A novel wavelet feature analysis method for transient voltage stability in power systems
By constructing a conservative stability domain using Daubechies wavelet transform and the minimum volume closed ellipse method, the problem of low false negative rate in transient voltage stability analysis of power systems is solved, achieving efficient transient voltage stability discrimination, which is applicable to the safety assessment and control of power systems.
Patent Information
- Application Number
- CN202611259970.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-19
- Publication Date
- 2026-11-17
AI Technical Summary
Existing methods for analyzing transient voltage stability in power systems still have room for improvement in meeting the requirement of low failure rate, especially under instability mechanisms such as multiple balance point switching and fast-slow dynamic decoupling. How to construct a discrimination strategy that combines low failure rate and interpretability is an urgent problem to be solved.
The node voltage time series data is converted into two-dimensional feature vectors using Daubechies wavelet transform. A conservative stable region is constructed using the minimum volume closed ellipse method. The discrimination rule of 'stable within the region and unstable outside the region' is adopted. A compact envelope region is constructed using known stable samples to achieve a low false negative rate.
It achieves transient voltage stability discrimination with low omission rate in full electromagnetic simulation data, provides an efficient discrimination scheme, is suitable for rapid screening of massive anticipated fault scenarios, and meets the safety requirements of engineering applications.
Smart Images

Figure CN123121664A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power system safety and stability analysis, assessment and control technology, and more specifically, to a novel wavelet feature analysis method for transient voltage stability of power systems. Background Technology
[0002] With a high proportion of new energy sources being connected to the grid via power electronic converters, transient voltage stability analysis of power systems faces new challenges. Instability mechanisms such as multi-balance point switching and fast / slow dynamic decoupling have received considerable attention in recent years. Against this backdrop, rapid screening of a large number of anticipated faults has become a crucial aspect of safe grid operation. Existing detection methods still have room for improvement in meeting the requirement of low false negative rates.
[0003] The effectiveness of stability discrimination largely depends on the quality of the system dynamics representation. Some studies have attempted to construct a feature space with clear mechanisms, such as using wavelet transform to extract physically meaningful features. These features include F1, which characterizes the energy of high-frequency transient impulses, and F2, which characterizes the energy structure relationship across time scales. Through these features, transient voltage dynamics can be mapped to a low-dimensional space. After obtaining the feature representation, how to design a matching discrimination strategy that achieves both low false negative rate and interpretability is a problem that requires further exploration.
[0004] In summary, in transient voltage stability discrimination, it is worthwhile to explore how to construct a discrimination mechanism that meets the requirements of low false negative rate while maintaining the interpretability of the decision-making process based on feature characterization techniques. Summary of the Invention
[0005] To address the shortcomings of existing technologies, this invention provides a novel wavelet feature analysis method for transient voltage stability in power systems.
[0006] According to one aspect of the present invention, a novel wavelet feature analysis method for transient voltage stability of a power system is provided, comprising: Acquire node voltage timing data of the target power grid after a preset fault occurs under a full electromagnetic transient simulation environment; The Daubechies wavelet pairs are used to convert the node voltage time series data into two-dimensional feature vectors with physical meaning; Select a sample set under known stable conditions from the two-dimensional feature vectors, and construct a conservative stability region based on the sample set using the minimum volume closed ellipse method; The feature vector of the sample to be judged is compared with the conservative stable region to determine whether the preset conditions are met. If the preset conditions are met, the feature vector of the sample to be judged is determined to be transient voltage stable; otherwise, it is determined to be transient voltage unstable.
[0007] Optionally, Daubechies wavelet pairs are used to convert the nodal voltage time-series data into physically meaningful two-dimensional feature vectors, including: Daubechies wavelet was used to perform multi-level discrete wavelet transform on the node voltage time series data to obtain several levels of approximation coefficients and detail coefficients. Based on the mapping relationship between the subband frequency corresponding to each level of detail coefficients and the physical dynamics of the power grid, several levels of approximation coefficients and detail coefficients are divided into at least three coefficient groups with different physical meanings. Based on the pre-constructed calculation expressions of the first and second characteristic quantities, a standard two-dimensional orthogonal characteristic space independent of specific operating conditions is constructed. The first characteristic quantity represents the weighted comprehensive level of the switching process dynamic energy and the control dynamic energy, and the second characteristic quantity represents the modulation intensity of the interaction between the switching process energy and the control dynamic and electromechanical coupling dynamic. The coordinates of the first and second eigenvalues of the coefficient set are calculated according to the calculation expression, and all coordinates are mapped to the standard two-dimensional orthogonal feature space to form a two-dimensional feature vector representing the node state.
[0008] Optionally, the coefficient set includes: First coefficient group G sw Corresponding to the dynamics of the power electronic switching process, the sub-band frequency range of the detailed coefficient layer contained in the first coefficient group needs to cover the integer or half-integer multiples of the switching frequency of the power electronic converter. Second coefficient group G ctrl Corresponding to the converter control dynamics, the subband frequency range of the detailed coefficient layer contained in the second coefficient group mainly reflects the response frequency band of the current loop and voltage loop controllers. Third coefficient group G em Corresponding to the electromechanical coupling dynamics of the power grid, the third coefficient group includes a detailed coefficient layer or a lowest frequency approximation coefficient layer, whose sub-band frequency range mainly reflects the synchronous unit rotor motion and low-frequency oscillation electromechanical transient processes.
[0009] Optionally, the expression for calculating the first feature is: In the formula, E (·) is a function for calculating the total energy of the coefficient group; ω sw and ω ctrl The preset weighting coefficients are used, and they satisfy the following conditions: ω sw + ω ctrl = 1; The calculation expression of the second feature quantity is: wherein, E ( G ctrl ) / E ( G em ) reflects the ratio of the control dynamic energy to the electromechanical coupling dynamic energy; S (·) is a nonlinear modulation function, configured to characterize the modulation effect of the switching process energy on the ratio relationship, and is specifically S(x)=log(1+x) or S(x)=sqrt(x).
[0010] Optionally, the minimum volume enclosing ellipse is determined by solving the following optimization problem: minimize log det(Σ) subject to (xi - μ)ᵀΣ -1 (xi - μ) ≤ 1, i = 1, 2, ..., N wherein, xi ∈ 2 is the two-dimensional feature vector of the i-th stable sample, μ ∈ 2 is the ellipse center, Σ ∈ 2x2 is a 2×2 positive definite matrix (covariance matrix), log det(Σ) represents the natural logarithm of the determinant of the matrix Σ, Σ is the positive definite matrix for constructing the conservative stability region, and N is the number of samples.
[0011] Optionally, the preset condition is: wherein, d is the Mahalanobis distance from the feature vector of the to-be-determined sample to the boundary of the conservative stability region; x is the two-dimensional feature vector of the to-be-determined sample, μ is the ellipse center, and Σ is the positive definite matrix for constructing the conservative stability region; when the to-be-determined sample satisfies the preset condition, it is determined to be transient voltage stable, otherwise it is determined to be transient voltage unstable.
[0012] Optionally, the method further comprises: setting a buffer area outside the boundary of the conservative stability region, defining two thresholds d1 and d2, where d1=1, d2>1, when the Mahalanobis distance d of the feature vector of the to-be-determined sample satisfies 1<d ≤ d2, the sample falls into the buffer area, triggering manual review or outputting a determination conclusion with confidence.
[0013] Optionally, it also includes: when a new stable operating condition sample is added, the new sample is incorporated into the original known stable sample set, the optimization problem is solved again, the elliptical boundary is updated, and the conservative stable domain is adaptively updated.
[0014] According to another aspect of the present invention, a novel wavelet feature analysis device for transient voltage stability of a power system is provided, comprising: The acquisition module is used to acquire node voltage timing data generated by the target power grid after a preset fault occurs under a full electromagnetic transient simulation environment. The conversion module is used to convert nodal voltage time-series data into physically meaningful two-dimensional feature vectors using Daubechies wavelet pairs. The module is used to select a sample set under known stable conditions from the two-dimensional feature vectors, and to construct a conservative stability region based on the sample set using the minimum volume closed ellipse method. The judgment module is used to compare the feature vector of the sample to be judged with the conservative stable region to determine whether the preset conditions are met. The determination module is used to determine whether the feature vector of the sample to be judged is transiently stable if the preset conditions are met, otherwise it is determined to be transiently unstable.
[0015] According to another aspect of the present invention, a computer-readable storage medium is provided, the storage medium storing a computer program for performing the methods described in any of the above aspects of the present invention.
[0016] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the method described in any of the preceding aspects of the present invention.
[0017] Therefore, this invention uses the node voltage time series data output by full electromagnetic simulation as a basis, extracts physically meaningful feature quantities through discrete wavelet transform, constructs a compact envelope region as a conservative stability domain in a two-dimensional feature space using known stable samples, and adopts the discrimination rule of "determining stability within the domain and instability outside the domain", thus forming a complete technical chain from data to discrimination. Attached Figure Description
[0018] Exemplary embodiments of the present invention can be more fully understood by referring to the following figures: Figure 1 This is a flowchart illustrating a novel wavelet feature analysis method for transient voltage stability in a power system provided by an exemplary embodiment of the present invention. Figure 2This is another flowchart illustrating a novel wavelet feature analysis method for transient voltage stability of a power system provided by an exemplary embodiment of the present invention. Figure 3 This is a schematic diagram of wavelet coefficient structured grouping based on physical frequency band correlation provided by an exemplary embodiment of the present invention; Figure 4 This is a schematic diagram of the wavelet feature space and conservative stability domain provided by an exemplary embodiment of the present invention; Figure 5 This is a schematic diagram of stability margin and buffer region provided by an exemplary embodiment of the present invention; Figure 6 This is a schematic diagram of the structure of a novel power system transient voltage stability wavelet feature analysis device provided in an exemplary embodiment of the present invention; Figure 7 This is the structure of an electronic device provided in an exemplary embodiment of the present invention. Detailed Implementation
[0019] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.
[0020] It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values of the components and steps described in these embodiments do not limit the scope of the invention.
[0021] Those skilled in the art will understand that the terms "first," "second," etc., in the embodiments of the present invention are only used to distinguish different steps, devices, or modules, and do not represent any specific technical meaning, nor do they indicate a necessary logical order between them.
[0022] It should also be understood that in the embodiments of the present invention, "multiple" can refer to two or more, and "at least one" can refer to one, two or more.
[0023] It should also be understood that any component, data or structure mentioned in the embodiments of the present invention can generally be understood as one or more unless explicitly defined or given contrary instructions in the context.
[0024] Furthermore, the term "and / or" in this invention is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this invention generally indicates that the preceding and following related objects have an "or" relationship.
[0025] It should also be understood that the description of the various embodiments in this invention emphasizes the differences between the various embodiments, and the similarities or similarities can be referred to each other. For the sake of brevity, they will not be described in detail.
[0026] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.
[0027] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.
[0028] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, they should be considered part of the specification.
[0029] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.
[0030] The embodiments of this invention can be applied to electronic devices such as terminal devices, computer systems, and servers, and can operate together with a wide range of other general-purpose or special-purpose computing system environments or configurations. Well-known examples of terminal devices, computing systems, environments, and / or configurations suitable for use with electronic devices such as terminal devices, computer systems, and servers include, but are not limited to: personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputer systems, mainframe computer systems, and distributed cloud computing environments including any of the above systems, etc.
[0031] Electronic devices such as terminal devices, computer systems, and servers can be described in the general context of computer system executable instructions (such as program modules) executed by a computer system. Typically, program modules can include routines, programs, object programs, components, logic, data structures, etc., which perform specific tasks or implement specific abstract data types. Computer systems / servers can be implemented in distributed cloud computing environments, where tasks are executed by remote processing devices linked through communication networks. In distributed cloud computing environments, program modules can reside on local or remote computing system storage media, including storage devices.
[0032] Exemplary methods Figure 1 This is a flowchart illustrating a novel wavelet feature analysis method for transient voltage stability in a power system provided by an exemplary embodiment of the present invention. This embodiment can be applied to electronic devices, such as... Figure 1 As shown, the new wavelet feature analysis method 100 for transient voltage stability of power systems includes the following steps: Step 101: Obtain the node voltage timing data generated after a preset fault occurs in the target power grid under a full electromagnetic transient simulation environment; Step 102: Use Daubechies wavelet pairs to convert the node voltage time series data into two-dimensional feature vectors with physical meaning; Step 103: Select a sample set under known stable conditions from the two-dimensional feature vector, and construct a conservative stability region based on the sample set using the minimum volume closed ellipse method. Step 104: Compare the feature vector of the sample to be judged with the conservative stable region to determine whether the preset conditions are met; Step 105: If the preset conditions are met, determine that the feature vector of the sample to be judged is transient voltage stable; otherwise, determine that it is transient voltage unstable.
[0033] Specifically, this invention provides a wavelet feature-based zero-miss detection method for transient voltage stability in full electromagnetic simulation based on a conservative stability domain. This method utilizes a constructed two-dimensional orthogonal wavelet feature space (F1, F2) and employs geometric methods such as single-class support vector machines or minimum volume closed ellipsoids to construct a conservative stability domain that encloses known stable samples. It then adopts a discrimination rule of "stable within the domain, unstable outside the domain," providing an optional solution for online screening and assessment of power grid stability.
[0034] The purpose of this invention is to provide a method and system for zero-miss detection of transient voltage stability wavelet features based on conservative stability domain in all-electromagnetic simulation. This method aims to solve the problem that existing stability discrimination methods cannot achieve a zero miss rate in engineering applications. Specific objectives include: A conservative stability domain discrimination mechanism based on geometric envelope is constructed. Wavelet transform is used to extract physically meaningful features, dynamically mapping transient voltages to a two-dimensional feature space. Based on this, using known stable samples, a compact envelope region is constructed as the conservative stability domain using geometric methods such as single-class support vector machines or minimum-volume closed ellipses. The discrimination rule adopts the logic of determining stability within the domain and instability outside the domain, ensuring that the requirement of a low false negative rate is reflected from the algorithm construction level.
[0035] An efficient discrimination scheme suitable for the characteristics of full electromagnetic simulation data is developed. This method has a simple structure, manageable computational burden, and can adapt to the high-dimensional time-series data characteristics of full electromagnetic simulation output. By constructing a stable sample set offline and then discriminating online, it can perform batch screening of massive amounts of anticipated fault scenarios, meeting the practical needs of engineering applications that balance discrimination efficiency and safety requirements.
[0036] This invention is based on the node voltage time series data output by full electromagnetic simulation. It extracts physically meaningful feature quantities through discrete wavelet transform, constructs a compact envelope region as a conservative stability domain in a two-dimensional feature space using known stable samples, and adopts the discrimination rule of "determining stability within the domain and instability outside the domain", thus forming a complete technical chain from data to discrimination.
[0037] refer to Figure 2 As shown, the method includes the following core steps: Step S1: Obtain the timing data of the node voltages in the full electromagnetic simulation. This step provides the raw input for the discrimination. Specifically, it involves acquiring the node voltage timing data generated after a preset fault occurs in a full electromagnetic transient simulation environment of the target power grid. Data sources include offline simulation software or online safety assessment programs from dispatch centers. The data sampling rate is typically no less than several kilohertz, covering the switching processes of power electronic equipment and the response of the control system. This step clarifies that the data object processed by the method is a known and complete simulation waveform.
[0038] Step S2: Wavelet Feature Extraction This step converts the raw voltage timing data into a two-dimensional feature vector with physical meaning.
[0039] The voltage time-series data is decomposed using discrete wavelet transform. One of the Daubechies, Symlets, or Coiflets wavelet families is selected. The number of decomposition levels is determined based on the sampling frequency and the switching frequency of the power electronic device of interest, typically satisfying the following: the upper limit of the subband frequency corresponding to the highest frequency detail coefficient is higher than the highest switching harmonic frequency of the power electronic device of interest, and the lower limit of the subband frequency corresponding to the lowest frequency detail coefficient is lower than the lower limit of the possible subsynchronous oscillation frequency.
[0040] like Figure 3 As shown, based on the mapping relationship between the sub-band frequencies corresponding to the detailed coefficients of each layer and the physical dynamics of the power grid, the decomposed coefficients are divided into three groups: First coefficient group G sw This corresponds to the dynamics of the power electronic switching process. The detailed coefficient layers contained in this group must cover the integer or half-integer multiples of the switching frequency of the power electronic converter (e.g., obtaining the nominal switching frequency based on the converter model). f sw Calculate k· f sw and (k+0.5)· f sw where k is a positive integer. G swThe corresponding sub-band frequency range must include at least one of the aforementioned frequency bands.
[0041] Second coefficient group G ctrl This corresponds to the converter control dynamics. The detailed coefficient layer contained in this group has a subband frequency range that mainly reflects the response frequency band of controllers such as the current loop and voltage loop (typically tens to hundreds of Hz).
[0042] Third coefficient group G em This corresponds to the electromechanical coupling dynamics of the power grid. The detailed coefficient layer or the lowest frequency approximation coefficient layer contained in this group mainly reflects the electromechanical transient processes (usually sub-Hz to several Hz) such as synchronous generator rotor motion and low-frequency oscillations in its sub-band frequency range.
[0043] Construct the first feature quantity F1 and the second feature quantity F2: Where E(·) is a function for calculating the total energy of the coefficient group, and can be either L1 norm or L2 norm; ω sw and ω ctrl The preset weighting coefficients are used, and they satisfy the following conditions: ω sw + ω ctrl = 1; The weighting coefficient is set according to the dominant equipment type in the power grid area: for new energy power plants that mainly use voltage source converters, it can be taken as 1. ω ctrl > ω sw For power grid areas containing a large number of thyristor-type devices, the following approach can be adopted: ω sw > ω ctrl This characteristic quantity reflects the weighted combined energy of the switching process and control dynamics.
[0044] in, E ( G ctrl ) / E ( G em This reflects the ratio of control dynamic energy to electromechanical coupling dynamic energy. S (·) is a nonlinear modulation function used to characterize the modulation effect of the switching process energy on this ratio relationship. Specifically, it can be taken as S(x) = log(1+x) or S(x) = sqrt(x). When E( G sw When ) increases, S(E( Gsw As the value of F2 increases, the dynamic interaction between fast and slow states intensifies. This characteristic reflects the modulation strength of the switching process energy on the control-electromechanical interaction relationship.
[0045] The voltage data of each node within the post-fault time window is converted into a two-dimensional feature vector (F1, F2). For steady state, both F1 and F2 are non-negative and are typically located in the region of the first quadrant near the origin.
[0046] Additional examples: To illustrate the feature extraction process, we will now use an example of using Daubechies 4th order wavelet (db4) to perform a 6-level decomposition of the bus voltage time series signal output from the full electromagnetic simulation of a new energy power station.
[0047] Given a sampling frequency fs = 10 kHz and an original signal duration of 2 seconds, including the period before and after the fault, a 6-level discrete wavelet transform is performed on the voltage signal V(t) to obtain the coefficient sequence: {cA6, cD6, cD5, cD4, cD3, cD2, cD1}. According to wavelet theory, the approximate frequency ranges corresponding to the coefficients at each level are as follows: cA6 (6th layer approximation coefficient): frequency approximately 0–78 Hz cD6 (6th layer detail factor): Frequency approximately 78–156 Hz cD5 (5th layer detail factor): Frequency approximately 156–312 Hz cD4 (4th layer detail factor): frequency approximately 312–625 Hz cD3 (third layer detail factor): frequency approximately 625–1250 Hz cD2 (second layer of detail): frequency approximately 1250–2500 Hz cD1 (Level 1 detail factor): Frequency approximately 2500–5000 Hz Grouping based on physical frequency band correlation: G sw (Switching process dynamics): Select cD1 and cD2, whose frequency range covers integer multiples of the frequency band of typical power electronic switches.
[0048] G ctrl (Converter control dynamics): Select cD3 and cD4, whose frequency range covers the response bands of the current loop and voltage loop.
[0049] G em (Electromechanical coupling dynamics): cD5 and cA6 are selected, and their frequency range covers electromechanical transient processes.
[0050] The energy function E(·) adopts the L2 norm (i.e., the sum of squares) for the coefficient sequence c. X Its energy EX = Σ (cX[i])^2.
[0051] Assuming the station primarily uses voltage source converters, take... ω ctrl = 0.7, ω sw = 0.3, the nonlinear modulation function is S(x) = log(1+x).
[0052] For the voltage data at a certain moment after the fault, the following can be calculated: E( G sw ) = 0.52 E( G ctrl ) = 0.38 E( G em = 2.15 but: F1 = 0.3×0.52 + 0.7×0.38 = 0.156 + 0.266 = 0.422 F2 = (0.38 / 2.15) × log(1+0.52) = 0.1768 × 0.419 = 0.074 The feature vector of this node is (F1, F2) = (0.422, 0.074).
[0053] Step S3: Construct a conservative stability region, such as Figure 4 As shown The core of this step is to transform the concept of zero missed cases into a conservative discrimination region in the feature space. This region is defined as a continuous region that tightly encloses all known stable samples, and its construction process follows the principle of prioritizing safety.
[0054] Select a sample set under known steady-state conditions, denoted as { ( F 1 i , F 2 i The sample set should cover the steady state under various operating modes and fault types, including normal operating conditions, recovery conditions after minor disturbances, and critical stable conditions.
[0055] The goal of constructing a conservative stability region is to find a region with the smallest area that contains all known stable samples. This region is usually elliptical in two-dimensional feature space because its shape is adjustable, its boundaries are smooth, and its mathematical expression is concise.
[0056] A conservative stability region is constructed using the minimum volume closed ellipse method. The ellipse is determined by solving the following optimization problem: Min log det(Σ) st(x i - μ ) T Σ -1 (x i - μ )≤1, i in, , Center of the ellipse The ellipse is a 2×2 positive definite matrix, and its shape and orientation determine the geometric characteristics of the ellipse. The optimal solution to this optimization problem yields an ellipse with the smallest area, ensuring that all stable sample points lie inside the ellipse or on its boundary. This elliptical region is the conservative stability region.
[0057] Ellipse center Reflects the overall distribution and shape matrix of the stable sample set. This determines the extent to which the ellipse extends in each direction. When the stable sample distribution is relatively concentrated, the area of the ellipse is smaller; when the stable sample distribution is relatively dispersed, the area of the ellipse increases accordingly, but still maintains the characteristics of a tight envelope.
[0058] Additional examples: To illustrate the ellipse construction process, assume the stable sample set contains feature points under three typical conditions: (0.05, 0.06), (0.07, 0.05), and (0.06, 0.08). By solving the optimization problem described above, the center of the ellipse is obtained. and shape matrix Related analytical expressions: .
[0059] The equation of the ellipse is: This ellipse encloses all three stable samples and has the smallest area. If new stable samples are added later, the optimization problem can be solved again to obtain the updated elliptical boundary.
[0060] Step S4: Zero-leakage detection This step compares the feature vector of the sample to be judged with the conservative stable region and outputs the judgment conclusion.
[0061] For the node to be judged, its feature vector (F1, F2) is input into the conservative stability region. If the point satisfies the elliptic inequality: it is determined as transient voltage stable; if the point is located outside the ellipse, that is, the above inequality is not satisfied, it is determined as transient voltage instability.
[0062] Since the conservative stability region strictly envelopes all known stable samples, and unstable samples are usually distributed outside the range of stable samples in the feature space, this discrimination rule makes the probability of stable samples being misjudged as unstable low from the construction principle, and the probability of unstable samples being judged as unstable is high at the same time.
[0063] Step S5: Output of discrimination conclusion and optional extension This step outputs the discrimination conclusion, and can be attached with auxiliary information.
[0064] The basic outputs include: Discrimination conclusion: stable or unstable; Feature vector coordinates: (F1, F2).
[0065] As an optional step, the method may further include the following extended functions: Stability margin calculation: Calculate the Mahalanobis distance from the feature vector of the sample to be discriminated to the ellipse boundary, and output it as the stability margin index. A larger distance value indicates that the sample is farther away from the boundary of the stability region, with a higher degree of stability or a deeper degree of instability. The specific calculation formula is: When d ≤ 1, the sample is located within the stability region, and a smaller d indicates that it is closer to the center of the ellipse; when d>1, the sample is located outside the stability region, and a larger d indicates a greater degree of deviation.
[0066] As Figure 5 described, buffer zone setting: set a buffer zone outside the ellipse boundary, that is, define two thresholds d1 and d2 (d1=1, d2>1). When the Mahalanobis distance d of the sample satisfies 1<d ≤ d2, the sample falls into the buffer zone, which can trigger manual review or output a discrimination conclusion with confidence for further analysis by operating personnel.
[0067] Adaptive update of the conservative stability region: when new stable operating condition samples are added, re-execute step S3, add the new samples to the training set, and update the ellipse boundary. The update can adopt an incremental method, that is, adjust on the basis of the original ellipse, or recalculate periodically.
[0068] In a specific embodiment of the present invention, for the full electromagnetic simulation data of a new energy station, the sampling frequency is 10kHz, Daubechies 4th-order wavelet (db4) is selected, and the decomposition level is 6. The grouping strategy is: G sw ={cD1,cD2}, G ctrl ={cD3,cD4},G em ={cD5,cA6}. The energy function uses the sum of squares of the L2 norm, and the weighting coefficients are set according to the equipment type. ω sw =0.3、 ω ctrl =0.7, and the nonlinear modulation function is S(x)=log(1+x).
[0069] A minimum-volume closed ellipse was constructed using 200 stable operating condition samples. After solving the optimization problem, the ellipse center μ = (0.08, 0.09) and the shape matrix Σ = [[0.0025, 0.0008], [0.0008, 0.0032]]. This elliptical region encloses all 200 stable samples, and its major axis makes an angle of approximately 15 degrees with the F2 axis, reflecting a slightly larger degree of dispersion of the stable samples in the F2 direction.
[0070] 100 unstable samples were used for discrimination testing. The feature vectors of all unstable samples were located outside the ellipse, with a minimum Mahalanobis distance of 1.12 and a maximum of 3.85. In the test with 100 stable samples, all samples were located inside the ellipse, achieving zero missed detections. For newly added stable samples, if they were located inside the ellipse, no update was needed; if they were located outside the ellipse, an adaptive update process was triggered to recalculate the ellipse boundary so that it encompassed the new sample.
[0071] When the wavelet basis type is changed to db5 and the number of decomposition layers is increased to 5, constructing elliptical regions using the same procedure still achieves a stable and compact envelope for the samples. This demonstrates that the proposed method has good parameter adaptability and can be configured according to different simulation scenarios and accuracy requirements.
[0072] In summary, the embodiments of this disclosure provide a wavelet feature-based method and system for zero-missing-criteria transient voltage stability assessment based on a conservative stability domain. This method, through wavelet feature extraction and the construction of a minimum-volume closed ellipse, forms a complete technical chain from data to discrimination, providing an alternative solution for transient voltage stability assessment that balances low missing-criteria rate and interpretability.
[0073] Compared with existing technologies, the zero-omission judgment method and system for transient voltage stability wavelet features based on conservative stability domain in all electromagnetic simulation provided by this invention has the following beneficial effects: 1. Reduce the risk of missed detections from a structural perspective. This invention extracts physical features using discrete wavelet transform, dynamically mapping transient voltages to a two-dimensional space. Based on known stable samples, a compact envelope region is constructed using the minimum volume closed ellipse method, serving as a conservative stability region. The judgment rule is: a sample falling within the ellipse is considered stable, while falling outside the ellipse is considered unstable. Since the elliptical region strictly encloses all known stable samples, and unstable samples are usually distributed outside the range of stable samples in the feature space, this design, from an algorithmic construction perspective, reduces the probability of stable samples being misclassified as unstable, providing a feasible technical path for meeting the requirement of low false negative rates in engineering.
[0074] 2. The calculation structure is simple and suitable for batch screening. This invention addresses the characteristics of fully electromagnetic simulation data, employing discrete wavelet transform for feature extraction to separate dynamic information from different physical processes. The construction of the conservative stability region can be completed offline; online discrimination only requires calculating the relative position of the sample to be judged to the elliptical boundary, resulting in minimal computational overhead. The overall method has a clear structure, and parameters can be flexibly configured according to the simulation scenario and accuracy requirements, making it suitable for rapid screening of massive amounts of anticipated fault scenarios.
[0075] Exemplary device Figure 6 This is a schematic diagram of the structure of a novel power system transient voltage stability wavelet feature analysis device provided in an exemplary embodiment of the present invention. Figure 6 As shown, the device 600 includes: The acquisition module 610 is used to acquire node voltage timing data generated by the target power grid after a preset fault occurs under a full electromagnetic transient simulation environment. The conversion module 620 is used to convert nodal voltage time-series data into physically meaningful two-dimensional feature vectors using Daubechies wavelet pairs; Module 630 is used to select a sample set under known stable conditions from the two-dimensional feature vectors, and to construct a conservative stable region based on the sample set using the minimum volume closed ellipse method. The judgment module 640 is used to compare the feature vector of the sample to be judged with the conservative stable region to determine whether the preset conditions are met. The determination module 650 is used to determine whether the feature vector of the sample to be judged is transiently stable if the preset conditions are met, otherwise it is determined to be transiently unstable.
[0076] Optionally, Daubechies wavelet pairs are used to convert the nodal voltage time-series data into physically meaningful two-dimensional feature vectors, including: Daubechies wavelet was used to perform multi-level discrete wavelet transform on the node voltage time series data to obtain several levels of approximation coefficients and detail coefficients. Based on the mapping relationship between the subband frequency corresponding to each level of detail coefficients and the physical dynamics of the power grid, several levels of approximation coefficients and detail coefficients are divided into at least three coefficient groups with different physical meanings. Based on the pre-constructed calculation expressions of the first and second characteristic quantities, a standard two-dimensional orthogonal characteristic space independent of specific operating conditions is constructed. The first characteristic quantity represents the weighted comprehensive level of the switching process dynamic energy and the control dynamic energy, and the second characteristic quantity represents the modulation intensity of the interaction between the switching process energy and the control dynamic and electromechanical coupling dynamic. The coordinates of the first and second eigenvalues of the coefficient set are calculated according to the calculation expression, and all coordinates are mapped to the standard two-dimensional orthogonal feature space to form a two-dimensional feature vector representing the node state.
[0077] Optionally, the coefficient set includes: First coefficient group G sw Corresponding to the dynamics of the power electronic switching process, the sub-band frequency range of the detailed coefficient layer contained in the first coefficient group needs to cover the integer or half-integer multiples of the switching frequency of the power electronic converter. Second coefficient group G ctrl Corresponding to the converter control dynamics, the subband frequency range of the detailed coefficient layer contained in the second coefficient group mainly reflects the response frequency band of the current loop and voltage loop controllers. Third coefficient group G em Corresponding to the electromechanical coupling dynamics of the power grid, the third coefficient group includes a detailed coefficient layer or a lowest frequency approximation coefficient layer, whose sub-band frequency range mainly reflects the synchronous unit rotor motion and low-frequency oscillation electromechanical transient processes.
[0078] Optionally, the expression for calculating the first feature is: In the formula, E (·) is a function for calculating the total energy of the coefficient group; ω sw and ω ctrl The preset weighting coefficients are used, and they satisfy the following conditions: ω sw + ω ctrl = 1; The expression for calculating the second characteristic is: in, E ( G ctrl ) / E ( Gem ) reflects the ratio of control dynamic energy to electromechanical coupling dynamic energy; S (·) is a nonlinear modulation function, which is used to characterize the modulation effect of switching process energy on the ratio relationship, and is specifically S(x)=log(1+x) or S(x)=sqrt(x).
[0079] Optionally, the minimum volume enclosing ellipse is determined by solving the following optimization problem: minimize log det(Σ) subject to (xi - μ)ᵀΣ -1 (xi - μ) ≤ 1, i = 1, 2, ..., N where xi ∈ 2 is the two-dimensional feature vector of the i-th stable sample, μ ∈ 2 is the ellipse center, Σ ∈ 2x2 is a 2×2 positive definite matrix (covariance matrix), log det(Σ) represents the natural logarithm of the determinant of the matrix Σ, Σ is the positive definite matrix for constructing the conservative stability region, and N is the number of samples.
[0080] Optionally, the preset condition is: wherein, d is the Mahalanobis distance from the feature vector of the sample to be discriminated to the boundary of the conservative stability region; x is the two-dimensional feature vector of the sample to be discriminated, μ is the ellipse center, and Σ is the positive definite matrix for constructing the conservative stability region; when the sample to be discriminated meets the preset condition, it is determined that the transient voltage is stable, otherwise it is determined that the transient voltage is unstable.
[0081] Optionally, the method further includes: setting a buffer area outside the boundary of the conservative stability region, and defining two thresholds d1 and d2, where d1=1, d2>1. When the Mahalanobis distance d of the feature vector of the sample to be discriminated satisfies 1<d ≤ d2, the sample falls into the buffer area, triggering manual review or outputting a discrimination conclusion with confidence.
[0082] Optionally, the method further includes: when a new stable operating condition sample is added, incorporating the new sample into the original known stable sample set, re-solving the optimization problem, updating the ellipse boundary, and realizing adaptive update of the conservative stability region.
[0083] Exemplary electronic devices Figure 7 is the structure of the electronic device provided by an exemplary embodiment of the present invention. As shown Figure 7As shown, the electronic device 70 includes one or more processors 71 and a memory 72.
[0084] The processor 71 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.
[0085] The memory 72 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 71 may execute the program instructions to implement the methods of the software programs of the various embodiments of the present invention described above, and / or other desired functions. In one example, the electronic device may also include an input device 73 and an output device 74, these components being interconnected via a bus system and / or other forms of connection mechanisms (not shown).
[0086] In addition, the input device 73 may also include, for example, a keyboard, a mouse, etc.
[0087] The output device 74 can output various information to the outside. The output device 74 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.
[0088] Of course, for the sake of simplicity, Figure 7 Only some of the components of this electronic device relevant to the present invention are shown, omitting components such as buses, input / output interfaces, etc. In addition, the electronic device may include any other suitable components depending on the specific application.
[0089] Exemplary computer program products and computer-readable storage media In addition to the methods and apparatus described above, embodiments of the present invention may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the methods according to various embodiments of the present invention described in the "Exemplary Methods" section above.
[0090] The computer program product can be written in any combination of one or more programming languages to perform the operations of the embodiments of the present invention. The programming languages include object-oriented programming languages such as Java and C++, as well as conventional procedural programming languages such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0091] Furthermore, embodiments of the present invention may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps of the methods according to various embodiments of the present invention described in the "Exemplary Methods" section above.
[0092] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, system, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.
[0093] The basic principles of the present invention have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in the present invention are merely examples and not limitations, and should not be considered as essential features of each embodiment of the present invention. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the present invention to the necessity of employing the aforementioned specific details.
[0094] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For system embodiments, since they largely correspond to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.
[0095] The block diagrams of devices, systems, devices, and systems involved in this invention are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, systems, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.
[0096] The methods and systems of the present invention may be implemented in many ways. For example, they may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order of steps for the methods is for illustrative purposes only, and the steps of the methods of the present invention are not limited to the order specifically described above unless otherwise specifically stated. Furthermore, in some embodiments, the present invention may also be implemented as a program recorded on a recording medium, the program comprising machine-readable instructions for implementing the methods according to the present invention. Thus, the present invention also covers recording media storing programs for performing the methods according to the present invention.
[0097] It should also be noted that in the systems, apparatus, and methods of the present invention, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered equivalents of the present invention. The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the invention. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of the invention. Therefore, the invention is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.
[0098] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the invention to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.
Claims
1. A novel wavelet feature analysis method for transient voltage stability in power systems, characterized in that, include: Acquire node voltage time-series data generated after a preset fault occurs in the target power grid under a full electromagnetic transient simulation environment; The node voltage time series data are converted into a two-dimensional feature vector with physical meaning using Daubechies wavelet pairs; Select a sample set under known stable conditions from the two-dimensional feature vectors, and construct a conservative stability region based on the sample set using the minimum volume closed ellipse method; The feature vector of the sample to be judged is compared with the conservative stable region to determine whether the preset conditions are met. If the preset conditions are met, the feature vector of the sample to be judged is determined to be transiently stable; otherwise, it is determined to be transiently unstable.
2. The method according to claim 1, characterized in that, The node voltage time series data is converted into a physically meaningful two-dimensional feature vector using Daubechies wavelet pairs, including: The node voltage time series data were subjected to multi-level discrete wavelet transform using Daubechies wavelet to obtain several levels of approximation coefficients and detail coefficients. Based on the mapping relationship between the subband frequency corresponding to each level of detail coefficients and the physical dynamics of the power grid, the approximation coefficients and detail coefficients of the aforementioned levels are divided into at least three coefficient groups with different physical meanings. Based on the pre-constructed calculation expressions of the first and second characteristic quantities, a standard two-dimensional orthogonal characteristic space independent of specific operating conditions is constructed. The first characteristic quantity represents the weighted comprehensive level of the switching process dynamic energy and the control dynamic energy, and the second characteristic quantity represents the modulation intensity of the interaction between the switching process energy and the control dynamic and electromechanical coupling dynamic. The coordinates of the first and second feature values of the coefficient group are calculated according to the calculation expression, and all coordinate points are mapped to the standard two-dimensional orthogonal feature space to form the two-dimensional feature vector representing the node state.
3. The method according to claim 2, characterized in that, The coefficient set includes: First coefficient group G sw Corresponding to the dynamics of the power electronic switching process, the sub-band frequency range of the detailed coefficient layer contained in the first coefficient group needs to cover the integer or half-integer multiples of the switching frequency of the power electronic converter. Second coefficient group G ctrl Corresponding to the converter control dynamics, the subband frequency range of the detailed coefficient layer contained in the second coefficient group mainly reflects the response frequency band of the current loop and voltage loop controllers. Third coefficient group G em Corresponding to the electromechanical coupling dynamics of the power grid, the third coefficient group includes a detailed coefficient layer or a lowest frequency approximation coefficient layer, whose sub-band frequency range mainly reflects the synchronous unit rotor motion and low-frequency oscillation electromechanical transient processes.
4. The method according to claim 3, characterized in that, The expression for calculating the first feature is: In the formula, E (·) is a function for calculating the total energy of the coefficient group; ω sw and ω ctrl The preset weighting coefficients are used, and they satisfy the following conditions: ω sw + ω ctrl = 1; The expression for calculating the second characteristic is: in, E ( G ctrl ) / E ( G em This reflects the ratio of control dynamic energy to electromechanical coupling dynamic energy. S (·) is a nonlinear modulation function used to characterize the modulation effect of the switching process energy on the ratio relationship, specifically S(x)=log(1+x) or S(x)=sqrt(x).
5. The method according to claim 1, characterized in that, The minimum volume closed ellipse is determined by solving the following optimization problem: minimize log det(Σ) subject to (xi - μ)ᵀΣ -1 (xi - μ) ≤ 1, i = 1, 2, ..., N In the formula, xi ∈ 2 Let μ be the two-dimensional feature vector of the i-th stable sample. 2 Let be the center of the ellipse, Σ ∈ 2x2 It is a 2×2 positive definite matrix (covariance matrix), log det(Σ) represents the natural logarithm of the determinant of matrix Σ, Σ is the positive definite matrix for constructing the conservative stability region, and N is the number of samples.
6. The method according to claim 1, characterized in that, The preset conditions are: In the formula, d Σ is the Mahalanobis distance from the feature vector of the sample to be judged to the boundary of the conservative stable domain; x is the two-dimensional feature vector of the sample to be judged, μ is the center of the ellipse, and Σ is the positive definite matrix for constructing the conservative stable domain; when the sample to be judged meets the preset conditions, it is determined to be transient voltage stable, otherwise it is determined to be transient voltage unstable.
7. The method according to claim 6, characterized in that, Also includes: A buffer region is set outside the boundary of the conservative stable region, and two thresholds d1 and d2 are defined, where d1=1 and d2>1. When the Mahalanobis distance d of the feature vector of the sample to be judged satisfies 1< d ≤ d2, the sample falls into the buffer region, triggering manual review or outputting a judgment conclusion with confidence.
8. The method according to claim 1, characterized in that, Also includes: When a new stable operating condition sample is added, the new sample is incorporated into the existing known stable sample set, the optimization problem is solved again, the elliptical boundary is updated, and the conservative stable domain is adaptively updated.
9. A novel wavelet feature analysis device for transient voltage stability in a power system, characterized in that, include: The acquisition module is used to acquire node voltage timing data generated by the target power grid after a preset fault occurs under a full electromagnetic transient simulation environment. The conversion module is used to convert the node voltage time series data into a two-dimensional feature vector with physical meaning using Daubechies wavelet pairs; The construction module is used to select a sample set under known stable conditions from the two-dimensional feature vector, and to construct a conservative stable region based on the sample set using the minimum volume closed ellipse method. The judgment module is used to compare the feature vector of the sample to be judged with the conservative stable region to determine whether the preset conditions are met. The determination module is used to determine whether the feature vector of the sample to be judged is transiently stable if the preset conditions are met, otherwise it is determined to be transiently unstable.
10. The apparatus according to claim 9, characterized in that, The node voltage time series data is converted into a physically meaningful two-dimensional feature vector using Daubechies wavelet pairs, including: The node voltage time series data were subjected to multi-level discrete wavelet transform using Daubechies wavelet to obtain several levels of approximation coefficients and detail coefficients. Based on the mapping relationship between the subband frequency corresponding to each level of detail coefficients and the physical dynamics of the power grid, the approximation coefficients and detail coefficients of the aforementioned levels are divided into at least three coefficient groups with different physical meanings. Based on the pre-constructed calculation expressions of the first and second characteristic quantities, a standard two-dimensional orthogonal characteristic space independent of specific operating conditions is constructed. The first characteristic quantity represents the weighted comprehensive level of the switching process dynamic energy and the control dynamic energy, and the second characteristic quantity represents the modulation intensity of the interaction between the switching process energy and the control dynamic and electromechanical coupling dynamic. The coordinates of the first and second feature values of the coefficient group are calculated according to the calculation expression, and all coordinate points are mapped to the standard two-dimensional orthogonal feature space to form the two-dimensional feature vector representing the node state.
11. The apparatus according to claim 10, characterized in that, The coefficient set includes: The first coefficient group Gsw corresponds to the dynamics of the power electronic switching process. The sub-band frequency range of the detailed coefficient layer contained in the first coefficient group needs to cover the integer or half-integer multiples of the switching frequency of the power electronic converter. The second coefficient group Gctrl corresponds to the converter control dynamics. The sub-band frequency range of the detailed coefficient layer contained in the second coefficient group mainly reflects the response frequency band of the current loop and voltage loop controllers. The third coefficient group Gem corresponds to the electromechanical coupling dynamics of the power grid. The third coefficient group includes a detailed coefficient layer or the lowest frequency approximation coefficient layer, and its sub-band frequency range mainly reflects the synchronous unit rotor motion and low-frequency oscillation electromechanical transient processes.
12. A computer-readable storage medium, characterized in that, The storage medium stores a computer program for performing the method described in any one of claims 1-8.
13. An electronic device, characterized in that, The electronic device includes: processor; Memory used to store the processor's executable instructions; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method described in any one of claims 1-8.
