Power grid oscillation source detection method and system based on spatial Bayesian model

Abstract

The invention discloses a power grid oscillation source detection method and system based on a spatial Bayesian model. The method comprises the following steps: carrying out data acquisition and preprocessing on an electrical characteristic time sequence signal of each node in a power grid; node time-frequency feature extraction is carried out on the preprocessed electrical feature time sequence signals; constructing a power grid node space Bayesian prior model according to the topological line connection relationship between the power grid nodes to obtain the prior probability that the nodes have oscillation risks; constructing a comprehensive likelihood function of the node as an oscillation source according to the time-frequency characteristics of the node; and performing Bayesian fusion to calculate posterior probability distribution of each node serving as an oscillation source by using the prior probability that the nodes have oscillation risks and the comprehensive likelihood function that the nodes serve as the oscillation source. According to the method, power grid oscillation source detection is carried out based on the spatial Bayesian model, the most probable oscillation source position is given, observation node probability statistics of the whole network and all areas is achieved, and uncertainty information of quantitative power grid oscillation source detection is provided for dispatchers.

Description

Technical Field

[0001] This invention relates to the field of oscillation source detection technology in power systems, and specifically to a power grid oscillation source detection method and system based on a spatial Bayesian model. Background Technology

[0002] With the deployment of synchronous phasor measurement units (PMUs), waveform measurement units (WMUs), and wide-area measurement systems (WAMS), more and more nodes in power systems can acquire high-precision voltage and current measurement data in real time. Based on this data, a relatively mature oscillation monitoring technology system has been constructed, mainly including: 1) Oscillation detection, which uses methods such as spectral analysis, time-frequency analysis, wavelet analysis, and modal analysis to identify the existence of continuous oscillations or low-damping modes; 2) Oscillation feature extraction, which extracts features such as oscillation frequency, damping ratio, amplitude, and duration; 3) Alarm and recording: once oscillation is detected, an alarm is generated according to a preset threshold, and relevant data is recorded and archived. In the field of oscillation monitoring, common time-frequency analysis methods include, but are not limited to: Short-Time Fourier Transform (STFT), Wavelet Transform (WT), Wavelet Packet Analysis, S-Transform and its variants, filter banks, and methods combining bandpass filtering with Hilbert Transform. These methods can display the energy distribution of the signal on the time-frequency plane, thereby characterizing the occurrence time, duration, and frequency changes of oscillations. When analyzing multiple nodes simultaneously, the oscillation energy variation process of each node within a certain frequency band, the dynamic change of phase difference between nodes over time, and a rough estimate of the energy diffusion or propagation mode between nodes can be obtained. Existing technologies for locating oscillation sources can be broadly categorized as follows: 1) Model- and modal analysis-based methods, which calculate the eigenvectors and participation factors of each mode based on a linearized system model and eigenvalue analysis, identifying the terminal or node that contributes most to a particular oscillation mode; some methods also combine damping sensitivity analysis to assess the influence of different nodes or devices on damping. 2) Energy balance or power flow-based methods, which use generalized energy functions or power imbalance theory to calculate energy injection and absorption points during oscillation; and determine the region where the oscillation source is located by analyzing the energy flow direction. 3) Measurement-driven methods based on phase difference and amplitude ratio, which use voltage / current phase measurements from a PMU to construct "phase gradient" or "phase cone" criteria, and make empirical judgments based on the relative magnitude and decay rate of oscillation amplitudes at different nodes. 4) Machine learning or deep learning-based methods utilize large amounts of simulated or historical perturbation data to train classification models (e.g., "a node is a source / non-source"); convolutional neural networks, recurrent neural networks, or graph neural networks are used to fit complex mapping relationships. Beyond power systems, spatial statistics and spatial Bayesian models have been widely applied in epidemiology, environmental science, and traffic safety. Typical examples include: the ICAR (Intrinsic Conditional Autoregressive) model, which constrains the difference in random effects between adjacent regions based on regional adjacency relationships, forming a spatially smooth random field; and the BYM (Besag-York-Mollié) model, which introduces unstructured random effects on top of ICAR, mixing structured spatial effects with independent noise to model disease incidence rates, accident risks, etc., in various regions.Its key features include: using the "adjacency matrix W" to characterize spatial correlation; treating "risk" as a spatially varying random variable; and integrating covariates (such as socioeconomic indicators) and historical data to output the risk distribution for each region. Therefore, how to introduce spatial Bayesian models into the assessment of power grid oscillation sources for the first time remains a critical technical problem that urgently needs to be solved. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a method and system for detecting power grid oscillation sources based on a spatial Bayesian model, which addresses the above-mentioned problems in the prior art. This invention aims to detect power grid oscillation sources based on a spatial Bayesian model, provide the most likely location of the oscillation source, realize the probability statistics of observation nodes in the whole network and each region, and provide dispatchers with quantitative information on the uncertainty of power grid oscillation source detection.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for detecting power grid oscillation sources based on a spatial Bayesian model includes the following steps: S1, to collect and preprocess the electrical characteristic time-series signals of each node in the power grid; S2, extract node time-frequency features from the preprocessed electrical characteristic time-series signal; S3. Construct a spatial Bayesian prior model of the power grid nodes based on the topological connection relationships between the power grid nodes to obtain the prior probability that node i has the risk of oscillation. Construct a comprehensive likelihood function for nodes as oscillation sources based on their time-frequency characteristics. S4, utilizing the prior probability that node i has the risk of oscillation. The combined likelihood function of the nodes as oscillation sources is used to perform Bayesian fusion calculation of the posterior probability distribution of each node as an oscillation source.

[0005] Optionally, the preprocessing in step S1 includes time alignment, filtering and noise reduction, removal of trend components, and normalization of amplitude. The removal of trend components is to subtract the average value of the electrical characteristic timing signals within a specified time window from the electrical characteristic timing signals.

[0006] Optionally, when extracting node time-frequency features in step S2, the extracted node time-frequency features include energy spectrum, response time difference, and phase difference, wherein the response time difference is the first time that the energy spectrum of the node and the nodes affected by it exceeds a preset threshold E at the target frequency. th The time difference between moments.

[0007] Optionally, the construction of the spatial Bayesian prior model of the power grid nodes based on the topological connection relationships between power grid nodes in step S3 includes: S3.1, Generate an adjacency matrix based on the topological connection relationships between power grid nodes: ; in, Adjacency matrix The element in the i-th row and j-th column; S3.2, Define the spatial structuring effect for each node i. To reflect the spatial smoothness of the risk of node i relative to its neighboring nodes, the spatial structuring effect. The smoothing constraint is: ; in, For spatial structuring effect Smoothing constraints, Indicating spatial structuring effect, and The spatial structuring effects of nodes i and j are respectively; and the spatial structuring effect of node i is... The smoothness constraint corresponds to a Gaussian prior: ; in, For spatial structuring effect Gaussian priors, Define the variance of the spatially structured effects; define the unstructured effects for each node i. To characterize the independent differences of each node, we use an independent Gaussian distribution model as follows: ; in, This indicates that the mean is 0 and the variance is 0. , an independent Gaussian distribution; S3.3, for each node i, based on its spatial structuring effect Unstructured effects Determine the definition of the overall prior risk value for node i: ; in, Let be the overall prior risk value of node i. Scaling factor These are the weighting coefficients for structured and unstructured effects, and they have... ; S3.4, use historical oscillation event data of the power grid to estimate the spatial structuring effect of each node i. Unstructured effects Weighting coefficients Variance of spatial structuring effect and the variance of the independent Gaussian distribution We obtain the prior probability that each node i has the risk of oscillation. .

[0008] Optionally, in step S3, the functional expression for constructing the comprehensive likelihood function of the node as an oscillation source based on the node's time-frequency characteristics is as follows: ; ; ; ; in, Let be the comprehensive likelihood function for node i. , and These are the likelihood function components for the energy spectrum, response time difference, and phase difference, respectively. , and These are the weighting coefficients for the energy spectrum, response time difference, and phase difference, respectively. Let i be the observed energy at the dominant oscillation frequency. The expected energy level of the source node. Let Variance be the variance of the energy spectrum. The variance of the response time difference, divide and sum The energy spectrum of node i and its affected node j at the target frequency exceeds the preset threshold E for the first time. th At that moment, The variance of the phase difference. Let be the phase difference between node i and the node j affected by it. Let be the expected value of the phase difference between node i and node j affected by it.

[0009] Optionally, in step S4, the function expression for calculating the posterior probability distribution of each node as an oscillation source using Bayesian fusion is as follows: ; in, Let be the posterior probability that node i is an oscillation source given data, where data represents the node time-frequency characteristics and oscillation start time of each node in the power grid. Let be the prior probability that node i has the risk of oscillation. Let be the prior probability that node j has the risk of oscillation. Let be the comprehensive likelihood function for node i. Let N be the comprehensive likelihood function of node j, and N be the number of nodes in the power grid.

[0010] Optionally, after step S4, the method further includes visualizing the posterior probability distribution of each node as an oscillation source. The visualization results include part or all of the node ranking table, topology probability map, regional probability map, and multi-source identification map. The node ranking table includes a sorted list of the posterior probabilities of each node i as an oscillation source. The topology probability map is an image showing the posterior probabilities of each node i as an oscillation source in the power grid node topology map. The regional probability map is an image of the posterior probability of each node i as an oscillation source within a given region. The multi-source identification map is an image of one or more nodes i marked as oscillation sources whose posterior probabilities exceed a preset threshold.

[0011] The present invention also provides a power grid oscillation source detection system based on a spatial Bayesian model, comprising a microprocessor and a memory interconnected thereto, wherein the microprocessor is programmed or configured to execute the power grid oscillation source detection method based on the spatial Bayesian model.

[0012] The present invention also provides a computer-readable storage medium storing a computer program or instructions that are programmed or configured to execute the power grid oscillation source detection method based on the spatial Bayesian model by a processor.

[0013] The present invention also provides a computer program product, including a computer program or instructions, which are programmed or configured to execute the power grid oscillation source detection method based on the spatial Bayesian model via a processor.

[0014] Compared with existing technologies, this invention mainly achieves the following beneficial effects: Based on a spatial Bayesian model, this invention comprehensively utilizes power grid topology, historical oscillation statistics, and real-time time-frequency characteristics to output the probability of power grid oscillation sources as a "posterior probability distribution of oscillation sources" within a unified Bayesian inference framework. Its advantages include: 1) This invention solves the problem of "only providing single-point location and lacking uncertainty characterization." By constructing a spatial Bayesian model and node likelihood functions, it outputs the posterior probability of all network nodes, forming an oscillation source probability distribution map, thus achieving a quantitative expression and ranking of source location uncertainty. 2) This invention solves the problem of "ignoring spatial correlation and failing to fully utilize power grid topology information." By introducing spatial Bayesian models such as the adjacency matrix W and ICAR / BYM, it models the risk of adjacent nodes through a spatial random field, achieving spatial smoothing and propagation of risk, thereby improving the stability and robustness of source location results. 3) This invention solves the problem of "being unable to utilize long-term historical data to form robust priors." The activity level, fault type, and operating condition information of each node in historical oscillation events can be mapped as model parameters and covariates to construct an updatable prior for node oscillation risk, enabling the system to learn from historical experience and update in real time. 4) This invention can solve the problem of "weak multi-source oscillation identification capability". Through the multi-peak structure of the Bayesian posterior distributed in the node space, it naturally supports multi-source scenarios: multiple nodes can simultaneously have high posterior probabilities, and the system can identify and display multiple suspected source regions and their relative probabilities. 5) This invention can improve the localization robustness in weak oscillation, short-term oscillation, and high-frequency oscillation scenarios. By reasonably characterizing the relationship between oscillation signals and noise in the likelihood function and combining spatial priors for constraints, even if the signal of a single event is weak, a relatively reliable source probability assessment can still be obtained through "prior + neighborhood information + accumulated history". 6) This invention can establish a unified "prior-likelihood-posterior" Bayesian inference logic closed loop. Using a spatial Bayesian model as a priori, a likelihood function is constructed based on the node oscillation response characteristics, and Bayesian formula is used for fusion inference. This allows for a unified mathematical characterization of the relationship between power grid topology, historical events, and real-time data, providing standardized probability outputs for subsequent risk assessment and pre-control strategy optimization. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the basic process of the method in an embodiment of the present invention.

[0016] Figure 2 This is a schematic diagram of the module division of the method in an embodiment of the present invention. Detailed Implementation

[0017] The core idea of ​​this invention is to model the oscillation risk of power grid nodes as a "spatial random field" constrained by the topology, based on multi-node synchronous measurement. A prior risk distribution is formed using historical oscillation data, and a likelihood function is constructed by combining the time-frequency response characteristics of each node in a specific oscillation event. Finally, Bayes' theorem is used to obtain the posterior probability distribution of each node as an oscillation source. To enable those skilled in the art to better understand the technical solution of this invention, the following will provide a more detailed description of the technical solution in conjunction with the accompanying drawings of the embodiments of this invention.

[0018] like Figure 1 As shown, the power grid oscillation source detection method based on the spatial Bayesian model in this embodiment includes the following steps: S1, to collect and preprocess the electrical characteristic time-series signals of each node in the power grid; S2, extract node time-frequency features from the preprocessed electrical characteristic time-series signal; S3. Construct a spatial Bayesian prior model of the power grid nodes based on the topological connection relationships between the power grid nodes to obtain the prior probability that node i has the risk of oscillation. Construct a comprehensive likelihood function for nodes as oscillation sources based on their time-frequency characteristics. S4, utilizing the prior probability that node i has the risk of oscillation. The combined likelihood function of the nodes as oscillation sources is used to perform Bayesian fusion calculation of the posterior probability distribution of each node as an oscillation source.

[0019] like Figure 2As shown, the power grid oscillation source detection method based on the spatial Bayesian model in this embodiment is specifically divided into a data acquisition module (for data acquisition in step S1), a data processing module (for preprocessing in step S1), a time-frequency feature extraction module (for step S2), a spatial prior modeling module (for constructing the spatial Bayesian prior model of the power grid nodes in step S3), a likelihood calculation module (for constructing the comprehensive likelihood function in step S3), a Bayesian fusion module (for step S4), and a model update module. After the event ends, the model update module updates the spatial prior model according to the new event record, enabling the system to have long-term self-learning capability. The modules work collaboratively as follows: the data acquisition module interfaces with the power grid synchronous measurement system, acquiring data periodically or when events are triggered; the data preprocessing module cleans the data and outputs it in a unified format; the time-frequency feature extraction module can be implemented based on CPU or GPU, adapting to real-time or near-real-time computing needs; the spatial prior modeling module can run offline, periodically updating parameters using historical oscillation events; the likelihood calculation module runs online after each oscillation event occurs; the Bayesian fusion module typically has a small computational load and can be completed within seconds; the source probability evaluation module displays the probability results to the dispatcher or sends them to other safety control modules; and the model update module updates the spatial prior model based on new event records after the event ends, enabling the system to have long-term self-learning capabilities.

[0020] In step S1 of this embodiment, when data acquisition of the electrical characteristic timing signals of each node in the power grid is performed, the data acquisition module is connected to the synchronous measurement device in the power grid, including a PMU (Phasor Measurement Unit) and a WMU (Waveform Measurement Unit). The voltage and current timing data of each node in the power grid are collected by the PMU and WMU as electrical characteristic timing signals. The acquired original electrical characteristic timing signals can be recorded as: ; in, i Number the nodes. t For time, N This represents the total number of monitoring nodes.

[0021] The preprocessing in step S1 of this embodiment includes time alignment, filtering and noise reduction, removal of trend components, and amplitude normalization. Removal of trend components involves subtracting the average value (DC component) of the electrical characteristic time-series signal within a specified time window from the electrical characteristic time-series signal. The electrical characteristic time-series signal can be a voltage or current signal, etc. Specifically, in this embodiment, to ensure the accuracy of subsequent analysis, the data preprocessing module performs the following processing on the original signal: Time alignment: Using GPS or other precision clock sources, the data of each node is aligned to a unified time reference; Filtering and noise reduction: Bandpass filtering, wavelet denoising, and other methods are used to suppress power frequency offset, DC bias, and high-frequency noise; De-trending and normalization: Trend components are removed, and the amplitude is normalized to make the oscillation characteristics of different nodes comparable. The signal obtained after preprocessing is denoted as... This serves as the input for subsequent time-frequency analysis.

[0022] In step S2 of this embodiment, when extracting the node time-frequency features, the extracted node time-frequency features include energy spectrum, response time difference, and phase difference. The response time difference is the first time that the energy spectrum of the node and the nodes affected by it at the target frequency exceeds a preset threshold E. th The time difference between moments.

[0023] In this embodiment, the time-frequency feature extraction module extracts signals from each node. A time-frequency transformation is performed to obtain the energy distribution and phase information in the time-frequency plane. This invention is not limited to a specific time-frequency analysis method; Short-Time Fourier Transform (STFT), Wavelet Transform (WT), S-Transform, etc., can be used. Taking the S-Transform as an example, the node... i The time-frequency representation can be written as: ; in, For nodes i In frequency f and time t Complex time-frequency coefficients under the given conditions.

[0024] Therefore, the oscillation energy spectrum and the oscillation phase trajectory are defined as follows:

[0025] ; In specific calculations, this invention can select a set of frequencies of interest based on the oscillation frequency band (such as low frequency, subsynchronous, or high frequency). Features are extracted at these frequencies.

[0026] To reflect the relative response between nodes, this embodiment further calculates the phase difference and energy difference between nodes, for example: ; ; in This refers to the current dominant frequency or frequency band center of the oscillation. Through analysis... and By observing the changes over time, information about the possible propagation direction of the source node can be obtained.

[0027] In addition, the "start time" of oscillation at each node is extracted, i.e., the node is defined. i In frequency The energy on the surface first exceeds a certain threshold E th The time is: This allows us to obtain the response time difference between nodes: ; The aforementioned energy spectrum, phase difference, and response time difference characteristics together constitute the input of the likelihood function in step S3.

[0028] The spatial prior modeling module constructs a spatial Bayesian model of "node oscillation risk" using power grid topology information. In this embodiment, step S3, constructing the spatial Bayesian prior model of the power grid nodes based on the topological line connection relationships between power grid nodes, includes: S3.1, Generate an adjacency matrix based on the topological connection relationships between power grid nodes: ; in, Adjacency matrix The element in the i-th row and j-th column; S3.2 In this embodiment, ICAR (Intrinsic Conditional Auto Regression) is used to describe spatial correlation, introducing two stochastic effects: defining a spatial structuring effect for each node i. To reflect the spatial smoothness of the risk of node i relative to its neighboring nodes, the spatial structuring effect. The smoothing constraint is: ; in, For spatial structuring effect Smoothing constraints, Indicating spatial structuring effect, and The spatial structuring effects of nodes i and j are respectively; and the spatial structuring effect of node i is... The smoothness constraint corresponds to a Gaussian prior: ; in, For spatial structuring effect Gaussian priors, Define the variance of the spatially structured effects; define the unstructured effects for each node i. To characterize the independent differences of each node, we use an independent Gaussian distribution model as follows: ; in, This indicates that the mean is 0 and the variance is 0. , an independent Gaussian distribution; S3.3, for each node i, based on its spatial structuring effect Unstructured effects Determine the definition of the overall prior risk value for node i: ; in, Let be the overall prior risk value of node i. Scaling factor These are the weighting coefficients for structured and unstructured effects, and they have... ; S3.4, use historical oscillation event data of the power grid to estimate the spatial structuring effect of each node i. Unstructured effects Weighting coefficients Variance of spatial structuring effect and the variance of the independent Gaussian distribution We obtain the prior probability that each node i has the risk of oscillation. This prior can be automatically updated through historical oscillation events.

[0029] Assuming node i is the source of this oscillation, it should exhibit specific characteristics in terms of energy, phase, and response timing. Therefore, likelihood function components based on oscillation energy, response time, and phase difference are constructed. In this embodiment, the likelihood function components of the energy spectrum are: ; in, Let i be the observed energy at the dominant oscillation frequency. The expected energy level of the source node (which can be obtained statistically from historical source events or adaptively estimated). The variance of the energy spectrum; In this embodiment, the likelihood function component of the response time difference is: ; in, The variance of the response time difference, divide and sum The energy spectrum of node i and its affected node j at the target frequency exceeds the preset threshold E for the first time. th At the moment when multiple nodes are significantly later than node i When oscillation begins, it will increase. The value of .

[0030] In oscillation propagation, the phase of nodes near the source typically exhibits a certain gradient characteristic. A phase error metric can be defined for a specific moment or time window; therefore, in this embodiment, the likelihood function components of the phase difference are: ; in, The variance of the phase difference. Let be the phase difference between node i and the node j affected by it. Let be the expected value of the phase difference between node i and the affected node j, that is, the expected pattern of the phase difference between node i and node j when node i is the source, obtained from historical events or simulations.

[0031] In step S3, the expression for the comprehensive likelihood function of the nodes as oscillation sources, constructed based on the time-frequency characteristics of the nodes, is as follows: ; in, Let be the comprehensive likelihood function for node i. , and These are the likelihood function components for the energy spectrum, response time difference, and phase difference, respectively. , and These are the weighting coefficients for the energy spectrum, response time difference, and phase difference, respectively. These weighting coefficients can be determined using historical data or experience.

[0032] The Bayesian fusion module fuses the prior probability P(i) obtained in step S3 with the likelihood function L(i) to calculate the posterior probability of each node as an oscillation source. In this embodiment, the function expression for calculating the posterior probability distribution of each node as an oscillation source using Bayesian fusion in step S4 is as follows: ; in, Let be the posterior probability that node i is an oscillation source given data, where data represents the node time-frequency characteristics and oscillation start time of each node in the power grid. Let be the prior probability that node i has the risk of oscillation. Let be the prior probability that node j has the risk of oscillation. Let be the comprehensive likelihood function for node i. Let be the comprehensive likelihood function of node j, and N be the number of nodes in the power grid. Based on the above Bayesian fusion inference method, when a node has high historical risk (large prior) and its current event response pattern also matches the source characteristics (large likelihood), its posterior probability will be significantly increased. If a node has high historical risk, but its current response is clearly inconsistent with the source characteristics, the likelihood term will "correct" its probability to avoid blindly relying on the prior. If the current event signal is weak and the noise is large, the prior will account for a larger proportion in the posterior, and the system can still output a relatively smooth source probability distribution, improving robustness.

[0033] The oscillation source probability assessment module organizes and displays the posterior probability distribution. For example... Figure 1 As shown, as an optional implementation, after step S4 in this embodiment, the posterior probability distribution of each node as an oscillation source is visualized. The visualization results include part or all of the node ranking table, topology probability map, regional probability map, and multi-source identification map. The node ranking table includes a sorted list of the posterior probabilities of each node i as an oscillation source. The topology probability map is an image showing the posterior probabilities of each node i as an oscillation source in the power grid node topology map. The regional probability map is an image of the posterior probabilities of each node i as an oscillation source within a given region. The multi-source identification map is an image of one or more nodes i marked as oscillation sources whose posterior probabilities exceed a preset threshold. This embodiment can not only provide the most likely location of the oscillation source, but also provide the probability ranking and regional probability map of all observed nodes in the entire network, thereby providing dispatchers with quantified uncertainty information. For example, in this embodiment, the node sorting table, topology probability map, regional probability map, and multi-source identification map are generated as follows: 1) Node sorting table: All nodes are arranged from high to low according to their posterior probability, and the top few are listed as key targets for investigation; 2) Topology probability map: On the power grid topology map, the probability is represented by the size of the node circle or the shade of gray, forming an "oscillation source probability heat map"; 3) Regional probability map: Nodes are assigned to different regions (such as substations, sub-regions, provinces), and the probabilities of nodes in each region are summed or weighted averaged to generate the regional oscillation source probability; 4) Multi-source identification map: When there are multiple obvious peaks in the posterior distribution, multiple high-probability nodes or regions can be marked simultaneously to indicate the possible existence of multi-source oscillations. Through the above outputs, not only is the most likely source location given, but also complete probability distribution information is provided, providing dispatchers with a more comprehensive and quantitative basis for decision-making.

[0034] In summary, most existing technologies only make deterministic judgments on single oscillation events, relying either on idealized system models and modal analysis or on single energy, phase, or threshold rules. They cannot systematically utilize historical information and grid topology, nor can they quantify uncertainty. Therefore, they are prone to misjudgment or unstable results in complex scenarios such as multi-source oscillations, weak oscillations, and high-frequency oscillations. This embodiment's grid oscillation source detection method based on a spatial Bayesian model first constructs a spatial Bayesian prior model based on the grid topology, transforming long-term statistical and structural information such as "a node has a long-term high frequency of oscillations and strong coupling with surrounding nodes" into a computable node risk prior. Then, it constructs a likelihood function "with nodes as sources" through multi-node time-frequency characteristics, reflecting the degree of matching between the actual response of each node and the physical behavior of the source node in this event. Because the prior and likelihood are integrated within a unified Bayesian framework, the system can achieve a balance between historical and real-time information. Even if the signal of a single event is weak or has high noise, robust source probability assessments can be obtained through prior and neighborhood constraints. This embodiment of the power grid oscillation source detection method based on the spatial Bayesian model has the following technical effects: 1) It elevates the traditional problem of "deterministic localization" of a single oscillation source to a new paradigm of "probability assessment of oscillation sources" for all nodes in the entire network. This embodiment outputs the probability distribution of each node as an oscillation source, as well as the node ranking and regional probability map, rather than just giving the judgment result of a single source node. 2) For the first time, the oscillation risk of power grid nodes is regarded as a "spatial random field," and the spatial correlation between nodes is constructed based on the power grid topology. By introducing structured spatial effects and unstructured random effects, a spatial Bayesian prior distribution of node risk is formed, and this prior can be continuously updated using historical oscillation events. 3) Based on multi-node synchronous measurement, the node likelihood function is constructed using the time-frequency characteristics of each node, such as oscillation energy, phase change, and oscillation start time, to determine "the reasonableness of the current observation when a certain node is the current oscillation source." 4) The spatial Bayesian model provides the node risk prior, the likelihood function is constructed using the node time-frequency characteristics, and the posterior probability of each node as an oscillation source is obtained through Bayesian updates. This framework unifies power grid topology information, historical oscillation statistics, and real-time oscillation response information. 5) The probabilistic framework of this embodiment does not force the assumption of a "unique oscillation source," allowing the posterior distribution to present multiple high-probability nodes or regions, naturally supporting multi-source oscillation identification. Simultaneously, by leveraging spatial prior and neighborhood information constraints, even if a single event signal is weak or noisy, a relatively stable source probability estimate can be obtained, improving adaptability to complex scenarios such as weak oscillations, short-term oscillations, and high-frequency power electronic oscillations. Furthermore, the power grid oscillation source detection method based on the spatial Bayesian model in this embodiment outputs the posterior probability distribution of oscillation sources for all nodes in the entire network, rather than a hard decision for a single source node.This probabilistic output directly addresses the actual decision-making needs in scheduling and operation and maintenance: on the one hand, dispatchers can sort and grade nodes or regions according to their probability; on the other hand, when there are multi-source oscillations or regional oscillations, this embodiment will not forcibly "select only one point," but will present potential source regions in the form of multiple high-probability peaks, thus better reflecting the actual situation of oscillation mechanisms in complex power grids. In summary, this embodiment overcomes the problems of insufficient utilization of spatial correlation, inability to quantify uncertainty, difficulty in determining in multi-source scenarios, and poor robustness to weak oscillations in existing methods through a complete technical solution of spatial Bayesian prior, time-frequency likelihood modeling, and Bayesian fusion. It achieves a more reliable, interpretable, and self-learning evolutionary power grid oscillation source probability assessment capability.

[0035] Furthermore, this embodiment also provides a power grid oscillation source detection system based on a spatial Bayesian model, including a microprocessor and a memory interconnected, wherein the microprocessor is programmed or configured to execute the power grid oscillation source detection method based on the spatial Bayesian model. The present invention also provides a computer-readable storage medium storing a computer program or instructions programmed or configured to execute the power grid oscillation source detection method based on the spatial Bayesian model via a processor. The present invention also provides a computer program product including a computer program or instructions programmed or configured to execute the power grid oscillation source detection method based on the spatial Bayesian model via a processor.

[0036] Those skilled in the art will understand that the technical solutions provided by this invention may take the form of a method, system, or computer program product. Therefore, this invention may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this invention may take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, produce an implementation of the flowchart... Figure 1 One or more processes and / or boxes Figure 1The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0037] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for power grid oscillation source detection based on spatial Bayesian model, characterized in that, The method comprises the following steps: S1, collecting and preprocessing the electrical characteristic time series signals of each node in the power grid; S2, extracting the node time-frequency characteristics of the preprocessed electrical characteristic time series signals; S3, constructing a power grid node space Bayesian prior model according to a topological line connection relationship between grid nodes to obtain a prior probability that node i has an oscillation risk ; constructing a comprehensive likelihood function of a node as an oscillation source according to node time-frequency characteristics; S4, using the prior probability that node i has the risk of oscillation and the integrated likelihood function of the node as the source of oscillation, the posterior probability distribution of each node as the source of oscillation is calculated by Bayesian fusion.

2. The method for power grid oscillation source detection based on spatial Bayesian model according to claim 1, characterized in that, The preprocessing in step S1 comprises time alignment, filtering and denoising, removing the trend component, and normalizing the amplitude, wherein the removing the trend component is subtracting the average value of the electrical characteristic time series signals in a specified size time window from the electrical characteristic time series signals.

3. The method for power grid oscillation source detection based on spatial Bayesian model according to claim 1, characterized in that, The node time-frequency features extracted in step S2 include energy spectrum, response time difference and phase difference, wherein the response time difference is the time difference between the time when the energy spectrum of the node and the nodes affected by the node first exceeds a preset threshold E th at the target frequency.

4. The method for power grid oscillation source detection based on spatial Bayesian model according to claim 1, characterized in that, The step S3 of constructing the power grid node space Bayesian prior model according to the topological line connection relationship between the nodes comprises: S3.1, generating an adjacency matrix according to the topological line connection relationship between the nodes: ； wherein is the ith row jth column element in the adjacency matrix is the ith row jth column element in the adjacency matrix S3.2 Define a spatial structuring effect for each node i to reflect that the risk of node i is spatially smoothed with respect to neighboring nodes. The smoothing constraint for the spatial structuring effect of node i is: ​ ； where the smoothness constraint for the spatial structuring effect of the spatial structuring effect, denotes the spatial structuring effect, and are the spatial structuring effects of nodes i and j, respectively; and the spatial structuring effect of node i the smoothness constraint for the spatial structuring effect corresponds to a Gaussian-type prior: ； where, for the spatially structured effects Gaussian priors, for the variance of the spatially structured effects; for each node i define unstructured effects to capture the independent variation of the nodes themselves and model with an independent Gaussian distribution of the form: ； wherein, represents an independent Gaussian distribution with mean 0 and variance of 1. S3.3 For each node i, determine its overall prior risk value, based on its spatially structured effects and unstructured effects Definition of the overall prior risk value for node i: ； wherein, is the overall prior risk value for node i, is a scaling factor, is a weight coefficient for structured and unstructured effects, and has ; S3.4, estimate the spatially structured effect for each node i using historical oscillation event data of the power grid , unstructured effect , weight coefficient , variance of the spatially structured effect , and variance of the independent Gaussian distribution , obtain the prior probability that each node i has an oscillation risk .

5. The method for power grid oscillation source detection based on spatial Bayesian model according to claim 1, characterized in that, The function expression of the step S3 of constructing the comprehensive likelihood function of the node as an oscillation source according to the node time-frequency characteristics is: ； ； ； ； wherein, is the combined likelihood function for node i, , and are the likelihood function components for the energy spectrum, the response time difference, and the phase difference, respectively, , and are the weight coefficients for the energy spectrum, the response time difference, and the phase difference, respectively; is the observed energy of node i at the dominant frequency of oscillation, is the expected energy level of the source node, is the variance of the energy spectrum, is the variance of the response time difference, is the variance of the phase difference, are the times at which the energy spectrum between node i and the node j affected by it first exceeds a preset threshold E th at the target frequency, is the variance of the phase difference, is the phase difference between node i and the node j affected by it, is the expected value of the phase difference between node i and the node j affected by it.

6. The method for power grid oscillation source detection based on spatial Bayesian model according to claim 1, characterized in that, The function expression of the step S4 of performing Bayesian fusion calculation on the posterior probability distribution of each node as an oscillation source is: ； wherein, is the posterior probability that node i is the source of oscillation under given data, data is the node time-frequency features of each node in the power grid and the oscillation start time; is the prior probability that node i has the risk of oscillation, is the prior probability that node j has the risk of oscillation, is the comprehensive likelihood function of node i, is the comprehensive likelihood function of node j, and N is the number of nodes in the power grid.

7. The method of claim 1, wherein the method further comprises: After the step S4, the method further comprises visualizing the posterior probability distribution of each node as an oscillation source, and the visualization result comprises part or all of a node ranking table, a topological probability graph, a regional probability graph, and a multi-source identification graph, wherein the node ranking table comprises a ranking list of the posterior probability of each node i as an oscillation source, the topological probability graph is an image showing the posterior probability of each node i as an oscillation source in a power grid node topology graph, the regional probability graph is an image of the posterior probability of each node i as an oscillation source in a given region, and the multi-source identification graph is an image of one or more nodes i whose posterior probability as an oscillation source exceeds a preset threshold.

8. A power grid oscillation source detection system based on spatial Bayesian model, comprising a microprocessor and a memory connected to each other, characterized in that, The microprocessor is programmed or configured to perform the power grid oscillation source detection method based on the spatial Bayesian model according to any one of claims 1-7.

9. A computer-readable storage medium having stored therein a computer program or instructions, characterized in that, The computer program or instructions are programmed or configured to perform the power grid oscillation source detection method based on the spatial Bayesian model according to any one of claims 1-7 by the processor.

10. A computer program product comprising computer programs or instructions, characterized in that, The computer program or instructions are programmed or configured to perform the power grid oscillation source detection method based on the spatial Bayesian model according to any one of claims 1-7 by the processor.

Images