A method, system, terminal and dielectric for extracting transient overvoltage features

By constructing an initial input feature set and using the Spearman correlation coefficient method to remove redundant features, combined with the p-value test method to determine the final input feature quantity, the problem of input feature redundancy affecting the power system transient voltage stability assessment model is solved, thus improving the performance and prediction efficiency of the assessment model.

CN118094189BActive Publication Date: 2026-11-10XI AN JIAOTONG UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410256477.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-03-06
Publication Date
2026-11-10
Estimated Expiration
2044-03-06

AI Technical Summary

Technical Problem

Redundancy exists among input features in existing technologies, affecting the performance and prediction efficiency of power system transient voltage stability assessment models. Furthermore, relying on human expertise to select variable features is difficult to adapt to the complex characteristics of large power grids.

Method used

By acquiring input features, an initial input feature set is constructed. Redundant features are eliminated using the Spearman correlation coefficient method, and the final input features are determined by combining the p-value test method. Key response features reflecting the transient overvoltage level are then extracted.

Benefits of technology

It reduces redundancy among high-dimensional electrical characteristic input features, improves the performance and prediction efficiency of the evaluation model, and adapts to the complex characteristics of large power grids.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118094189B_ABST
    Figure CN118094189B_ABST
Patent Text Reader

Abstract

This invention relates to the field of transient voltage technology in power systems, and discloses a method, system, terminal, and medium for extracting transient overvoltage features. The method includes acquiring input feature quantities; constructing an initial input feature set based on the input feature quantities; removing input feature quantities from the initial input feature set to obtain input feature quantities with low redundancy; and performing correlation feature verification on the input feature quantities with low redundancy to obtain the final input feature quantities. This process completes the transient overvoltage feature extraction, reduces redundancy between high-dimensional electrical characteristic input features, and obtains the final input feature quantities by verifying the correlation analysis values, thereby improving the performance and prediction efficiency of the evaluation model.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of transient voltage technology in power systems, specifically to a method, system, terminal, and medium for extracting transient overvoltage features. Background Technology

[0002] Under the strategic framework of "dual carbon" goals, the widespread inclusion of renewable energy and the commissioning of multiple ultra-high-voltage direct current (UHVDC) transmission projects have significantly altered the operating characteristics of the power system. Increased uncertainty between power sources and loads, coupled with the high proportion of power electronic devices, has profoundly impacted the dynamic characteristics of the power grid. Furthermore, reduced fault tolerance has exacerbated the risk of transient voltage instability, posing a threat to the safe and stable operation of the power system. Therefore, an effective strategy for rapidly implementing Transient Voltage Stability Assessment (TVSA) is urgently needed.

[0003] With the development of big data theory and artificial intelligence technology, and the rapid popularization of measurement equipment, data-driven artificial intelligence methods have made it possible to achieve rapid transient voltage stability assessment. However, due to the large scale and numerous components of actual power systems, high-dimensional electrical characteristics can lead to redundancy among input features, severely affecting the performance and prediction efficiency of assessment models. In recent years, in order to fully explore the key features under system disturbances, dynamic time series data have been applied to transient stability analysis, which increases the computational load and the risk of overfitting. Therefore, to promote the application of artificial intelligence methods in power system transient voltage stability assessment, it is still necessary to explore suitable feature selection strategies to reduce the dimensionality of the original features.

[0004] Traditional methods typically involve manually selecting key features and relying on expert experience to choose variable characteristics for stability analysis. However, due to the unclear fault characteristics of new equipment and the unclear transient voltage instability mechanisms, relying on manual expertise to select variable characteristics is increasingly difficult to adapt to the complex characteristics of large power grids. Summary of the Invention

[0005] In order to overcome the defects of the prior art, the present invention aims to provide a transient overvoltage feature extraction method, system, terminal and medium to solve the technical problem that redundancy exists between input features in the prior art, which affects the performance of the evaluation model and the prediction efficiency.

[0006] This invention is achieved through the following technical solution:

[0007] In a first aspect, the present invention provides a method for extracting transient overvoltage features, including...

[0008] Obtain the input feature values, and construct an initial input feature set based on the input feature values;

[0009] The input features in the initial input feature set are removed to obtain input features with low redundancy.

[0010] The input features with low redundancy are subjected to correlation feature testing to obtain the final input features, thus completing the transient overvoltage feature extraction.

[0011] Preferably, the input characteristic quantities are obtained by combining the main influencing factors of grid voltage instability or transient overvoltage level, wherein the input characteristic quantities include DC correlation response characteristic quantities characterizing before and after the fault and power output characteristic correlation response characteristic quantities of the new energy system.

[0012] Furthermore, key influencing factors include the ratio of renewable energy and conventional generating units connected to the power grid at the sending end, the transmission power of the DC system, and the output characteristics of the DC and renewable energy systems during faults.

[0013] Preferably, the input features in the initial input feature set are removed to obtain input features with low redundancy. The specific process is as follows:

[0014] S1, calculate several correlation coefficients between the input feature quantities in the initial input feature set and the degree of voltage change, and make a first condition judgment on the correlation coefficients. Perform a first stage elimination on the input feature quantities corresponding to the correlation coefficients that do not meet the first condition judgment to obtain the input feature quantities that are highly correlated with the degree of voltage change.

[0015] S2: For input features with high correlation to voltage changes, calculate several correlation coefficients pairwise. Then, perform a second condition judgment on these correlation coefficients. For any two input features corresponding to correlation coefficients that do not meet the second condition judgment, perform a second-stage elimination to obtain input features with low redundancy.

[0016] Furthermore, the Spearman correlation coefficient method was used to calculate the correlation coefficient, and the formula for calculating the Spearman correlation coefficient is as follows:

[0017]

[0018] In the formula: X i and Y i These represent the levels corresponding to the i-th data group, and These are the average ranks of variables x and y, respectively.

[0019] Furthermore, in S1, the specific process of making a first conditional determination on several correlation coefficients is as follows:

[0020] A first threshold is set. When the absolute values ​​of several correlation coefficients between the input feature quantities in the initial input feature set and the voltage change degree are less than the first threshold, the input feature quantities corresponding to the correlation coefficients are removed to obtain the input feature quantities that are highly correlated with the voltage change degree.

[0021] In S2, the specific process of making a second conditional determination on several correlation coefficients is as follows:

[0022] A second threshold is set. When the absolute value of several correlation coefficients obtained by pairwise calculation of the remaining input features is greater than the second threshold, either of the two input features corresponding to that correlation coefficient is removed to obtain input features with low redundancy.

[0023] Preferably, the p-value test method is used to perform correlation feature testing on input features with low redundancy. A threshold is set, and input features with p values ​​less than the threshold are taken as the final input features.

[0024] Secondly, the present invention provides a transient overvoltage feature extraction system, comprising:

[0025] The dataset construction module is used to obtain input feature quantities and construct an initial input feature set based on the input feature quantities.

[0026] The elimination module is used to eliminate input features in the initial input feature set to obtain input features with low redundancy.

[0027] The verification module is used to perform correlation feature verification on input features with low redundancy to obtain the final input features and complete the extraction of transient overvoltage features.

[0028] Thirdly, the present invention provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the steps of the transient overvoltage feature extraction method described above.

[0029] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the steps of the transient overvoltage feature extraction method described above.

[0030] Compared with the prior art, the present invention has the following beneficial technical effects:

[0031] This invention provides a method, system, terminal, and medium for extracting transient overvoltage features. It involves creating an initial input feature set from the acquired input features, then removing input features from the initial set to obtain input features with low redundancy. Input features corresponding to correlation coefficients that do not meet the judgment criteria are also removed, reducing redundancy among high-dimensional electrical characteristic input features. Finally, the final input features are obtained by verifying the correlation analysis values, thereby improving the performance and prediction efficiency of the evaluation model. Attached Figure Description

[0032] Figure 1 This is a flowchart of the method in this invention;

[0033] Figure 2 This is a schematic diagram of a DC power transmission system with wind power in an embodiment of the present invention;

[0034] Figure 3 This is a schematic diagram illustrating the changing characteristics of the system response in an embodiment of the present invention;

[0035] Figure 4 This is the structure of a local DC transmission system in an embodiment of the present invention;

[0036] Figure 5 This is a Spearman correlation heatmap from an embodiment of the present invention;

[0037] Figure 6 This refers to the deviation between the true value and the predicted value under different feature extraction states in the embodiments of the present invention;

[0038] Figure 7 This is a schematic diagram of the system principle structure in this invention;

[0039] In the diagram: 1-Dataset construction module; 2-Removal module; 3-Testing module. Detailed Implementation

[0040] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0041] The present invention will now be described in further detail with reference to the accompanying drawings:

[0042] The purpose of this invention is to provide a method, system, terminal, and medium for extracting transient overvoltage features, in order to solve the technical problem that redundancy exists between input features in the prior art, affecting the performance of evaluation models and prediction efficiency.

[0043] Example 1

[0044] Taking the extraction of key response features reflecting the transient overvoltage level at the sending end as an example, the following method is adopted: Figure 4The simulation example of a typical wind-thermal bundled DC transmission system in a localized area verifies the effectiveness of the proposed method. The test system is built on a 500kV grid structure, containing 210 nodes with a base capacity of 100MW. The installed capacity of wind turbines and conventional turbines is 4470MW and 8460MW, respectively. The system includes one ±500kV DC line with a rated active power transmission capacity of 5000MW.

[0045] See Figure 1 In one embodiment of the present invention, a method for extracting transient overvoltage features is provided, the specific process of which is as follows:

[0046] First, based on the analysis results of the main influencing factors of transient overvoltage levels and the selection principles of input data, relevant response features that can characterize the output characteristics of DC, new energy, and conventional units were selected to construct an initial input feature set, as shown in Table 1. Considering factors such as the proportion of renewable energy access and DC transmission power, multiple simulations were performed to complete sampling under DC commutation failure faults, obtaining multiple sets of data corresponding to different input features and transient overvoltage levels. For the operating mode, the output of conventional units and wind farms were adjusted at load levels of 90%, 100%, and 110%, respectively, and the proportion of new energy access was set to 30%, 35%, 40%, 45%, 50%, and 55%. The load model consisted of static load and induction motors in certain proportions, including 40%, 50%, and 60%. In addition, the DC transmission power was adjusted in 10% increments within the range of 60% to 100% of the rated power. For fault types, commutation failures included single-pole and double-pole faults. The number of commutation failures was set to 1, 2, and 3, respectively. Meanwhile, the duration of a single commutation failure ranges from 0.15s to 0.25s, with three durations set in 0.05s increments. The simulation time is 5s, and the total number of samples is 4480.

[0047] Table 1 Initial Input Feature Construction

[0048]

[0049]

[0050] Based on the proposed method, correlation analysis is performed on the sampled data to remove low-correlation features and redundant features. The p-value test is then used to verify the significance level of the correlation between each feature, thus obtaining the final input features.

[0051] The correlation heatmap of some input characteristics and the degree of transient overvoltage change is shown below. Figure 5 As shown, the darker the color of the square, the higher the correlation between the two corresponding features. It can be seen that the selected features all meet the proposed feature selection principle.

[0052] To verify the effectiveness of the proposed feature selection method and the rationality of the selected key input features, a classic BP neural network was used to perform regression prediction of transient overvoltage amplitude. The selected features were used as input, and the transient overvoltage amplitude of the corresponding scenario was used as the output. 80% of the samples were used as the training dataset, and the remaining 20% ​​as the test dataset. Root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) were used as performance evaluation metrics. The formulas for each metric are as follows:

[0053]

[0054]

[0055]

[0056] In the formula: n represents the total number of test samples, j represents the j-th test sample, y' represents the predicted value, and y represents the actual value. Each index reflects the overall error level of overvoltage amplitude prediction; the lower the evaluation index value, the better the model prediction effect.

[0057] Table 2 compares the prediction performance of the BP neural network under different feature extraction states. As can be seen from Table 2, the model's prediction performance is poor with the initial input feature set. However, as the feature extraction process progresses, the model shows improvement across various metrics, demonstrating the effectiveness of the proposed method.

[0058] Table 2. Prediction results under different feature extraction states.

[0059]

[0060] Furthermore, the deviation between the true and predicted values ​​of the BP network under different feature extraction states is as follows: Figure 6 As shown, using the key response features extracted by the method in this paper as input yields good regression prediction results on the test set, with predicted values ​​closer to the true values, further validating the effectiveness of the proposed method.

[0061] Prior to transient overvoltage feature extraction methods, this invention analyzes the mechanism of transient overvoltage at the sending end caused by commutation failure faults in AC systems based on the dynamic characteristics of renewable energy and asynchronous motors during faults.

[0062] by Figure 3 Taking the HVDC system containing wind power as an example, under normal operating conditions, the reactive power balance relationship of the sending-end system can be expressed as:

[0063] ΔQ=Q ac +Q w +Q cr -Q dr =0 (4)

[0064] In the formula: Q cr Q ac and Q w These refer to the reactive power output of the AC filter and reactive power compensation device, the sending-end AC system, and the wind farm, respectively; Q dr This refers to the reactive power consumption of the rectifier station.

[0065] The reactive power consumption of the rectifier station is as follows:

[0066]

[0067] In the formula: P dr E represents the active power of DC transmission. r N1 is the effective value of the no-load line voltage of the rectifier-side transformer, N1 is the number of 6-pulse converters in the rectifier station, α is the emission angle of the rectifier station, and X is the effective value of the no-load line voltage of the rectifier-side transformer. r For the leakage reactance of the rectifier transformer, I d Let be the direct current. From equation (2), it can be seen that the reactive power consumed by the rectifier is related to the emission angle α, and increases with the increase of the emission angle. The reactive power consumption of the rectifier station is generally between 40% and 60% of the DC transmission power. Therefore, higher DC power transmission requires increased reactive power compensation on the rectifier side.

[0068] When a severe short-circuit fault occurs in the receiving-end AC system, causing DC commutation failure, the receiving-end inverter turns on, resulting in a rapid drop in DC voltage and a rapid rise in DC current. The sending-end rectifier station activates its low-voltage limit control unit, increasing the emission angle to suppress the rise in DC current. During this process, the reactive power consumption of the rectifier station increases with the increase in emission angle, leading to a voltage drop near the sending-end AC system.

[0069] Under the regulation of the low-voltage limit control unit, the emission angle continuously increases, exceeding 90°, causing the rectifier to shut down and protecting the equipment. The DC current gradually decreases, possibly even reaching zero, causing a momentary blockage in the DC system. As the DC current decreases, the reactive power consumption of the rectifier station also gradually decreases. When the DC current approaches zero, the reactive power consumption of the rectifier station becomes negligible. During this process, the AC filter and reactive power compensation device of the rectifier station remain operational, and the reactive power consumption of the inverter is far lower than the reactive power generated by the filter. Therefore, a large amount of reactive power is fed back to the sending-end system, leading to an increase in the transient voltage of the sending-end AC system. Furthermore, during the DC current recovery phase, the DC current remains low, and the reactive power consumption of the sending-end rectifier station remains low, which may lead to a certain degree of overvoltage.

[0070] The above analysis shows that after commutation failure, the sending-end system exhibits a transient characteristic of first decreasing and then increasing the sending-end bus voltage, which can lead to transient overvoltage. The changes in DC current, bus voltage, and emission angle during commutation failure are as follows: Figure 2 As shown.

[0071] Secondly, based on the mechanism of transient overvoltage at the sending end caused by commutation failure in the AC system, the relationship between the change in AC bus voltage at the sending end and the reactive power output of the AC filter and reactive power compensation device, the reactive power consumed on the DC rectifier side, the reactive power transmitted by the wind farm, and the short-circuit capacity at the sending end is analyzed to uncover the main influencing factors reflecting the level of transient overvoltage.

[0072] During the commutation failure period, the change in AC bus voltage at the sending end can be expressed as:

[0073]

[0074] In the formula: ΔU t S indicates the degree of change in the sending-end bus voltage. cr This represents the short-circuit capacity of the converter bus on the rectifier side. As the transient overvoltage mechanism indicates, during the transient voltage rise period, ΔQ > 0, and the reactive power consumed by the rectifier station is relatively small, resulting in a large amount of reactive power being fed back into the sending-end system. During the transient process, Q... ac This can be approximated as a steady-state value. Therefore, the degree of change in the sending-end bus voltage ΔUt is mainly affected by the reactive power Q output by the AC filter and reactive power compensation device. cr The reactive power consumption Q on the DC rectifier side dr Reactive power Q transmitted by wind farm w and the short-circuit capacity S at the sending end cr Influence.

[0075] The degree of change in bus voltage ΔU t With Q cr Q dr Q w ,S cr Analyze the relationships between them:

[0076] (1) The impact of AC filters and reactive power compensation devices

[0077] AC filters and reactive power compensation devices are typically capacitors. Considering their inherent reactive-voltage dynamic characteristics, Q cr It can be represented as follows:

[0078]

[0079] In the formula: Q crN The reactive power compensation capacity of the AC filter and reactive power compensation device under normal operating conditions depends mainly on the transmission power of the DC system under normal operating conditions. The reactive power compensation increases with the increase of transmission power, which may lead to more severe overvoltage levels.

[0080] (2) The impact of DC systems

[0081] According to the expression for reactive power consumed by the rectifier station in the transient overvoltage mechanism, when DC commutation fails, the reactive power consumed on the rectifier side mainly depends on the DC current I. d emission angle α and DC transmission power P dr The reactive power absorbed by the rectifier during commutation failure varies with changes in DC voltage and current. When the DC current and voltage drop to their limits, the rectifier will absorb minimal reactive power. At this point, the reactive power generated by the DC system has the most severe impact on the transient voltage of the transmitting AC system, leading to a transient overvoltage peak at the transmitting bus.

[0082] (3) Impact of wind power generation systems

[0083] Following a commutation fault, near-field wind farms may experience low-voltage ride-through protection during voltage reduction. During low-voltage ride-through, wind turbines employ a "reactive power priority" control mode to provide the system with as much reactive power as possible while avoiding a significant reduction in active power output, as well as excessive generator speed and disconnection from the power system due to a significant reduction in active power output. The changes in active and reactive power of wind turbines during low-voltage ride-through can be expressed as follows:

[0084] ΔQ w =[k q (0.9-U w )-1]Q w0 (8)

[0085] ΔP w =(k p -1)P w0 (9)

[0086] In the formula: k q and k p These represent the reactive and active power output coefficients of the fan during low-voltage breakdown, respectively; P w0 and Q w0 U represents the rated active power and reactive power of the wind turbine, respectively. w This refers to the turbine terminal voltage. During low voltage ride-through, the active power of the turbine decreases significantly (ΔP). w >0), reactive power output increases (ΔQ) w >0), which leads to a reactive power surplus on the wind farm side after the fault. This surplus reactive power may flow back to the sending-end system and work together with the surplus reactive power of the converter station, thereby causing the transient overvoltage level at the sending end to deteriorate further. It can be seen that the transient overvoltage level is related to various response characteristics of the wind power system.

[0087] (4) Impact of short-circuit capacity at the sending end

[0088] The strength of a sending-end system is typically measured by the ratio of the short-circuit capacity of the sending-end converter bus to its rated DC power, known as the short-circuit ratio. Based on the definition of the short-circuit ratio, its calculation formula can be expressed as:

[0089]

[0090] Generally, the higher the system strength, the lower the transient overvoltage level. For a DC system with a fixed rated transmission power, the short-circuit capacity of the sending-end bus directly affects the system's transient overvoltage level. Under the same conditions, conventional units generally have a larger short-circuit capacity and stronger support capability, while renewable energy units have a relatively smaller short-circuit capacity and weaker support capability. Therefore, the proportion of conventional units and renewable energy in the DC near-field has a significant impact on the short-circuit capacity. After large-scale renewable energy integration, the short-circuit capacity of the sending-end system decreases, which will lead to more severe transient overvoltages during faults.

[0091] Based on the above analysis, the main influencing factors that reflect the transient overvoltage level are: the ratio of new energy and conventional units connected to the sending end, the transmission power of the DC system, and the output characteristics of DC, new energy and other systems during the fault.

[0092] In this embodiment, the transient overvoltage feature extraction uses Spearman correlation coefficient to extract key response features that reflect the grid voltage safety and stability level, thereby improving the calculation accuracy and speed of the response-driven algorithm.

[0093] Data-driven methods typically require constructing a sample set, with the key being the selection of appropriate input features. These input features need to fully reflect the system's operational state and be obtainable online. Furthermore, the dimensionality of the input features should not be too high. In cases of topological changes, too many features can lead to the curse of dimensionality, resulting in excessively long online update times. Therefore, it is necessary to determine suitable model input features offline. Considering the ease of data acquisition, the reasonableness of data dimensionality, and the ability of the data to reflect the system state, this invention proposes a feature selection method based on the Spearman correlation coefficient.

[0094] The Spearman correlation coefficient measures the correlation between two variables. It is calculated based on the order of the original data to reduce reliance on linear relationships. The Spearman correlation coefficient can be used to assess correlation for discrete data or data that does not conform to a normal distribution. Assume there are n sets of data, including (x1, y1), (x2, y2), ..., (x... i y i ), ..., (x n y n When the Spearman correlation coefficient is 0, the formula for calculating the correlation coefficient is as follows:

[0095]

[0096] In the formula: X i and Y i These represent the levels corresponding to the i-th data group, and Let x and y be the average ranks of variables x and y, respectively. From equation (8), we know that the Spearman correlation coefficient ranges from [-1, 1]. When R... s When R approaches 0, it indicates that variables x and y are uncorrelated; when R... s A p-value close to 1 or -1 indicates a strong correlation between variables x and y. After obtaining the correlation analysis values, hypothesis testing (p-value test) is needed to verify the strength and significance of the correlation between the features. If the p-value is greater than 0.05, there is no significant difference, i.e., no correlation. If the p-value is less than 0.05, it indicates a significant difference between the features, and a strong correlation.

[0097] In summary, this invention elucidates the mechanism of transient overvoltage generation in the sending-end system under commutation failure faults based on the dynamic characteristics of renewable energy and asynchronous motors during faults, providing theoretical support for clarifying the analysis of transient voltage instability under large-scale grid connection of new energy.

[0098] This invention combines the mechanism of transient overvoltage at the sending end caused by commutation failure in AC systems, and based on the transient characteristics after commutation failure, it identifies the main influencing factors reflecting the level of transient overvoltage, providing a basis for the selection of input features for feature extraction methods.

[0099] This invention proposes a feature extraction method based on Spearman correlation coefficient, which combines the main influencing factors reflecting transient overvoltage levels. This method extracts key response features that reflect the voltage safety and stability level of the power grid, improves the calculation accuracy and speed of the response-driven algorithm, helps reduce redundancy among high-dimensional electrical characteristic input features, improves the performance and prediction efficiency of the evaluation model, and is more adaptable to the complex characteristics of large power grids.

[0100] Example 2

[0101] according to Figure 7 As shown, the present invention also provides a transient overvoltage feature extraction system, comprising:

[0102] Dataset construction module 1 is used to obtain input feature quantities and construct an initial input feature set based on the input feature quantities;

[0103] Elimination module 2 is used to eliminate input features in the initial input feature set to obtain input features with low redundancy.

[0104] The verification module 3 is used to perform correlation feature verification on input feature quantities with low redundancy to obtain the final input feature quantity and complete the transient overvoltage feature extraction.

[0105] Example 3

[0106] The present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, such as a transient overvoltage feature extraction program.

[0107] When the processor executes the computer program, it implements the steps of the transient overvoltage feature extraction method described above, for example:

[0108] Obtain the input feature values, and construct an initial input feature set based on the input feature values;

[0109] The input features in the initial input feature set are removed to obtain input features with low redundancy.

[0110] The input features with low redundancy are subjected to correlation feature testing to obtain the final input features, thus completing the transient overvoltage feature extraction.

[0111] Alternatively, when the processor executes the computer program, it implements the functions of each module in the above system, for example:

[0112] Dataset construction module 1 is used to obtain input feature quantities and construct an initial input feature set based on the input feature quantities;

[0113] Elimination module 2 is used to eliminate input features in the initial input feature set to obtain input features with low redundancy.

[0114] The verification module 3 is used to perform correlation feature verification on input feature quantities with low redundancy to obtain the final input feature quantity and complete the transient overvoltage feature extraction.

[0115] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the mobile terminal.

[0116] For example, the computer program can be divided into a dataset construction module 1, a elimination module 3, and a testing module 4; the specific functions of each module are as follows:

[0117] Dataset construction module 1 is used to obtain input feature quantities and construct an initial input feature set based on the input feature quantities;

[0118] Elimination module 2 is used to eliminate input features in the initial input feature set to obtain input features with low redundancy.

[0119] The verification module 3 is used to perform correlation feature verification on input feature quantities with low redundancy to obtain the final input feature quantity and complete the transient overvoltage feature extraction.

[0120] The mobile terminal can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The mobile terminal may include, but is not limited to, a processor and memory.

[0121] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the mobile terminal, connecting various parts of the mobile terminal via various interfaces and lines.

[0122] The memory can be used to store the computer program and / or module. The processor implements various functions of the mobile terminal by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0123] The memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a given function (such as sound playback or image playback). The data storage area may store data created based on the use of the phone (such as audio data or a phonebook). Furthermore, the memory may include high-speed random access memory (RAM) and non-volatile memory, such as hard disks, RAM, plug-in hard disks, SmartMediaCards (SMC), Secure Digital (SD) cards, flash cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0124] Example 4

[0125] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the transient overvoltage feature extraction method.

[0126] If the modules / units integrated in the mobile terminal are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0127] Based on this understanding, all or part of the processes in the above-described method can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of the transient overvoltage feature extraction method described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form.

[0128] The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0129] It should be noted that the content contained in the computer-readable medium may be appropriately added to or subtracted from the content as required by the legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium may not include electrical carrier signals and telecommunication signals.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for extracting transient overvoltage features, characterized in that, include: Obtain the input feature values, and construct an initial input feature set based on the input feature values; Among them, the input characteristic quantities are obtained by combining the main influencing factors of grid voltage instability or transient overvoltage level. The input characteristic quantities include DC-related response characteristic quantities before and after the fault and power output characteristic related response characteristic quantities of new energy system. The main influencing factors include the ratio of new energy and conventional units connected to the sending end, the transmission power of the DC system, and the output characteristics of the DC and new energy systems during faults. The input features in the initial input feature set are removed to obtain input features with low redundancy. The process of removing input features from the initial input feature set to obtain input features with low redundancy is as follows: S1, calculate several correlation coefficients between the input feature quantities in the initial input feature set and the degree of voltage change, and make a first condition judgment on the correlation coefficients. Perform a first stage elimination on the input feature quantities corresponding to the correlation coefficients that do not meet the first condition judgment to obtain the input feature quantities that are highly correlated with the degree of voltage change. S2, calculate several correlation coefficients for each pair of input feature quantities with high correlation to voltage change, and make a second condition judgment on several correlation coefficients. For any one of the two input feature quantities corresponding to the correlation coefficient that does not meet the second condition judgment, perform a second stage elimination to obtain input feature quantities with low redundancy. The specific process for determining the first condition for several correlation coefficients is as follows: A first threshold is set. When the absolute values ​​of several correlation coefficients between the input feature quantities in the initial input feature set and the voltage change degree are less than the first threshold, the input feature quantities corresponding to the correlation coefficients are removed to obtain the input feature quantities that are highly correlated with the voltage change degree. In S2, the specific process of making a second conditional determination on several correlation coefficients is as follows: A second threshold is set. When the absolute value of several correlation coefficients obtained by pairwise calculation of the remaining input features is greater than the second threshold, either of the two input features corresponding to the correlation coefficient is removed to obtain input features with low redundancy. The correlation feature test is performed on the input feature quantities with low redundancy to obtain the final input feature quantities, thus completing the transient overvoltage feature extraction. Among them, the p-value test method is used to perform correlation feature testing on input features with low redundancy. A threshold is set, and input features with p values ​​less than the threshold are taken as the final input features.

2. The transient overvoltage feature extraction method according to claim 1, characterized in that, The correlation coefficient was calculated using the Spearman correlation coefficient method. The formula for calculating the Spearman correlation coefficient is as follows: In the formula: and They represent the first i The level corresponding to the group data and These are the average ranks of variables x and y, respectively.

3. A transient overvoltage feature extraction system, characterized in that, A method for extracting transient overvoltage features according to any one of claims 1-2 includes: Dataset construction module (1) is used to obtain input feature quantities and construct an initial input feature set based on the input feature quantities; The elimination module (2) is used to eliminate the input features in the initial input feature set to obtain input features with low redundancy. The verification module (3) is used to perform correlation feature verification on the input feature quantities with low redundancy to obtain the final input feature quantity and complete the transient overvoltage feature extraction.

4. A mobile terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the transient overvoltage feature extraction method as described in any one of claims 1-2.

5. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the transient overvoltage feature extraction method as described in any one of claims 1-2.

Citation Information

Patent Citations

  • Active-voltage mutual correlation degree island detection method based on adaptive threshold selection

    CN109031045A

  • Multi-infeed direct current mutual influence evaluation method and system considering direct current change domain

    CN112364497A