Power grid operation risk assessment method and system considering time segment coupling characteristics

Abstract

The invention discloses a power grid operation risk assessment method and system considering time period coupling characteristics, and relates to the technical field of power grid operation risk assessment, and the method comprises the following steps: obtaining operation data of a target power grid in a plurality of continuous time periods, and constructing a time-space state tensor based on the operation data; introducing a time period memory kernel function to obtain a cross-time period evolution memory curve of a node state; constructing a risk diffusion path diagram based on the evolutionary memory curve and the power grid power flow equation; establishing a critical energy accumulation function based on the risk diffusion path diagram, and triggering a node risk source mark when accumulation exceeds a preset threshold value; and generating a space-time risk evolution spectrum based on the risk source mark, and outputting a risk trajectory and a threshold breakthrough time point of each node in a continuous time period. According to the method, the cross-time evolution memory curve and the dynamic risk diffusion path diagram are constructed, so that high-risk nodes of continuous-time-period tail events are recognized in advance, and the problem of lack of cross-time-period accumulation risk quantification is solved.

Description

Technical Field

[0001] This invention relates to the field of power grid operation risk assessment technology, and more specifically, to a power grid operation risk assessment method and system that considers time-segment coupling characteristics. Background Technology

[0002] With the large-scale integration of new energy sources and increased load fluctuations, the risks in power grid operation exhibit a clear time-dependent coupling characteristic. Under conditions of sustained low output from new energy sources or continuous peak loads, the risks of the power grid system not only manifest in a single time period but also accumulate over multiple consecutive time periods, forming potential cumulative risks. Ultimately, this may lead to "threshold breaches" at critical nodes or lines, affecting the safe and stable operation of the system.

[0003] Existing power grid risk assessment methods mainly rely on single-period analysis, typically quantifying risks for instantaneous extreme events or short-term disturbances, such as voltage drop analysis, frequency stability assessment, or short-term load overload determination. These methods often ignore the cumulative effect of risks over continuous periods, lack cross-period conditional risk measurement, and are difficult to accurately identify high-risk nodes that gradually accumulate under continuous load or low renewable energy output scenarios.

[0004] Furthermore, existing methods do not adequately consider the spatiotemporal propagation characteristics, typically relying on single-node or global system indicators, and thus fail to characterize the dynamic transmission and cumulative effects of risks between nodes and across continuous time periods. Therefore, existing technologies suffer from problems such as untimely risk identification, insufficient early warning windows, and inaccurate node prioritization when facing the cumulative risks of tail-end events over continuous time periods, making it difficult to meet the needs of modern smart grids for cross-time period operational risk assessment.

[0005] The above-disclosed technical solutions have at least the following technical problems: under the condition of continuous low output of new energy or continuous peak load, the risks are superimposed in multiple time periods, which may eventually lead to "threshold breakthrough". Existing methods mostly focus on extreme scenarios in a single time period, and lack cross-time period cumulative risk indicators.

[0006] To address the above problems, this invention proposes a solution. Summary of the Invention

[0007] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method and system for assessing power grid operation risk based on time-segment coupling characteristics. By constructing cross-time period evolution memory curves and dynamic risk diffusion path diagrams, and establishing a critical energy accumulation function to generate a spatiotemporal risk evolution spectrum, the method enables cumulative risk analysis of tail events in continuous time periods and early identification of high-risk nodes. This addresses the problem in the prior art that it only focuses on extreme events in a single time period and lacks cross-time period cumulative risk quantification and spatiotemporal coupling analysis.

[0008] To achieve the above objectives, the present invention provides the following technical solution: On the one hand, the power grid operation risk assessment method based on time-segment coupling characteristics includes the following steps: acquiring the operation data of the target power grid in several consecutive time periods, and constructing a time-space state tensor based on the operation data and introducing a time period memory kernel function to obtain the cross-time period evolution memory curve of the node state; constructing a risk diffusion path diagram based on the evolution memory curve and the power flow equation; establishing a critical energy accumulation function based on the risk diffusion path diagram, gradually superimposing the tail events of the consecutive time periods, and triggering node risk source marking when the accumulation exceeds a preset threshold; generating a spatiotemporal risk evolution spectrum based on the risk source marking, and outputting the risk trajectory of each node in the consecutive time periods and the threshold breakthrough time point.

[0009] In a preferred embodiment, the step of acquiring the operating data of the target power grid over several consecutive time periods and constructing a time-space state tensor based on the operating data specifically involves: acquiring the operating data of each node of the power grid over consecutive time periods; organizing the node operating parameters over consecutive time periods according to the time order and spatial topology information to form a basic time-space state tensor; introducing the uncertainty disturbance component of the operating data into the basic time-space state tensor, and generating a multi-instance tensor set through Monte Carlo sampling.

[0010] In a preferred embodiment, the step of introducing a time-period memory kernel function to obtain the cross-time-period evolution memory curve of the node state specifically involves: inputting the multi-instance time-space state tensor of a continuous time period into the memory kernel function; the memory kernel function includes exponential decay type, long-tail type, or event-weighted type, used to characterize the decay of short-term disturbances, the continuation of tail events, and the differentiated effects of different event categories, respectively; based on the weighting result of the memory kernel function, the cross-time-period evolution memory value of the node in the current time period is calculated; the cross-time-period evolution memory value is calculated one by one on the multi-instance tensor set to generate a set of multi-instance cross-time-period evolution memory curves of the node state.

[0011] In a preferred embodiment, the construction of the risk diffusion path diagram based on the evolutionary memory curve and the power flow equation specifically involves: defining a propagation weight function based on the cross-time-period evolutionary memory curve and the power flow sensitivity matrix of the nodes; dynamically correcting the propagation weight function using a gradient differentiable update method, wherein the correction process aims at the system operating risk function and iteratively updates it by combining the memory curve modulation factor and the power flow sensitivity partial derivative; introducing a state-related diffusion adjustment factor during the correction process, automatically expanding the propagation range when the power grid operating stress increases and automatically shrinking the propagation range when the operating risk decreases; and generating a dynamic risk diffusion path diagram based on the corrected propagation weight function.

[0012] In a preferred embodiment, generating a dynamic risk diffusion path map based on the modified propagation weight function specifically involves: assembling the modified propagation weight function into a spatiotemporal propagation weight matrix according to the power grid topology; starting from a node marked as a risk source, calculating its risk diffusion probability on adjacent nodes based on the propagation weight matrix to generate an initial risk diffusion probability distribution; filtering an initial risk diffusion path set based on the initial risk diffusion probability distribution; iterating the propagation weight matrix over continuous time periods, starting from the initial risk diffusion path set, calculating the diffusion amplitude and direction of risk at different time steps to obtain a cross-time period risk propagation trajectory; introducing a risk threshold monitoring mechanism during the iteration process, merging the risk propagation trajectories after iteration in each time period into the path set to generate a unified spatiotemporal risk diffusion path map.

[0013] In a preferred embodiment, the step of establishing a critical energy accumulation function based on the risk diffusion path graph and progressively superimposing events at the tail end of a continuous time period specifically involves: mapping the risk propagation intensity of each node in the risk diffusion path graph to an equivalent energy component, wherein the equivalent energy component reflects the risk contribution of the node at the end of the time period; progressively superimposing the equivalent energy components over a continuous time period to construct a critical energy accumulation function, which is used to characterize the cumulative effect of risk across time periods; and introducing a dynamic threshold determination mechanism during the accumulation process, triggering node risk source marking when the critical energy accumulation value exceeds a preset threshold.

[0014] In a preferred embodiment, the step of generating a spatiotemporal risk evolution spectrum based on risk source marking and outputting the risk trajectory and threshold breakthrough point of each node in a continuous time period specifically involves: filtering out a subset of risk trajectories corresponding to nodes that have been marked as risk sources based on existing node risk trajectories in the dynamic risk diffusion path diagram; monitoring the critical energy accumulation threshold of the filtered node risk trajectories and recording the time point when each node risk trajectory first exceeds the threshold, marking it as a cross-time period threshold breakthrough point; integrating the node risk trajectories and the corresponding cross-time period threshold breakthrough points to generate a spatiotemporal risk evolution spectrum; the evolution spectrum includes: node risk intensity curves and cross-time period threshold breakthrough points; and extracting the time window and spatial range of risk warning based on the spatiotemporal risk evolution spectrum.

[0015] In a preferred embodiment, the step of outputting the risk trajectory and threshold breakthrough time of each node within a continuous time period further includes: identifying nodes with high risk levels or threshold breakthroughs within a continuous time period based on the node risk intensity curve and cross-time period threshold breakthrough points in the spatiotemporal risk evolution spectrum; prioritizing the identified high-risk nodes according to their risk intensity, threshold breakthrough time, and risk accumulation trend to form a node risk level sequence; generating maintenance optimization suggestions for high-risk nodes based on the node risk level sequence, combined with the node's location in the power grid topology, maintenance resource constraints, and historical maintenance records; and outputting the maintenance optimization suggestions, including maintenance priority order, expected risk mitigation effect, and recommended time window.

[0016] On the other hand, the power grid operation risk assessment system that considers the coupling characteristics of time periods includes the following modules: a state tensor construction module: used to acquire the operation data of the target power grid in several consecutive time periods and construct a time-space state tensor based on the operation data; an evolutionary memory extraction module: used to introduce a time period memory kernel function to obtain the cross-time period evolutionary memory curve of the node state; a risk diffusion path generation module: used to construct a risk diffusion path diagram based on the evolutionary memory curve and the power flow equation; a risk source marking module: used to establish a critical energy accumulation function based on the risk diffusion path diagram, gradually superimpose the tail events of the continuous time periods, and trigger the node risk source marking when the accumulation exceeds a preset threshold; and a spatiotemporal risk evolution spectrum generation module: used to generate a spatiotemporal risk evolution spectrum based on the risk source marking, and output the risk trajectory of each node in the continuous time period and the threshold breakthrough time point.

[0017] The technical effects and advantages of the power grid operation risk assessment method and system based on time-segment coupling characteristics of this invention are as follows: 1. This invention constructs a cross-time-period evolution memory curve and combines it with a power flow sensitivity matrix to form a dynamic risk diffusion path diagram, thereby achieving time-space coupled analysis of power grid node risks. This method can cumulatively model the evolution of node states over continuous time periods, while considering power flow constraints and topological relationships between nodes, enabling dynamic transmission and adaptive adjustment of risks in space and time. This allows for the identification of key nodes that may become risk sources, especially in risk accumulation scenarios under conditions of continuous low output from new energy sources or continuous peak loads.

[0018] 2. This invention quantifies the cumulative effect of cross-period tail events and marks node risks by establishing a critical energy accumulation function and generating a spatiotemporal risk evolution spectrum. The method can monitor whether the accumulated risk over a continuous period exceeds a preset threshold, and mark high-risk nodes when the threshold is exceeded. It also provides risk intensity curves and cross-period threshold breakthrough points, providing a basis for the maintenance optimization and operation and maintenance strategy prioritization of high-risk nodes, thereby supporting early warning and risk prevention and control measures. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the power grid operation risk assessment method based on time-segment coupling characteristics of the present invention. Figure 2 This is a schematic diagram of the power grid operation risk assessment system based on time-segment coupling characteristics according to the present invention; Figure 3 A schematic diagram of the memory curve of the cross-period evolution of node risk; Figure 4 For each node, there is a risk intensity curve; Figure 5 A heatmap of risk evolution. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1, Figure 1 The present invention provides a power grid operation risk assessment method based on time-segment coupling characteristics, comprising the following steps: S1, acquire the operating data of the target power grid in several consecutive time periods, and construct a time-space state tensor based on the operating data; The time-space state tensor is a multidimensional data structure used to characterize the operating state of each node in a target power grid over a continuous period of time. It can simultaneously represent the continuous evolution characteristics of node states in the time dimension and the interrelationship of nodes in the power grid topology in the spatial dimension.

[0022] Specifically, the time-space state tensor includes the following characteristics: Time dimension: One dimension of the tensor corresponds to several consecutive time periods (such as minute level, hour level or scheduling cycle), which is used to record the operating parameters of each node at different time points, such as voltage amplitude, power flow, energy storage status and new energy output, and can reflect the evolution trend of node status over time.

[0023] Spatial dimension: Another dimension of the tensor corresponds to the power grid nodes and their topological relationships. It is used to reflect the power transmission, power flow constraints and risk propagation channels between nodes, and to realize the spatial characterization of the overall power grid structure.

[0024] Multi-instance feature: When considering uncertainties (such as new energy fluctuations and load forecasting errors), a multi-instance tensor set can be formed by introducing a disturbance component into the tensor and performing multiple samplings. This set can be used to describe the possible evolution of node states under different disturbance scenarios.

[0025] In this embodiment, the step of acquiring the operating data of the target power grid over several consecutive time periods and constructing a time-space state tensor based on the operating data specifically involves: The operation data of each node of the power grid is acquired over a continuous period of time. The operation data includes node voltage, branch power flow, new energy output and energy storage status. The node operation parameters within a continuous time period are organized according to the time sequence and spatial topology information to form a basic time-space state tensor; Uncertainty perturbation components of the running data are introduced into the basic time-space state tensor, and a multi-instance tensor set is generated through Monte Carlo sampling.

[0026] S2, introduce the time period memory kernel function to obtain the cross-time period evolution memory curve of the node state; A time-period memory kernel function is a function structure used to characterize the decay-cumulative effect of the state of a power grid node across different time periods. It superimposes the state information of past time periods onto the current time period after decaying or amplifying it according to a certain rule.

[0027] In this embodiment, the introduction of a time-period memory kernel function to obtain the cross-time-period evolution memory curve of the node state is specifically as follows: The multi-instance time-space state tensor of continuous time periods is input into the memory kernel function, which is used to attenuate or amplify the node state of historical time periods, and is used to simulate the temporal attenuation effect or cumulative amplification effect of risk energy respectively. The memory kernel functions include exponential decay type, long tail type and event weighted type, which are used to characterize the decay of short-term disturbances, the continuation of tail events and the differentiated effects of different event categories, respectively. Based on the weighted results of the memory kernel function, the cross-time evolution memory value of the node in the current time period is calculated; The cross-time period evolution memory value is calculated one by one on the multi-instance tensor set to generate a set of multi-instance cross-time period evolution memory curves of node states.

[0028] Figure 3 A schematic diagram of the cross-period evolution memory curve of node risk is given to intuitively present the decay / cumulative effect of node risk state at different time periods. Through three typical "memory kernel function action schematic curves", the influence of different types of memory mechanisms on risk energy and the evolution law of node state over time are shown. Exponential decay type (blue curve): This reflects the exponential decay characteristic of risk memory. As time steps (horizontal axis) advance, the cross-period evolutionary memory value (vertical axis) rises rapidly and then gradually falls back, eventually stabilizing with fluctuations. This indicates that risk energy decays rapidly over time, retaining only short-term memory effects, and the long-term impact of past risks on the current state is weak.

[0029] Long-tail type (orange curve): The curve shows a continuous upward or high-level fluctuation trend, indicating that the risk energy gradually accumulates and decays slowly over time. The risks of the past period have a significant long-term impact on the current state, and the risks are prone to form a "long-tail" type of continuous memory.

[0030] Event-weighted type (green curve): This shows the characteristic of risk memory being accumulated by key events / periods. The curve grows slowly in the early stage and accelerates in the later stage, reflecting that as node risks evolve, due to the "weighting effect" of specific events or stages, the risk energy is continuously superimposed, and the memory effect is gradually strengthened over time, with the cumulative effect being particularly prominent in the later stage.

[0031] The cross-time-period evolutionary memory value is specifically as follows:

[0032] in, This represents the cross-time period evolution memory value of node i at time t. The total number of consecutive time periods. For memory kernel function, For node i in the historical time period The state vector.

[0033] S3, based on evolutionary memory curves and power grid flow equations, constructs a risk diffusion path diagram; In this embodiment, the construction of the risk diffusion path diagram based on the evolutionary memory curve and the power grid flow equation specifically involves: Based on the node-based cross-period evolution memory curve and power flow sensitivity matrix, a propagation weighting function is defined. The propagation weight function is dynamically corrected by a gradient differentiable update method. The correction process is aimed at the system operation risk function and iteratively updated by combining the memory curve modulation factor and the power flow sensitivity partial derivative. During the correction process, a state-dependent diffusion adjustment factor is introduced, which automatically expands the propagation range when the grid operating stress increases and automatically shrinks the propagation range when the operating risk decreases. A dynamic risk diffusion path diagram is generated based on the modified propagation weight function to characterize the spatiotemporal transmission relationship of risk in the power grid topology.

[0034] The dynamic correction of the propagation weight function is specifically as follows:

[0035]

[0036]

[0037] in, For the corrected propagation weights, Let be the propagation weight from node i to node j at time t. To update the step size for gradient updates, This is a pre-defined spatiotemporal loss function constructed based on risk energy distribution. These are the corresponding elements of the power flow sensitivity matrix. As a diffusion regulator, The time modulation function derived for the memory curve. The preset weight adjustment factor, For node time sequence dependency, , These are the cross-time evolutionary memory curves for nodes i and j, respectively.

[0038] In this embodiment, the generation of a dynamic risk diffusion path graph based on the modified propagation weight function specifically includes: The modified propagation weight function is assembled into a spatiotemporal propagation weight matrix according to the power grid topology, where the matrix elements reflect the risk transmission intensity between nodes. Starting with the node marked as a risk source, calculate its risk diffusion probability on neighboring nodes based on the propagation weight matrix to generate an initial risk diffusion probability distribution; Based on the initial risk diffusion probability distribution, a set of initial risk diffusion paths is selected; The propagation weight matrix is ​​iteratively propagated over continuous time periods. Starting from the initial set of risk propagation paths, the magnitude and direction of risk propagation at different time steps are calculated to obtain the risk propagation trajectory across time periods. A risk threshold monitoring mechanism is introduced during the iteration process. When the propagation intensity exceeds the set threshold, the path branches are expanded. When the propagation intensity is lower than the shrinkage threshold, invalid paths are pruned, thereby ensuring that the path graph can adaptively shrink or expand. The risk propagation trajectories after each time period iteration are merged into the path set to ultimately generate a unified spatiotemporal risk diffusion path map.

[0039] The intensity of risk transmission is specifically as follows:

[0040] The risk diffusion probability is specifically:

[0041] in, For the intensity of risk transmission, For the corrected propagation weights, This represents the state offset of node i at time t (voltage deviation, power imbalance, or risk energy residual). This represents the probability of risk diffusion.

[0042] S4. Based on the risk diffusion path diagram, a critical energy accumulation function is established to gradually superimpose tail events in continuous time periods, and trigger node risk source marking when the accumulation exceeds a preset threshold. In this embodiment, the step of establishing a critical energy accumulation function based on the risk diffusion path diagram and gradually superimposing tail events over a continuous period is specifically as follows: The risk propagation intensity of each node in the risk diffusion path diagram is mapped to an equivalent energy component, which reflects the risk contribution of the node at the end of the time period. The equivalent energy components are gradually superimposed over a continuous time period to construct a critical energy accumulation function, which is used to characterize the cumulative effect of risk across time periods. A dynamic threshold determination mechanism is introduced during the accumulation process. When the critical energy accumulation value exceeds the preset threshold, the node risk source is marked to achieve early warning of cross-time period risks.

[0043] The critical energy accumulation function is specifically:

[0044]

[0045] in, For node i in time period The critical energy accumulation value, The preset memory modulation coefficient, To accumulate window length, Let be the equivalent energy component of node i in time period t. The mapping coefficient from risk intensity to energy. The duration of the period. Let represent the risk transmission intensity of node i during time period t.

[0046] S5 generates a spatiotemporal risk evolution spectrum based on risk source marking, and outputs the risk trajectory and threshold breakthrough time of each node in a continuous period.

[0047] In this embodiment, the step of generating a spatiotemporal risk evolution spectrum based on risk source markers and outputting the risk trajectory and threshold breakthrough point of each node within a continuous time period specifically involves: Based on the existing node risk trajectories in the dynamic risk diffusion path diagram, a subset of risk trajectories corresponding to nodes that have been marked as risk sources is selected. The critical energy accumulation threshold of the screened node risk trajectory is monitored, and the time point when the risk trajectory of each node first exceeds the threshold is recorded and marked as the cross-time threshold breakthrough point. By integrating the node risk trajectory and the corresponding cross-time threshold breakthrough point, a spatiotemporal risk evolution spectrum is generated; The evolutionary spectrum includes: Node risk intensity curve: depicts the change in risk level of each node over a continuous period of time; such as Figure 4 The diagram shows the risk intensity curves for each node. The horizontal axis represents time period, and the vertical axis represents risk intensity. Different curves correspond to different nodes and their respective memory mechanisms. Dashed lines represent the risk thresholds set for each node, and red-marked points indicate the moments when the risk intensity exceeds the threshold. This diagram can intuitively reflect the risk evolution trend and exceedance events at different nodes, providing a basis for the generation of subsequent risk warning and control strategies. Cross-time threshold breach point: The moment when the risk of a node first exceeds the critical threshold. Based on the spatiotemporal risk evolution spectrum, the time window and spatial range of risk warning are extracted to provide a reference for risk prevention and control across time periods and multiple nodes.

[0048] The output of the risk trajectory and threshold breakthrough time points of each node within a continuous time period also includes: By utilizing the node risk intensity curves and cross-period threshold breakthrough points in the spatiotemporal risk evolution spectrum, nodes with high risk levels or threshold breakthroughs within continuous time periods can be identified. Based on the risk intensity, threshold breach time, and risk accumulation trend of the nodes, the identified high-risk nodes are prioritized to form a node risk level sequence. Based on the node risk level sequence, combined with the node's location in the power grid topology, operation and maintenance resource constraints, and historical maintenance records, maintenance optimization suggestions are generated for high-risk nodes. The maintenance optimization suggestions are output, including maintenance priority, expected risk mitigation effect and recommended time window, to guide the adjustment of operation and maintenance strategies and risk prevention and control.

[0049] Example 2, Figure 2 The present invention provides a power grid operation risk assessment system that considers time-segment coupling characteristics, comprising the following modules: State Tensor Construction Module: Used to obtain the operating data of the target power grid over several consecutive time periods and construct a time-space state tensor based on the operating data; Evolutionary memory extraction module: used to introduce time-period memory kernel function to obtain cross-time-period evolutionary memory curve of node state; Risk diffusion path generation module: used to construct risk diffusion path diagrams based on evolutionary memory curves and power grid flow equations; Risk source marking module: It is used to establish a critical energy accumulation function based on the risk diffusion path map, gradually superimpose the tail events of continuous time periods, and trigger node risk source marking when the accumulation exceeds a preset threshold; Spatiotemporal risk evolution spectrum generation module: used to generate a spatiotemporal risk evolution spectrum based on risk source markers, and output the risk trajectory and threshold breakthrough time of each node in a continuous time period.

[0050] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0051] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0052] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0053] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0054] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0055] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for assessing power grid operation risk based on time-segment coupling characteristics, characterized in that, Includes the following steps: Obtain operational data of the target power grid over several consecutive time periods, and construct a time-space state tensor based on the operational data; By introducing a time-period memory kernel function, the cross-time-period evolution memory curve of the node state is obtained; Based on evolutionary memory curves and power grid flow equations, a risk diffusion path diagram is constructed; Based on the risk diffusion path diagram, a critical energy accumulation function is established to gradually superimpose tail events in continuous time periods, and trigger node risk source marking when the accumulation exceeds a preset threshold. Based on risk source labeling, a spatiotemporal risk evolution spectrum is generated, and the risk trajectory and threshold breakthrough time of each node in a continuous time period are output.

2. The power grid operation risk assessment method based on time-segment coupling characteristics according to claim 1, characterized in that, The process of acquiring operational data of the target power grid over several consecutive time periods and constructing a time-space state tensor based on the operational data specifically involves: Obtain the operational data of each node in the power grid over a continuous period of time; The node operation parameters within a continuous time period are organized according to the time sequence and spatial topology information to form a basic time-space state tensor; Uncertainty perturbation components of the running data are introduced into the basic time-space state tensor, and a multi-instance tensor set is generated through Monte Carlo sampling.

3. The power grid operation risk assessment method based on time-segment coupling characteristics according to claim 2, characterized in that, The introduction of a time-period memory kernel function yields the cross-time-period evolution memory curve of the node state, specifically as follows: Input the multi-instance time-space state tensor of continuous time intervals into the memory kernel function; The memory kernel functions include exponential decay type, long tail type and event weighted type, which are used to characterize the decay of short-term disturbances, the continuation of tail events and the differentiated effects of different event categories, respectively. Based on the weighted results of the memory kernel function, the cross-time evolution memory value of the node in the current time period is calculated; The cross-time period evolution memory value is calculated one by one on the multi-instance tensor set to generate a set of multi-instance cross-time period evolution memory curves of node states.

4. The power grid operation risk assessment method based on time-segment coupling characteristics according to claim 3, characterized in that, The risk diffusion path diagram constructed based on evolutionary memory curves and power grid flow equations is as follows: Based on the node-based cross-period evolution memory curve and power flow sensitivity matrix, a propagation weighting function is defined. The propagation weight function is dynamically corrected by a gradient differentiable update method. The correction process is aimed at the system operation risk function and iteratively updated by combining the memory curve modulation factor and the power flow sensitivity partial derivative. During the correction process, a state-dependent diffusion adjustment factor is introduced, which automatically expands the propagation range when the grid operating stress increases and automatically shrinks the propagation range when the operating risk decreases. A dynamic risk diffusion path diagram is generated based on the modified propagation weight function.

5. The power grid operation risk assessment method based on time-segment coupling characteristics according to claim 4, characterized in that, The generation of the dynamic risk diffusion path graph based on the modified propagation weight function is specifically as follows: The modified propagation weight function is assembled into a spatiotemporal propagation weight matrix according to the power grid topology; Starting with the node marked as a risk source, calculate its risk diffusion probability on neighboring nodes based on the propagation weight matrix to generate an initial risk diffusion probability distribution; Based on the initial risk diffusion probability distribution, a set of initial risk diffusion paths is selected; The propagation weight matrix is ​​iteratively propagated over continuous time periods. Starting from the initial set of risk propagation paths, the magnitude and direction of risk propagation at different time steps are calculated to obtain the risk propagation trajectory across time periods. A risk threshold monitoring mechanism is introduced during the iteration process to merge the risk propagation trajectories after each iteration in the path set and generate a unified spatiotemporal risk diffusion path map.

6. The power grid operation risk assessment method based on time-segment coupling characteristics according to claim 5, characterized in that, The critical energy accumulation function, based on the risk diffusion path diagram, is established to progressively superimpose tail events over a continuous time period. Specifically: The risk propagation intensity of each node in the risk diffusion path diagram is mapped to an equivalent energy component, which reflects the risk contribution of the node at the end of the time period. The equivalent energy components are gradually superimposed over a continuous time period to construct a critical energy accumulation function, which is used to characterize the cumulative effect of risk across time periods. A dynamic threshold determination mechanism is introduced during the accumulation process. When the critical energy accumulation value exceeds the preset threshold, the node risk source is marked.

7. The power grid operation risk assessment method based on time-segment coupling characteristics according to claim 6, characterized in that, The process of generating a spatiotemporal risk evolution spectrum based on risk source markers, and outputting the risk trajectory and threshold breakthrough time points of each node within a continuous time period, specifically involves: Based on the existing node risk trajectories in the dynamic risk diffusion path diagram, a subset of risk trajectories corresponding to nodes that have been marked as risk sources is selected. The critical energy accumulation threshold of the screened node risk trajectory is monitored, and the time point when the risk trajectory of each node first exceeds the threshold is recorded and marked as the cross-time threshold breakthrough point. By integrating the node risk trajectory and the corresponding cross-time threshold breakthrough point, a spatiotemporal risk evolution spectrum is generated; The evolution spectrum includes: node risk intensity curves and cross-time threshold breakthrough points; The time window and spatial range for risk warning are extracted based on the spatiotemporal risk evolution spectrum.

8. The power grid operation risk assessment method based on time-segment coupling characteristics according to claim 7, characterized in that, The output of the risk trajectory and threshold breakthrough time points of each node within a continuous time period also includes: Based on the node risk intensity curves and cross-time threshold breakthrough points in the spatiotemporal risk evolution spectrum, nodes with high risk levels or threshold breakthroughs within continuous time periods are identified. Based on the risk intensity, threshold breach time, and risk accumulation trend of the nodes, the identified high-risk nodes are prioritized to form a node risk level sequence. Based on the node risk level sequence, combined with the node's location in the power grid topology, operation and maintenance resource constraints, and historical maintenance records, maintenance optimization suggestions are generated for high-risk nodes. Output the maintenance optimization suggestions, including maintenance priority, expected risk mitigation effect, and recommended time window.

9. A system for assessing power grid operation risk using the time-segment coupling characteristics as described in any one of claims 1-8, characterized in that, Includes the following modules: State Tensor Construction Module: Used to obtain the operating data of the target power grid over several consecutive time periods and construct a time-space state tensor based on the operating data; Evolutionary memory extraction module: used to introduce time-period memory kernel function to obtain cross-time-period evolutionary memory curve of node state; Risk diffusion path generation module: used to construct risk diffusion path diagrams based on evolutionary memory curves and power grid flow equations; Risk source marking module: It is used to establish a critical energy accumulation function based on the risk diffusion path map, gradually superimpose the tail events of continuous time periods, and trigger node risk source marking when the accumulation exceeds a preset threshold; Spatiotemporal risk evolution spectrum generation module: used to generate a spatiotemporal risk evolution spectrum based on risk source markers, and output the risk trajectory and threshold breakthrough time of each node in a continuous time period.

Images