the basic features and curvature features in an embodiment;

[0031] Figure 6 Curve graph of the effective voltage value changing with time in another embodiment;

[0032] Figure 7 Schematic flowchart of the steps for determining the basic features in an embodiment;

[0033] Figure 8 Curve graph of the effective voltage value changing with time in another embodiment;

[0034] Figure 9 Schematic flowchart of the steps for determining the sag classification model in an embodiment;

[0035] Figure 10 Schematic diagram of the optimization part of the sag classification model in an embodiment;

[0036] Figure 11 Schematic flowchart of the steps for determining the sag classification model in another embodiment;

[0037] Figure 12 Schematic diagram of the steps for determining the sag classification model in another embodiment;

[0038] Figure 13 Schematic block diagram of the voltage sag identification device in an embodiment;

[0039] Figure 14 Internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0040] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0041] The voltage sag identification method provided by the embodiments of the present application can be applied to, for example Figure 1In the application environment shown. Among them, the power system 102 communicates with the computer device 104 through a network. Among them, the computer device 104 can be a terminal or a server. Among them, the data storage system can store the data that the server needs to process. The data storage system can be integrated on the server, or can be placed in the cloud or on other network servers. If the computer device 104 receives the voltage sag signal sent by the power system 102, it performs short-time Fourier processing on multiple three-phase voltage data during the sag time period to obtain multiple processed target voltage data; among them, each sag moment during the sag time period corresponds to a target voltage data. The computer device 104 extracts features from the target voltage data corresponding to each sag moment respectively to obtain a basic feature for characterizing the voltage sag and a curvature feature for characterizing the curvature change of the voltage sag; among them, the basic feature includes at least one of three-phase unbalance degree, sag depth, and sag duration. The computer device 104 performs voltage sag identification processing on the basic feature and the curvature feature through a sag classification model to obtain the identification type of the voltage sag. Among them, the power system 102 can be, but is not limited to, various computers, laptops, tablets, etc. Among them, the terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, Internet of Things devices, and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented by an independent server or a server cluster composed of multiple servers.

[0042] In one embodiment, as Figure 2 shown, a method for identifying voltage sag is provided. Taking the computer device in Figure 1 as an example, the method includes the following steps:

[0043] Step 202, if a voltage sag signal sent by the power system is received, perform short-time Fourier processing on multiple three-phase voltage data during the sag time period to obtain multiple processed target voltage data; among them, each sag moment during the sag time period corresponds to a target voltage data.

[0044] Among them, the power system converts primary energy into electrical energy through a power generation power device, and then supplies the electrical energy to each user through power transmission, transformation, and distribution. Among them, the voltage sag signal is used to characterize the occurrence of a voltage sag phenomenon. The voltage sag (also known as voltage dip) is a phenomenon in which the root mean square value of the power frequency voltage at a certain point in the power system suddenly drops to 0.1 p.u. (per unit value) to 0.9 p.u., and returns to normal after a short duration of 10 ms to 1 min. Among them, the three-phase voltage data is the voltage between phases. Short-time Fourier is used to determine the frequency and phase of the sine wave in the local area of a time-varying signal.

[0045] Specifically, if the computer device receives a voltage sag signal sent by the power system, the computer device acquires multiple three-phase voltage data during the sag time period. The computer device determines a short-time Fourier expression based on the Fourier expression and performs an update process on the short-time Fourier expression to obtain an updated short-time Fourier expression. The computer device processes the multiple three-phase voltage data through the updated short-time Fourier expression to obtain target voltage data corresponding to each sag moment respectively. The target voltage data is characterized in the form of three-phase voltage.

[0046] For example, the computer device acquires a basic Fourier expression as follows:

[0047]

[0048] where X(k) is the data after Fourier transform, and X(n) is the voltage sag signal collected by the monitoring device. Expanding the above formula gives:

[0049]

[0050]

[0051] where the real part of the signal is and the imaginary part is Then the effective value X of the signal is obtained as follows:

[0052]

[0053] Based on the above basic Fourier expression, expanding at X k (0) and X k (1) gives the following two formulas:

[0054]

[0055]

[0056] where ω N = e (-2*j) / n , subtracting the above X k (0) and X k (1) gives the updated short-time Fourier expression as follows:

[0057]

[0058] where the updated short-time Fourier expression uses 1 complex multiplication in a recursive manner to achieve continuous updated short-time Fourier transform. The schematic diagram of this algorithm is as Figure 3As shown in the figure. The computer device obtains multiple three-phase voltage data during the sag time period. This three-phase voltage data has strong time series characteristics. Each cycle consists of 128 discrete signal points and exists in an Excel table in csv file format. The computer device processes the multiple three-phase voltage data through the updated short-time Fourier expression to obtain target voltage data corresponding to each sag moment respectively. This target voltage data is characterized in the form of three-phase voltage.

[0059] It should be noted that considering that the measured waveform is not a perfect sine wave, it is necessary to introduce Fourier transform to filter out the fundamental wave and each harmonic. The short-time Fourier transform is to perform windowing processing on the Fourier transform, that is, divide the time-domain signal into several segmented time regions. The updated short-time Fourier extracts the fundamental wave component and obtains the fundamental wave component through numerical transformation on this basis, which can truthfully reflect the frequency content of the signal and also truly reflect the law of the change of the frequency content over time.

[0060] Step 204: Extract features from the target voltage data corresponding to each sag moment respectively to obtain a basic feature for characterizing voltage sag and a curvature feature for characterizing the curvature change of voltage sag; among them, this basic feature includes at least one of three-phase unbalance degree, sag depth, and sag duration.

[0061] Among them, the three-phase unbalance degree is the degree of three-phase unbalance in a three-phase power system, expressed as the percentage of the root mean square value of the negative sequence component of voltage or current to the positive sequence component. The sag depth is the ratio of the absolute value of the difference between the root mean square value of the residual power frequency voltage and the system nominal voltage to the system nominal voltage in a voltage sag event, generally expressed as a percentage. The sag duration is the duration of the sudden voltage drop recorded with a set voltage sag magnitude as the threshold value.

[0062] Specifically, the computer device obtains the target voltage data corresponding to each sag moment respectively, and performs multiple basic index calculations on each target voltage data to obtain basic features corresponding to each basic index calculation respectively, where the basic index calculation includes three-phase unbalance degree calculation, sag depth calculation, and sag duration calculation. The computer device performs phase space vector calculation and curvature calculation respectively based on each target voltage data to obtain the phase space vector calculation result and the curvature calculation result. The computer device determines the curvature feature based on this phase space vector calculation result and the curvature machine calculation result.

[0063] Step 206: Through the sag classification model, perform voltage sag identification processing on this basic feature and this curvature feature to obtain the identification type of voltage sag.

[0064] Among them, the sag classification model is a model constructed based on the LightGBM (Light Gradient Boosting Machine) algorithm.

[0065] Specifically, the computer device obtains the basic feature and the curvature feature, and inputs the basic feature and the curvature feature into the sag classification model to obtain the recognition type of the voltage sag.

[0066] Among them, the recognition types of voltage sags include induction motor starting type, transformer switching type, and short - circuit fault type. Among them, the induction motor starting type, transformer switching type, and short - circuit fault type are all relatively rapid in the occurrence stage of voltage sags, and usually the effective value drops to the lowest stable value within 1 cycle. Among them, as Figure 4 shown, this figure shows the situation of the voltage effective value changing with time. Among them, the short - circuit fault type recovers from the voltage sag state to the effective value quickly, that is, the sag time is short; while the induction motor starting type and transformer switching type recover relatively slowly. In addition, the three - phase amplitudes of the voltage sags caused by the transformer switching type are always unequal.

[0067] In the above - mentioned voltage sag recognition method, if a voltage sag signal sent by the power system is received, short - time Fourier processing is performed on multiple three - phase voltage data within the sag time period to obtain multiple processed target voltage data; among them, each sag moment within the sag time period corresponds to a target voltage data. Feature extraction is performed on the target voltage data corresponding to each sag moment respectively to obtain a basic feature for characterizing the voltage sag and a curvature feature for characterizing the curvature change of the voltage sag; among them, the basic feature includes at least one of three - phase unbalance degree, sag depth, and sag duration. In this way, on the basis of the basic feature, combined with the curvature feature for characterizing the curvature change of the voltage sag, the detailed information of the voltage sag can be further expanded, thus greatly improving the effectiveness of feature extraction. Through the sag classification model, voltage sag recognition processing is performed on the basic feature and the curvature feature to obtain the recognition type of the voltage sag. In this way, by processing the basic feature and the curvature feature through the sag classification model, the change process of the voltage sag can be comprehensively and accurately reflected, so that the recognition type of the voltage sag can be accurately determined, and the accuracy of voltage sag recognition is greatly improved.

[0068] In one embodiment, as Figure 5 shown, the basic feature includes at least one of three - phase unbalance degree, sag depth, and sag duration. The feature extraction is performed on the target voltage data corresponding to each sag moment respectively to obtain a basic feature for characterizing the voltage sag and a curvature feature for characterizing the curvature change of the voltage sag, including:

[0069] Step S502: Determine the three-phase unbalance degree and the sag depth based on the target voltage data corresponding to each sag moment and the standard voltage data.

[0070] Specifically, the computer device obtains the target voltage data corresponding to each sag moment and the standard voltage data. The computer device determines the instantaneous voltage information corresponding to each sag moment based on the target voltage data corresponding to each sag moment. For each sag moment, the computer device calculates the three-phase unbalance degree and the sag depth respectively for the instantaneous voltage information corresponding to the corresponding sag moment based on the labeled voltage data, and obtains the three-phase unbalance degree and the sag depth corresponding to each sag moment.

[0071] For example, for the sag moment t, the computer device obtains the target voltage data A corresponding to the sag moment t, and determines the instantaneous voltage information a corresponding to the sag moment t based on the target voltage data A. The computer device calculates the three-phase unbalance degree and the sag depth respectively based on the instantaneous voltage information a and the standard voltage data, and obtains the three-phase unbalance degree and the sag depth of the sag moment t.

[0072] It can be understood that in this application, each three-phase unbalance degree corresponds to a sag moment, that is, this three-phase unbalance degree does not correspond to a long-period or short-period time period, but corresponds to a sag moment.

[0073] Step S504: Determine the voltage sag start time and the voltage sag end time based on each sag moment in the sag time period, and determine the sag duration based on the voltage sag start time and the voltage sag end time.

[0074] Specifically, the computer device obtains the sag time period, and based on each sag moment within the sag time period, the computer device takes the moment when the three-phase voltage data rapidly drops from the rated data to a predetermined value as the voltage sag start time, and takes the moment when the three-phase voltage data rises back to the predetermined value as the voltage sag end time. The computer device determines the sag duration based on the voltage sag start time and the voltage sag end time.

[0075] For example, as Figure 6 shown, the computer device obtains the sag time period, and based on each sag moment within the sag time period, the computer device takes the moment when the three-phase voltage data rapidly drops from the rated data U N to 0.9 p.u. as the voltage sag start time T1, and takes the moment when the three-phase voltage data rises back to 0.9 p.u. as the voltage sag end time T2. The computer device subtracts the voltage sag start time from the voltage sag end time to obtain the sag duration △T.

[0076] Step S506: For each sag moment, calculate the phase space radius for the three-phase voltage data corresponding to the respective sag moment to obtain the phase space vector corresponding to the respective sag moment.

[0077] Among them, the phase space is a space used to represent all possible states of a system, and each possible state of the system has a corresponding point in the phase space.

[0078] Specifically, for each sag moment, the computer device calculates the phase space radius for the three-phase voltage data corresponding to the respective sag moment to obtain the phase space vector corresponding to the respective sag moment. For example, the target voltage data A corresponding to the sag moment t contains instantaneous voltage information a, and the instantaneous voltage information a is a set of three target sub-voltage values, namely ν a (k), ν b (k), ν c (k) (corresponding to the per-unit values of the three-phase instantaneous voltages), and the phase space vector x(k) corresponding to the sag moment t is calculated using the following formula:

[0079]

[0080] Among them, taking the amplitude of the phase space vector can obtain R(k), that is, the phase space radius.

[0081] Step S508: Perform threshold screening on the phase space vectors corresponding to the respective sag moments to determine the target sag moment, and calculate the slope of the target voltage data corresponding to the target sag moment to obtain the curvature feature.

[0082] Specifically, the computer device obtains the phase space vectors corresponding to the respective sag moments and performs threshold screening on each phase space vector to obtain the screened phase space vectors. The computer device determines the target sag moment based on the screened phase space vectors. The computer device determines the target voltage data corresponding to the target sag moment and calculates the slope of the target voltage data to obtain the curvature feature.

[0083] In this embodiment, based on the target voltage data and the standard voltage data corresponding to each sag moment, the three-phase unbalance degree and the sag depth corresponding to each sag moment are determined. Thus, a three-phase unbalance degree and a sag depth with better real-time performance can be obtained, which is beneficial to obtaining more targeted basic features. Based on the voltage sag start time and the voltage sag end time corresponding to the sag time period, the sag duration can be accurately and quickly obtained. For each sag moment, by calculating the phase space radius of the three-phase voltage data corresponding to the corresponding sag moment, a phase space vector corresponding to the corresponding sag moment is obtained. In this way, the phase space vector can accurately reflect the change of the curvature during the voltage sag. By screening each phase space vector that truthfully reflects the curvature change, the target sag moment is obtained. By calculating the slope of the target voltage data corresponding to the target sag moment, the curvature feature is obtained. In this way, based on each phase space vector that can truthfully reflect the curvature change, the change process of the voltage sag can be clearly and accurately reflected through slope calculation, ensuring the effectiveness of feature extraction.

[0084] In one embodiment, as Figure 7 shown, the target voltage data includes a plurality of sub-voltage values. The determining of the three-phase unbalance degree and the sag depth based on the target voltage data and the standard voltage data corresponding to each sag moment includes:

[0085] Step S702, for each sag moment, based on the target voltage data corresponding to the corresponding sag moment, determine a plurality of target sub-voltage values corresponding to the corresponding sag moment.

[0086] Wherein, the target voltage data is three-phase voltage data, and each target voltage data includes multiple-phase voltages. Among them, each phase corresponds to a target sub-voltage value. Among them, three-phase electricity consists of a power supply composed of three AC electromotive forces with the same frequency, equal amplitude, and a phase difference of 120° in sequence.

[0087] Specifically, for each sag moment, the computer device determines a plurality of target sub-voltage values corresponding to the corresponding sag moment based on the target voltage data corresponding to the corresponding sag moment. For example, the target voltage data A corresponding to the sag moment t includes instantaneous voltage information a, and the instantaneous voltage information a is a set of three target sub-voltage values.

[0088] Step S704, for each sag moment, determine the maximum value and the minimum value from the plurality of target sub-voltage values corresponding to the corresponding sag moment, and determine a first difference corresponding to the corresponding sag moment based on the maximum value and the minimum value.

[0089] Specifically, for each sag moment, the computer device compares multiple target sub-voltage values corresponding to the respective sag moment to obtain a maximum value and a minimum value. For each sag moment, the computer device subtracts the minimum value from the maximum value to obtain a first difference corresponding to the respective sag moment.

[0090] Step S706: For each sag moment, based on the first difference corresponding to the respective sag moment and the standard voltage data, determine the three-phase unbalance degree corresponding to the respective sag moment.

[0091] Specifically, for each sag moment, the computer device divides the first difference corresponding to the respective sag moment by the standard voltage value to obtain a ratio value corresponding to the respective sag moment. The computer device uses the ratio value corresponding to the respective sag moment as the three-phase unbalance degree corresponding to the respective sag moment.

[0092] For example, for each sag moment, the computer device compares multiple target sub-voltage values corresponding to the respective sag moment to obtain a maximum value and a minimum value where is the target voltage data (i.e., the sag three-phase sag voltage amplitude), and a, b, c represent the three phases. For each sag moment, the computer device calculates the three-phase unbalance degree based on the maximum value and the minimum value, specifically using the following formula:

[0093]

[0094] It should be noted that according to the principle that the effective value is taken as a sliding window, the target voltage data is calculated at the monitoring point one period after the first point where the effective value is less than 0.95 p.u. U N is the rated operating voltage for monitoring data acquisition, i.e., the standard voltage data.

[0095] Step S708: For each sag moment, based on the minimum value corresponding to the respective sag moment and the standard voltage data, determine a second difference corresponding to the respective sag moment, and based on the second difference corresponding to the respective sag moment and the standard voltage data, determine the sag depth corresponding to the respective sag moment.

[0096] Specifically, for each sag moment, the computer device subtracts the minimum value corresponding to the respective sag moment from the standard voltage data to obtain a second difference corresponding to the respective sag moment. For each sag moment, the computer device divides the second difference corresponding to the respective sag moment by the standard voltage data to obtain the sag depth corresponding to the respective sag moment. For example, for each sag moment, the computer device is based on the minimum value U m (i.e., corresponding to ), the following formula is adopted:

[0097]

[0098] In this embodiment, for each sag moment, based on a plurality of target sub-voltage values corresponding to the corresponding sag moment, a more instantaneous three-phase unbalance degree is calculated through the three-phase unbalance degree. For each sag moment, based on the minimum value corresponding to the corresponding sag moment, a better real-time sag depth is calculated through the sag depth. Thus, a better real-time three-phase unbalance degree and sag depth can be obtained, which is beneficial to obtaining better targeted basic features.

[0099] In one embodiment, threshold screening is performed on the phase space vectors corresponding to each sag moment to determine the target sag moment, including: determining a first target vector smaller than the vector threshold from a plurality of phase space vectors, and determining a second target vector greater than or equal to the vector threshold from a plurality of phase space vectors. Determine the first sag moment corresponding to each first target vector, and determine the second sag moment corresponding to each second target vector. Determine the first target sag moment based on the time sequence of each first sag moment, and determine the second target sag moment based on the time sequence of each second sag moment. Based on the sampling period, determine a third target sag moment corresponding to the first target sag moment and a fourth target sag moment corresponding to the second target sag moment. Take the first target sag moment, the second target sag moment, the third target sag moment, and the fourth target sag moment as the target sag moment.

[0100] Specifically, the computer device obtains a plurality of phase space vectors, takes the amplitude of each phase space vector, and obtains the phase space radius corresponding to the corresponding phase space vector. The computer device takes the phase space vector corresponding to the phase space radius smaller than the vector threshold as the first target vector. The computer device takes the phase space vector corresponding to the phase space radius greater than or equal to the vector threshold as the second target vector. The computer device determines the first sag moment corresponding to each first target vector and determines the second sag moment corresponding to each second target vector. The computer device sorts each first sag moment in ascending order and takes the first sorted first sag moment as the first target sag moment. The computer device sorts each second sag moment in ascending order and takes the first sorted second sag moment as the second target sag moment. The computer device advances the first target sag moment by one sampling period to obtain the third target sag moment, and advances the second target sag moment by one sampling period to obtain the fourth target sag moment. The computer device takes the first target sag moment, the second target sag moment, the third target sag moment, and the fourth target sag moment as the target sag moment.

[0101] For example, the computer device obtains the phase space radius R(k) corresponding to each phase space vector, and takes the first sag moment when R(k) is less than 0.95 as the first target sag moment t1. The computer device takes the first sag moment when R(k) returns to above 0.95 as the second target sag moment t2. The computer device takes the sag moment of the first target sag moment t1 in the next period T (i.e., t3 = t1 + T) as the third target sag moment t3. The computer device takes the sag moment of the second target sag moment t2 in the previous period T (i.e., t4 = t2 - T) as the fourth target sag moment t4. The computer device takes the first target sag moment, the second target sag moment, the third target sag moment, and the fourth target sag moment as the target sag moments.

[0102] It should be noted that the phase space vector calculated based on the phase space radius can be represented by a circle. That is, if a voltage sag occurs, the calculated phase space vector is represented as a circle with a decreasing radius, and at this time, the radius instantaneously decreases. Therefore, the phase space vector can well represent the curve corresponding to the time point, that is, it can better reflect the curve details.

[0103] In this embodiment, by calculating the phase space vectors corresponding to each sag moment, the curves corresponding to each sag moment can be obtained. By screening the phase space vectors with a vector threshold, the target sag moments with better curve details can be obtained. In this way, based on each sag moment, the change of the curvature during the voltage sag can be accurately and precisely reflected.

[0104] In one embodiment, calculating the slope of the target voltage data corresponding to the target sag moment to obtain the curvature feature includes: for each target sag moment, determining the target curvature corresponding to the corresponding target sag moment based on the first adjacent sag moment and the first adjacent target voltage data corresponding to the corresponding target sag moment; where the target sag moment includes the first target sag moment, the second target sag moment, the third target sag moment, and the fourth target sag moment. For each sag moment, determining the curvature to be processed corresponding to the corresponding sag moment based on the second adjacent sag moment and the second adjacent target voltage data corresponding to the corresponding sag moment, and calculating the average value of the multiple curvatures to be processed to obtain the target mean. Combining the multiple target curvatures and the target mean to obtain the curvature feature.

[0105] Specifically, for each target sag moment, the computer device determines the first adjacent sag moment corresponding to the corresponding target sag moment, and determines the first adjacent target voltage data corresponding to each of the first adjacent sag moments. For each target sag moment, the computer device determines the first voltage difference based on the multiple first adjacent target voltage data, and determines the first time difference based on the multiple first adjacent sag moments. For each target sag moment, the computer device obtains the target curvature corresponding to the corresponding target sag moment based on the first voltage difference and the first time difference corresponding to the corresponding target sag moment. For each sag moment, the computer device calculates the curvature to be processed corresponding to the corresponding sag moment based on the second adjacent sag moment and the second adjacent target voltage data corresponding to the corresponding sag moment. The computer device calculates the average value of the multiple curvatures to be processed to obtain the target average value. The computer device combines the multiple target curvatures and the target average value to obtain the curvature feature.

[0106] For example, the curvature calculation formula is as follows:

[0107]

[0108] Where U i and U i+1 are the effective values of the i-th sampling point (i.e., the target sag moment) and the next point (i.e., the adjacent sag moment adjacent to the target sag moment), respectively. The k i is the curvature of the target sag moment. Wherein, the target sag moment includes the first target sag moment, the second target sag moment, the third target sag moment, and the fourth target sag moment, and the distribution of each target sag moment is as Figure 8 shown. Wherein, k i includes the sag moment information T i and the target voltage data U i . Where k1 and k2 are less than 0, k3 and k4 are greater than 0, and the target average value k5 is the average value of the sum of the absolute values of the slopes of the voltage sag curves during the whole process.

[0109] In this embodiment, based on the target voltage data corresponding to each target sag moment, the target curvature corresponding to each target sag moment can be accurately determined. Based on the adjacent target voltage data corresponding to each sag moment, the comprehensive situation of the curvature change during voltage sag, that is, the target average value, can be determined. In this way, by combining the target curvature corresponding to each target sag moment and the target average value that can represent the comprehensive change of the curvature, an effective curvature feature can be obtained, and thus, the change of the curvature during voltage sag can be accurately and precisely reflected.

[0110] In one embodiment, as Figure 9 shown, the determination steps of the sag classification model include:

[0111] In step S902, short-time Fourier processing is performed on multiple sample three-phase voltage data within the sample sag time period to obtain multiple sample voltage data, and feature extraction is performed on the multiple sample three-phase voltage data to obtain sample basic features and sample curvature features.

[0112] Specifically, the computer device obtains multiple sample three-phase voltage data within the sample sag time period, determines a short-time Fourier expression based on the Fourier expression, and performs an update process on the short-time Fourier expression to obtain an updated short-time Fourier expression. The computer device, based on the multiple sample three-phase voltage data, obtains sample voltage data corresponding to each sample sag moment through the updated short-time Fourier expression. The computer device calculates multiple basic indicators for the multiple sample voltage data to obtain sample basic features corresponding to each basic indicator calculation. The computer device performs phase space vector calculation and curvature calculation respectively based on the multiple sample voltage data to determine the sample curvature features.

[0113] In step S904, a sag classification model composed of multiple decision trees is constructed, and based on the multiple sample voltage data, iterative calculations of a depth restriction generation strategy are performed on the decision trees in the sag classification model until the iterative stop condition is met and then stopped.

[0114] Among them, the sag classification model is a model constructed based on the LightGBM (Light Gradient Boosting Machine) algorithm.

[0115] Specifically, the computer device constructs a sag classification model composed of multiple decision trees, and the computer device performs iterative calculations of a depth restriction generation strategy on the decision trees in the sag classification model based on the multiple sample voltage data until the iterative stop condition is met and then stopped. Among them, the iterative stop condition can be that the number of iterations reaches the iteration threshold, etc. Among them, the sag classification model is constructed based on a classification and recognition algorithm containing multiple decision trees, and the sag classification model is an optimized classification model, that is, the sag classification model is optimized through a histogram optimization part, a storage and memory optimization part, and a generation strategy optimization part.

[0116] Among them, histogram optimization improves the traditional pre-sorting method, divides continuous values into a series of discrete domains, and one domain corresponds to a histogram block in the dataset. The values in this interval are called bins, and then a histogram with these bins as the accuracy unit is used. Compared with the traditional pre-sorting, it simplifies the model expression, reduces memory usage. At the same time, the histogram has a certain regularization effect, reduces the recognition error, and can effectively improve the accuracy of voltage sag classification and recognition. The structure diagram of its histogram optimization is asFigure 10 As shown in a. Among them, when storing memory optimization describes data characteristics, the pre-sorting algorithm is also improved to store the data of each row and each column. When there are structures with highly similar data characteristics, the distinguishing boundaries between the features are directly blurred. For example Figure 10 For the grayscale image in b, the 4×4 data feature arrangement is directly optimized into data with a calculation cost of 0 for the whole image, that is, the discretized values are converted into integer storage. Therefore, the memory occupancy of the entire algorithm is reduced to 1 / 8 of the original. Among them, the growth strategy optimization part adopts the Leaf-wise tree growth strategy with depth limitation. This strategy can grow deeper leaves on the feature data, which means a more in-depth analysis of the features. However, it is prone to overfitting, resulting in a decrease in the accuracy of voltage sag classification and recognition. Therefore, it can improve the model optimization efficiency. The comparison diagram between the Leaf-wise tree growth strategy with depth limitation and the level-wise tree growth strategy is as Figure 10 shown in c.

[0117] Step S906: Obtain the trained sag classification model based on the decision trees in the last iterative calculation and the weights corresponding to each decision tree respectively.

[0118] Specifically, the computer device obtains the decision trees obtained from the last iterative calculation and the weights corresponding to each decision tree respectively, and performs weighted summation on each decision tree and the weights corresponding to each decision tree to determine the trained sag classification model.

[0119] For example, if the iteration stop condition is satisfied, the computer device obtains each decision tree in the last iterative calculation, that is, each trained decision tree. Among them, a decision tree can be regarded as a weak classifier. The computer device obtains the weights corresponding to each trained decision tree respectively, and processes each trained decision tree through weighted summation to obtain the trained sag classification model. Among them, the trained decision tree can be regarded as a strong classifier, that is, a strong classifier obtained by weighted integration of each trained weak classifier, as specifically shown below:

[0120] F m (x)=a0f0(x)+a1f1(x)+…+a i f i (x)+…+a m f m (x)

[0121] Among them, f i (x) represents a weak classifier, and a i represents the weight of the weak classifier.

[0122] Step S908: Train the trained sag classification model by using multiple test three-phase voltage data during the test sag time period to determine a sag classification model for voltage sag identification.

[0123] Specifically, the computer device obtains multiple test three-phase voltage data during the test sag time period, determines a short-time Fourier expression based on the Fourier expression, and performs an update process on the short-time Fourier expression to obtain an updated short-time Fourier expression. The computer device obtains test voltage data corresponding to each sample sag moment respectively through the updated short-time Fourier expression based on the multiple test three-phase voltage data. The computer device performs multiple basic index calculations on the multiple test voltage data to obtain test basic features corresponding to each basic index calculation respectively. The computer device performs phase space vector calculation and curvature calculation respectively based on the multiple test voltage data to determine test curvature features. The computer device trains the trained sag classification model based on the test basic features and test features to determine a sag classification model for voltage sag identification.

[0124] In this embodiment, sample basic features and sample curvature features for training the sag classification model are determined based on multiple sample voltage data obtained by short-time Fourier processing. Based on the sample basic features and sample curvature features, the sag classification model composed of a histogram optimization part, a storage memory optimization part, and a generation strategy optimization part is trained and tested to obtain a sag classification model for voltage sag identification. In this way, the machine learning speed is greatly improved based on the histogram optimization technology, and the memory occupancy is reduced by using the storage memory technology. In this way, while ensuring the recognition accuracy of voltage sags, the recognition speed of voltage sags can also be greatly improved.

[0125] In one embodiment, as Figure 11As shown, the computer device performs short-time Fourier processing on multiple sample three-phase voltage data during the sample voltage sag period to obtain multiple sample voltage data, and extracts features from the multiple sample three-phase voltage data to obtain sample basic features and sample curvature features. The computer device initializes the number of decision trees and sets the weights of the training samples; initializes the number of iterations and sets the maximum number of iterations. The computer device calculates the gradient value based on the extracted sample basic features and sample curvature features, and determines the optimal voltage sag classification point according to the training data. Each decision tree is trained using the depth-limited Leaf-wise tree growth strategy, and the weights are determined with the goal of minimizing the error. The computer device performs weighted summation on each decision tree to obtain the trained voltage sag classification model, which is a strong classifier. The test data is input into the trained voltage sag classification and recognition model to verify the recognition result, thereby determining the category of the voltage sag. Among them, the category is one of short-circuit faults, large motor startups, or transformer switching.

[0126] Among them, the voltage sag classification model contains a classification and recognition algorithm. This classification and recognition algorithm first performs sample normalization, that is, let the sample set be T, where x i is the sample, and y i is the label. The normalization process for the voltage sag data is as follows:

[0127]

[0128] where υ ij corresponds to the data expression after normalization. Based on this normalized expression, the initial gradient value is calculated, and the expression is as follows:

[0129]

[0130] Then, gradient boosting is performed based on this initial gradient value, that is, the residual of the voltage sag classification model is used to train a new classifier. After that, the trained classifier is added to the voltage sag classification model of the current iteration, and multiple iterations are performed to obtain the trained voltage sag classification model. Among them, the process of gradient training is as Figure 12 shown. Among them, the histogram is calculated based on the gradient value calculation formula, and the formula is as follows:

[0131]

[0132] In the formula

[0133] Finally, the trained voltage sag classification model is trained based on the test set samples to obtain a voltage sag classification model for voltage sag recognition.

[0134] In this embodiment, based on a plurality of sample voltage data obtained by short-time Fourier processing, sample basic features and sample curvature features for training a sag classification model are determined. Based on the sample basic features and sample curvature features, the sag classification model composed of a histogram optimization part, a storage memory optimization part, and a generation strategy optimization part is trained and tested to obtain a sag classification model for voltage sag identification. In this way, based on the histogram optimization technology, the speed of machine learning is greatly improved, and the memory occupancy is reduced by using the storage memory technology. In this way, while ensuring the recognition accuracy of voltage sags, the recognition speed of voltage sags can be greatly improved.

[0135] To facilitate a clearer understanding of the technical solution of this application, a more detailed embodiment is provided for explanation. If the computer device receives a voltage sag signal sent by the power system, the computer device acquires a plurality of three-phase voltage data during the sag time period. The computer device performs an update process on the short-time Fourier expression to obtain an updated short-time Fourier expression. The computer device processes the plurality of three-phase voltage data through the updated short-time Fourier expression to obtain target voltage data corresponding to each sag moment respectively. For each sag moment, the computer device determines a plurality of target sub-voltage values corresponding to the corresponding sag moment based on the target voltage data corresponding to the corresponding sag moment. For each sag moment, the computer device compares the plurality of target sub-voltage values corresponding to the corresponding sag moment to obtain a maximum value and a minimum value. For each sag moment, the computer device subtracts the minimum value from the maximum value to obtain a first difference corresponding to the corresponding sag moment. For each sag moment, the computer device divides the first difference corresponding to the corresponding sag moment by the standard voltage value to obtain a ratio value corresponding to the corresponding sag moment. The computer device uses the ratio value corresponding to the corresponding sag moment as the three-phase unbalance degree corresponding to the corresponding sag moment. For each sag moment, the computer device subtracts the minimum value corresponding to the corresponding sag moment from the standard voltage data to obtain a second difference corresponding to the corresponding sag moment. For each sag moment, the computer device divides the second difference corresponding to the corresponding sag moment by the standard voltage data to obtain a sag depth corresponding to the corresponding sag moment. Based on each sag moment within the sag time period, the computer device takes the moment when the three-phase voltage data rapidly drops from the rated data to a predetermined value as the voltage sag start moment, and takes the moment when the three-phase voltage data rises back to the predetermined value as the voltage sag end moment. The computer device determines the sag duration based on the voltage sag start moment and the voltage sag end moment. For each sag moment, the computer device calculates the phase space radius for the three-phase voltage data corresponding to the corresponding sag moment to obtain a phase space vector corresponding to the corresponding sag moment. The computer device takes the amplitude of each phase space vector to obtain a phase space radius corresponding to the corresponding phase space vector. The computer device uses the phase space vector corresponding to the phase space radius less than the vector threshold as the first target vector, and uses the phase space vector corresponding to the phase space radius greater than or equal to the vector threshold as the second target vector. The computer device determines the first sag moments corresponding to the respective first target vectors and determines the second sag moments corresponding to the respective second target vectors. The computer device sorts the respective first sag moments in ascending order and takes the first sorted first sag moment as the first target sag moment. The computer device sorts the respective second sag moments in ascending order and takes the first sorted second sag moment as the second target sag moment.The computer device delays the first target sag moment by one sampling period to obtain the third target sag moment, and advances the second target sag moment by one sampling period to obtain the fourth target sag moment. The computer device uses the first target sag moment, the second target sag moment, the third target sag moment, and the fourth target sag moment as the target sag moments. For each target sag moment, the computer device determines the first adjacent sag moment corresponding to the corresponding target sag moment, and determines the first adjacent target voltage data corresponding to each of the first adjacent sag moments. For each target sag moment, the computer device determines the first voltage difference based on the multiple first adjacent target voltage data, and determines the first time difference based on the multiple first adjacent sag moments. For each target sag moment, the computer device obtains the target curvature corresponding to the corresponding target sag moment based on the first voltage difference and the first time difference corresponding to the corresponding target sag moment. For each sag moment, the computer device calculates the curvature to be processed corresponding to the corresponding sag moment based on the second adjacent sag moment and the second adjacent target voltage data corresponding to the corresponding sag moment. The computer device calculates the average value of the multiple curvatures to be processed to obtain the target mean value. The computer device combines the multiple target curvatures and the target mean value to obtain the curvature feature.

[0136] Perform short-time Fourier processing on multiple sample three-phase voltage data within the sample sag time period to obtain multiple sample voltage data, and perform feature extraction on the multiple sample three-phase voltage data to obtain sample basic features and sample curvature features. Construct a sag classification model composed of multiple decision trees, and perform iterative calculations of the depth restriction generation strategy on the decision trees in the sag classification model based on the multiple sample voltage data until the iteration stop condition is met and then stop. Obtain the trained sag classification model based on the decision trees in the last iterative calculation and the weights corresponding to each decision tree respectively. The computer device obtains the basic feature and the curvature feature, and inputs the basic feature and the curvature feature into the sag classification model to obtain the recognition type of the voltage sag.

[0137] In this embodiment, if a voltage sag signal sent by the power system is received, short-time Fourier processing is performed on multiple three-phase voltage data during the sag time period to obtain multiple processed target voltage data; wherein, each sag moment in the sag time period corresponds to a target voltage data. Feature extraction is performed on the target voltage data respectively corresponding to each sag moment to obtain a basic feature for characterizing voltage sag and a curvature feature for characterizing the curvature change of voltage sag; wherein, the basic feature includes at least one of three-phase unbalance degree, sag depth, and sag duration. In this way, on the basis of the basic feature, combined with the curvature feature for characterizing the curvature change of voltage sag, the detailed information of voltage sag can be further expanded, thus greatly improving the effectiveness of feature extraction. Through the sag classification model, voltage sag identification processing is performed on the basic feature and the curvature feature to obtain the identification type of voltage sag. In this way, by processing the basic feature and the curvature feature through the sag classification model, the change process of voltage sag can be comprehensively and accurately reflected, so that the identification type of voltage sag can be accurately determined, and the accuracy of voltage sag identification is greatly improved. [[ID=> ]][[ID=>

[0138] ]]It should be understood that although the steps in the flowcharts involved in the above-described embodiments are displayed sequentially according to the arrows, these steps are not necessarily executed sequentially in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps is not strictly restricted by order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or steps or stages in other steps. [[ID=> ]][[ID=>

[0139] ]]Based on the same inventive concept, the embodiments of the present application also provide a voltage sag identification device for implementing the voltage sag identification method described above. The solution provided by this device for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the voltage sag identification device provided below can refer to the limitations on the voltage sag identification method in the above text, and will not be repeated here. [[ID=> ]][[ID=>

[0140] ]]In one embodiment, as [[ID=> Figure 13 ]]shown, a voltage sag identification device is provided, including: a processing module 1302, an extraction module 1304, and an identification module 1306, wherein: [[ID=> ]][[ID=>

[0141] ]]A processing module 1302, configured to perform short-time Fourier processing on multiple three-phase voltage data during a sag period if a voltage sag signal sent by a power system is received, so as to obtain multiple processed target voltage data; wherein, each sag moment during the sag period corresponds to a target voltage data.

[0142] An extraction module 1304, configured to extract features from the target voltage data respectively corresponding to each sag moment, so as to obtain a basic feature for characterizing voltage sag and a curvature feature for characterizing the curvature change of voltage sag; wherein, the basic feature includes at least one of three-phase unbalance degree, sag depth, and sag duration.

[0143] An identification module 1306, configured to perform voltage sag identification processing on the basic feature and the curvature feature through a sag classification model, so as to obtain an identification type of voltage sag.

[0144] In one embodiment, the extraction module 1304 is configured to determine the three-phase unbalance degree and the sag depth based on the target voltage data respectively corresponding to each sag moment and the standard voltage data. Based on each sag moment in the sag period, determine the voltage sag start moment and the voltage sag end moment, and determine the sag duration based on the voltage sag start moment and the voltage sag end moment. For each sag moment, calculate the phase space radius of the three-phase voltage data corresponding to the corresponding sag moment to obtain a phase space vector corresponding to the corresponding sag moment. Perform threshold screening on the phase space vectors respectively corresponding to each sag moment to determine the target sag moment, and calculate the slope of the target voltage data corresponding to the target sag moment to obtain the curvature feature.

[0145] In one embodiment, the extraction module 1304 is configured to, for each sag moment, determine multiple target sub-voltage values corresponding to the corresponding sag moment based on the target voltage data corresponding to the corresponding sag moment. For each sag moment, determine the maximum value and the minimum value from the multiple target sub-voltage values corresponding to the corresponding sag moment, and determine a first difference corresponding to the corresponding sag moment based on the maximum value and the minimum value. For each sag moment, determine the three-phase unbalance degree corresponding to the corresponding sag moment based on the first difference corresponding to the corresponding sag moment and the standard voltage data. For each sag moment, determine a second difference corresponding to the corresponding sag moment based on the minimum value corresponding to the corresponding sag moment and the standard voltage data, and determine the sag depth corresponding to the corresponding sag moment based on the second difference corresponding to the corresponding sag moment and the standard voltage data.

[0146] In one embodiment, the extraction module 1304 is configured to determine first target vectors smaller than a vector threshold from multiple phase space vectors, and determine second target vectors greater than or equal to the vector threshold from the multiple phase space vectors. Determine the first sag moments corresponding to the respective first target vectors, and determine the second sag moments corresponding to the respective second target vectors. Determine a first target sag moment based on the time sequence of the respective first sag moments, and determine a second target sag moment based on the time sequence of the respective second sag moments. Based on the sampling period, determine a third target sag moment corresponding to the first target sag moment, and determine a fourth target sag moment corresponding to the second target sag moment. Use the first target sag moment, the second target sag moment, the third target sag moment, and the fourth target sag moment as the target sag moments.

[0147] In one embodiment, for each target sag moment, the extraction module 1304 is configured to determine a target curvature corresponding to the respective target sag moment based on a first adjacent sag moment and first adjacent target voltage data corresponding to the respective target sag moment; wherein the target sag moments include a first target sag moment, a second target sag moment, a third target sag moment, and a fourth target sag moment. For each sag moment, determine a curvature to be processed corresponding to the respective sag moment based on a second adjacent sag moment and second adjacent target voltage data corresponding to the respective sag moment, and calculate an average value of the multiple curvatures to be processed to obtain a target average value. Combine the multiple target curvatures and the target average value to obtain a curvature feature.

[0148] In one embodiment, the recognition module 1306 is configured to perform short-time Fourier processing on multiple sample three-phase voltage data within a sample sag time period to obtain multiple sample voltage data, and perform feature extraction on the multiple sample three-phase voltage data to obtain sample basic features and sample curvature features. Construct a sag classification model composed of multiple decision trees, and perform iterative calculations of a depth-limited generation strategy on the decision trees in the sag classification model based on the multiple sample voltage data until the iteration stop condition is met and then stop. Obtain a trained sag classification model based on the decision trees in the last iterative calculation and the weights corresponding to the respective decision trees. Train the trained sag classification model by using multiple test three-phase voltage data within a test sag time period to determine a sag classification model for voltage sag recognition.

[0149] Each module in the above voltage sag recognition device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in a processor in a computer device in a hardware form or be independent of the processor, or can be stored in a memory in the computer device in a software form so that the processor can call and execute the operations corresponding to the above respective modules.

[0150] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structural diagram may be as shown in Figure 14 the following figure. The computer device includes a processor, a memory, and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store identification data of voltage sags. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a method for identifying voltage sags.

[0151] Those skilled in the art can understand that Figure 14 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0152] In one embodiment, a computer device is further provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0153] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0154] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0155] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data that have been authorized by the user or fully authorized by all parties.

[0156] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include Read-Only Memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0157] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0158] The above-described embodiments only represent several implementation manners of the present application. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for identifying voltage sags, characterized in that, The method includes: If a voltage sag signal sent by a power system is received, short-time Fourier processing is performed on a plurality of three-phase voltage data during the sag time period to obtain a plurality of processed target voltage data; wherein, each sag moment during the sag time period corresponds to a target voltage data, and the target voltage data includes a plurality of sub-voltage values; For each sag moment, based on the target voltage data corresponding to the corresponding sag moment, a plurality of target sub-voltage values corresponding to the corresponding sag moment are determined; For each sag moment, the maximum value and the minimum value are determined from the plurality of target sub-voltage values corresponding to the corresponding sag moment, and a first difference corresponding to the corresponding sag moment is determined based on the maximum value and the minimum value; For each sag moment, based on the first difference corresponding to the corresponding sag moment and the standard voltage data, the three-phase unbalance degree corresponding to the corresponding sag moment is determined; For each sag moment, based on the minimum value corresponding to the corresponding sag moment and the standard voltage data, a second difference corresponding to the corresponding sag moment is determined, and based on the second difference corresponding to the corresponding sag moment and the standard voltage data, the sag depth corresponding to the corresponding sag moment is determined; Based on each sag moment in the sag time period, the voltage sag start moment and the voltage sag end moment are determined, and the sag duration is determined based on the voltage sag start moment and the voltage sag end moment; For each sag moment, the phase space radius of the three-phase voltage data corresponding to the corresponding sag moment is calculated to obtain a phase space vector corresponding to the corresponding sag moment; Threshold screening is performed on the phase space vectors corresponding to each sag moment respectively to determine the target sag moment, and the slope of the target voltage data corresponding to the target sag moment is calculated to obtain the curvature feature; Through a sag classification model, voltage sag identification processing is performed on the basic features and the curvature feature to obtain the identification type of the voltage sag, and the basic features include at least one of the three-phase unbalance degree, the sag depth, and the sag duration.

2. The method according to claim 1, characterized in that, The threshold screening of the phase space vectors corresponding to each sag moment respectively to determine the target sag moment includes: Determining a first target vector smaller than the vector threshold from a plurality of phase space vectors, and determining a second target vector greater than or equal to the vector threshold from a plurality of phase space vectors; Determining the first sag moment corresponding to each first target vector respectively, and determining the second sag moment corresponding to each second target vector respectively; Determining the first target sag moment based on the time sequence of each first sag moment, and determining the second target sag moment based on the time sequence of each second sag moment; Based on the sampling period, determining a third target sag moment corresponding to the first target sag moment and determining a fourth target sag moment corresponding to the second target sag moment; Taking the first target sag moment, the second target sag moment, the third target sag moment, and the fourth target sag moment as the target sag moment.

3. The method according to claim 1, characterized in that The calculating the slope of the target voltage data corresponding to the target sag moment to obtain the curvature feature includes: For each target sag moment, a target curvature corresponding to the corresponding target sag moment is determined based on the first adjacent sag moment and the first adjacent target voltage data corresponding to the corresponding target sag moment; wherein, the target sag moment includes a first target sag moment, a second target sag moment, a third target sag moment, and a fourth target sag moment; For each sag moment, a curvature to be processed corresponding to the corresponding sag moment is determined based on the second adjacent sag moment and the second adjacent target voltage data corresponding to the corresponding sag moment, and an average value calculation is performed on multiple curvatures to be processed to obtain a target average value; Multiple target curvatures and the target average value are combined to obtain a curvature feature.

4. The method according to any one of claims 1 to 3, characterized in that, The determination steps of the sag classification model include: Performing short-time Fourier processing on multiple sample three-phase voltage data within a sample sag time period to obtain multiple sample voltage data, and performing feature extraction on the multiple sample three-phase voltage data to obtain sample basic features and sample curvature features; Constructing a sag classification model composed of multiple decision trees, and performing iterative calculations of a depth-limiting generation strategy on the decision trees in the sag classification model based on the multiple sample voltage data until the iteration stop condition is met and then stopping; Based on the decision trees in the last iterative calculation and the weights corresponding to each decision tree respectively, obtaining a trained sag classification model; Training the trained sag classification model through multiple test three-phase voltage data within a test sag time period to determine a sag classification model for voltage sag identification.

5. An identification device for voltage sags, characterized in that, The device includes: A processing module, configured to, if a voltage sag signal sent by a power system is received, perform short-time Fourier processing on multiple three-phase voltage data within a sag time period to obtain multiple processed target voltage data; wherein, each sag moment within the sag time period corresponds to a target voltage data, and the target voltage data includes multiple sub-voltage values; An extraction module, for each sag moment, based on the target voltage data corresponding to the corresponding sag moment, determine a plurality of target sub-voltage values corresponding to the corresponding sag moment; for each sag moment, determine the maximum value and the minimum value from the plurality of target sub-voltage values corresponding to the corresponding sag moment, and determine a first difference corresponding to the corresponding sag moment based on the maximum value and the minimum value; for each sag moment, based on the first difference corresponding to the corresponding sag moment and the standard voltage data, determine the three-phase unbalance degree corresponding to the corresponding sag moment; for each sag moment, based on the minimum value corresponding to the corresponding sag moment and the standard voltage data, determine a second difference corresponding to the corresponding sag moment, and based on the second difference corresponding to the corresponding sag moment and the standard voltage data, determine the sag depth corresponding to the corresponding sag moment; based on each sag moment in the sag time period, determine the voltage sag start moment and the voltage sag end moment, and determine the sag duration based on the voltage sag start moment and the voltage sag end moment; for each sag moment, calculate the phase space radius of the three-phase voltage data corresponding to the corresponding sag moment to obtain a phase space vector corresponding to the corresponding sag moment; perform threshold screening on the phase space vectors corresponding to each sag moment respectively to determine the target sag moment, and calculate the slope of the target voltage data corresponding to the target sag moment to obtain the curvature feature; An identification module, for performing voltage sag identification processing on the basic features and the curvature feature through a sag classification model to obtain the identification type of the voltage sag, where the basic features include at least one of the three-phase unbalance degree, the sag depth, and the sag duration.

6. The device according to claim 5, characterized in that The extraction module is used to determine a first target vector smaller than the vector threshold from a plurality of phase space vectors, and determine a second target vector greater than or equal to the vector threshold from a plurality of phase space vectors; determine the first sag moment corresponding to each first target vector, and determine the second sag moment corresponding to each second target vector; determine the first target sag moment based on the time sequence of each first sag moment, and determine the second target sag moment based on the time sequence of each second sag moment; based on the sampling period, determine a third target sag moment corresponding to the first target sag moment and determine a fourth target sag moment corresponding to the second target sag moment; use the first target sag moment, the second target sag moment, the third target sag moment, and the fourth target sag moment as the target sag moment.

7. The device according to claim 5, characterized in that, The extraction module is configured to, for each target sag moment, determine a target curvature corresponding to the corresponding target sag moment based on a first adjacent sag moment and first adjacent target voltage data corresponding to the corresponding target sag moment; wherein the target sag moments include a first target sag moment, a second target sag moment, a third target sag moment, and a fourth target sag moment; for each sag moment, determine a curvature to be processed corresponding to the corresponding sag moment based on a second adjacent sag moment and second adjacent target voltage data corresponding to the corresponding sag moment, and calculate an average value of the multiple curvatures to be processed to obtain a target average value; combine the multiple target curvatures and the target average value to obtain a curvature feature.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 4.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 4.

10. A computer program product, comprising a computer program, characterized in that, When this computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 4.

Images