Intelligent early warning methods, devices and equipment for cross-regional electricity spot market purchase risks

Abstract

This invention discloses an intelligent early warning method, device, and equipment for cross-regional electricity spot market purchase risks, relating to the field of electricity market and power grid operation safety technology. The method includes the following steps: extracting risk indicator data affecting inter-provincial spot electricity purchases to form a risk indicator dataset; constructing a dynamic weighted risk transmission network based on the risk indicator dataset; inputting any risk indicator from the risk indicator dataset into the risk transmission network for simulation and prediction of the probability distribution of risk intensity at each node in the future; and outputting strategic early warning instructions based on the probability distribution of risk intensity through an optimized decision-making model. This application overcomes the shortcomings of existing technologies, such as difficulty in quantifying cross-regional risk transmission, delayed early warning, and lack of enforceability, achieving accurate prediction, advanced perception, and proactive defense against inter-provincial market risks.

Description

Technical Field

[0001] This invention relates to the field of electricity market and power grid operation safety technology, and more specifically, to a method, device and equipment for intelligent early warning of electricity purchase risks in cross-regional electricity spot markets. Background Technology

[0002] With the advancement of the global energy transition, especially against the backdrop of a gradually increasing proportion of renewable energy, the trading models of the electricity market and the operation of the power system are undergoing profound changes. The electricity spot market, as a trading mechanism based on real-time supply and demand balance, is gradually becoming an important component of national electricity markets. In inter-regional electricity markets, electricity flow is not only affected by the operation of the local power grid, but also closely related to a variety of factors such as long-distance transmission networks, market demand fluctuations, renewable energy generation fluctuations, and weather. These factors make the risks of electricity purchase in the spot market more complex, especially given the intertwined risk indicator data involved in inter-regional electricity purchases, rendering traditional risk assessment and early warning methods inadequate.

[0003] Currently, existing technologies typically use static or linear models to describe the propagation path of risks in cross-regional electricity markets. These models fail to take into account the complex nonlinear coupling relationships between risk indicators in the power system, as well as the dynamic and time-varying risk transmission mechanisms between different regions, leading to biases in comprehensive risk assessments. Summary of the Invention

[0004] To overcome the aforementioned deficiencies in the prior art, embodiments of the present invention provide a method, apparatus, and equipment for intelligent early warning of electricity purchase risks in cross-regional electricity spot markets. By constructing a dynamic weighted risk transmission network, risk indicator data is extracted in real time, and the future risk intensity distribution is simulated and deduced. Furthermore, strategic early warning instructions are generated by optimizing the decision-making model. This solves the problem of deviations in comprehensive risk assessment caused by insufficient extraction of risk indicator data and a single risk transmission path in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] The intelligent early warning method for cross-regional electricity spot market purchase risks includes the following steps: extracting risk indicator data affecting inter-provincial spot electricity purchases to form a risk indicator dataset; constructing a dynamic weighted risk transmission network based on the risk indicator dataset; inputting any risk indicator from the risk indicator dataset into the risk transmission network for simulation and prediction of the probability distribution of risk intensity at each node in the future; and outputting strategic early warning instructions based on the probability distribution of risk intensity by optimizing the decision model.

[0007] In a preferred embodiment, the step of extracting risk indicator data affecting inter-provincial spot electricity purchases to form a risk indicator dataset specifically involves: acquiring transaction data, power grid operation data, external environment data, and risk indicator data to construct a multi-source database; calculating the transmission congestion margin and power deviation of each inter-provincial interconnection line based on the power grid topology and power flow data in the power grid operation data; calculating the congestion surplus index of each inter-provincial interconnection line based on the transmission congestion margin and power deviation; dynamically determining the overall operating status of the inter-provincial market based on the congestion surplus index; and extracting corresponding risk indicator data from the multi-source fusion database according to the operating status and organizing it by time series to form a risk indicator dataset.

[0008] In a preferred embodiment, the construction of a dynamically weighted risk transmission network based on a risk index dataset specifically involves: dividing the power system into multiple regions, each region being defined as a network node; defining connection edges between nodes based on the grid topology and transaction coupling relationships between regions; calculating the real-time risk intensity value of each network node based on the risk index dataset; dynamically calculating the transmission resistance of each connection edge using a risk transmission barrier function, and mapping this transmission resistance to the dynamic weight of the connection edge; the barrier function characterizing the ease with which risk is transmitted across the connection edge; and generating a dynamically weighted risk transmission network based on the network nodes, the real-time risk intensity values ​​of each network node, the connection edges, and the dynamic weights of the connection edges.

[0009] In a preferred embodiment, the calculation steps for the transmission resistance are as follows: based on the day-ahead clearing results and real-time operating status of the spot market, identify the main power transmission paths and directions in the dynamic weighted risk transmission network; for the connecting edges on the main power transmission paths, distinguish between the forward and reverse directions of power transmission, and calculate the asymmetric dynamic risk transmission barrier values ​​respectively; for the forward barrier value consistent with the clearing power direction, it is determined based on the transmission margin of the physical channel; for the reverse barrier value opposite to the clearing power direction, it is calculated by introducing a penalty coefficient based on the transmission margin; and the forward and reverse barrier values ​​are weighted and combined to generate the transmission resistance.

[0010] In a preferred embodiment, the step of inputting any risk indicator from the risk indicator dataset into the risk transmission network for simulation and prediction of the future risk intensity probability distribution of each node specifically involves: constructing a time-varying system matrix based on the weight values ​​of each connecting edge in the dynamic risk transmission network; performing singular value decomposition on the time-varying system matrix to obtain multiple transmission modes and their corresponding mode weights; calculating the activation factor of each transmission mode based on real-time system state data from multi-source data; quantifying the risk indicator into initial risk energy and injecting it into the source nodes directly affected by it, and performing several rounds of Monte Carlo simulation; and statistically analyzing the simulation results of all rounds to generate the future risk intensity probability distribution of each node.

[0011] In a preferred embodiment, the Monte Carlo simulation method specifically comprises: calculating the transfer probability of risk propagation from the current node to its neighboring nodes based on dynamic weights; adjusting the transfer probability according to the pattern weights and corresponding activation factors; determining the transmission path and direction of risk through random sampling based on the adjusted transfer probability; and dissipating the initial risk energy according to the dynamic risk transmission barrier function during the transmission process until the simulation time reaches a preset future period or the risk energy is lower than a preset energy threshold, thus completing one round of simulation.

[0012] In a preferred embodiment, the step of outputting strategic early warning instructions based on the probability distribution of risk intensity through an optimized decision model specifically involves: calculating the risk debt equivalent based on the probability distribution of risk intensity of each node and the real-time available transmission capacity margin, and determining the set of nodes requiring mandatory intervention; calculating the physical sensitivity weight of the nodes requiring mandatory intervention based on their physical operational constraints; constructing and solving an equilibrium model using the risk debt equivalent as the commodity quantity, and using the physical sensitivity weight and the real-time electricity price difference between nodes as market parameters to obtain the equilibrium price; and generating executable strategic early warning instructions based on the equilibrium price.

[0013] In a preferred embodiment, the step of generating executable strategic early warning instructions based on the equilibrium price specifically involves: comparing the equilibrium price with the real-time marginal electricity price of each node to calculate a risk hedging coefficient; combining the risk hedging coefficient with a physical sensitivity weight to generate a risk adjustment priority score; sorting the nodes requiring intervention according to the priority score to determine a control priority sequence; allocating control tasks to each node based on the risk debt equivalent as the total control amount and according to the priority score ratio of the nodes in the control priority sequence; and encapsulating the control tasks of each node into executable strategic early warning instructions.

[0014] A smart early warning device for cross-regional electricity spot market purchase risks includes: a data extraction module for extracting risk indicator data affecting inter-provincial spot electricity purchases to form a risk indicator dataset; a network construction module for constructing a dynamically weighted risk transmission network based on the risk indicator dataset; a risk simulation module for inputting any risk indicator from the risk indicator dataset into the risk transmission network for simulation and prediction of the probability distribution of risk intensity at each node in the future; and an optimization decision module for outputting strategic early warning instructions based on the probability distribution of risk intensity through an optimized decision model.

[0015] A smart early warning device for electricity purchase risks in a cross-regional electricity spot market includes: a memory and a processor; the memory is used to store programs; the processor is used to execute programs to implement the smart early warning method for electricity purchase risks in a cross-regional electricity spot market.

[0016] This invention constructs a dynamic weighted risk transmission network, abstracting the complex interactions between power grids in different regions into a unified network. This method can simultaneously simulate the dynamic process of risk transmission through multiple paths, breaking through the simplistic assumptions of traditional methods that treat risk as an isolated event or transmits only along a single physical path. This allows for a more realistic and accurate depiction of the actual propagation behavior of risk in cross-regional markets. Furthermore, simulations on the dynamic weighted risk transmission network can obtain the probability distribution of risk intensity for each node in future time periods. This transforms the early warning result from a simple binary "yes" or "no" judgment, providing the probability of risk occurrence and its statistical distribution. This probabilistic output provides decision-makers with quantitative information about the degree of uncertainty of future risks, supporting more scientific and precise differentiated prevention and control decisions, thereby significantly improving the accuracy and real-time performance of risk prediction. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the intelligent early warning method for cross-regional electricity spot market purchase risks according to the present invention.

[0018] Figure 2 This is a schematic diagram of the intelligent early warning device for cross-regional electricity spot market purchase risks according to the present invention. Detailed Implementation

[0019] 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.

[0020] Example 1, Figure 1This invention presents an intelligent early warning method for electricity purchase risks in cross-regional electricity spot markets, comprising the following steps:

[0021] S1, extract risk indicator data affecting inter-provincial spot electricity purchases to form a risk indicator dataset;

[0022] In this embodiment, risk indicator data affecting inter-provincial spot electricity purchases are extracted to form a risk indicator dataset, specifically:

[0023] By acquiring transaction data, power grid operation data, external environment data, and risk indicator data using existing technologies, the collected multi-source data is preprocessed, including data cleaning, data standardization, and data fusion, to form a multi-source fusion database.

[0024] Based on the power grid topology and power flow data in the multi-source fusion database, the transmission congestion margin and power deviation of each inter-provincial tie line are calculated, specifically:

[0025]

[0026]

[0027] in, This represents the blocking margin of the connection line i→j. This indicates the rated transmission capacity of the connecting line. This indicates the actual power flow of the tie line. For power deviation, For the scheduled power.

[0028] Calculate the congestion surplus index based on transmission congestion margin and power deviation. Specifically:

[0029]

[0030] in, To prevent small constants from being divided by zero, This is a weighting coefficient used to balance the contribution of margin and bias to risk.

[0031] The blocking surplus index is compared with the preset blocking surplus index threshold range to dynamically determine the overall operating status of the inter-provincial market. If the blocking surplus index is greater than the preset blocking surplus index threshold range, it is a high-risk state; if the blocking surplus index is within the preset blocking surplus index threshold range, it is a tense state; and if the blocking surplus index is less than the preset blocking surplus index threshold range, it is a safe state.

[0032] Based on the identified market operating state, different feature extraction algorithms are selected to extract corresponding risk indicator data from a multi-source fusion database. For example, in a safe state, risk indicators focus on long-term and macroeconomic factors, using the Pearson correlation coefficient method to extract the top N risk indicators with the highest correlation to electricity price volatility from the database. In a tense state, risk indicators focus on real-time balance factors, using the mutual information method to extract the top M risk indicators with the largest mutual information content with real-time electricity prices from the database. In a high-risk state, risk indicators focus on physical security factors, directly extracting the most critical factors according to preset rules in the expert experience base. All extracted risk indicators are organized by time series and standardized again to form the risk indicator dataset.

[0033] S2, based on the risk indicator dataset, constructs a dynamic weighted risk transmission network;

[0034] In this embodiment, a dynamic weighted risk transmission network is constructed based on the risk indicator dataset, specifically as follows:

[0035] The power system is divided into multiple regions, each defined as a network node. The connection edges between nodes are defined based on the coupling relationship between the power grid topology and inter-provincial transactions. For regional nodes that do not have direct transactions or grid connections, they are considered as isolated nodes or connected through indirect paths.

[0036] For each network node Based on the relevant line congestion index and the node's own risk indicators, the real-time risk intensity value is calculated. :

[0037]

[0038] in, Let be the set of neighboring nodes of node i. The risk propagation weight between node i and its neighbor node j. Number the risk indicators. The weight of the k-th risk indicator. Let be the k-th risk indicator for node i.

[0039] For each connecting edge Construct a dynamic risk transmission barrier function Physical blocking item Economic driving factors and direction penalty items The transmission resistance is dynamically calculated using a three-part weighted average. The potential barrier function is used to characterize the ease with which risk propagates across the connecting edge, and the specific formula is as follows:

[0040]

[0041]

[0042]

[0043]

[0044] in, , Let be the actual marginal prices of nodes i and j, respectively. This represents the average real-time marginal electricity price at the node. For symbolic functions, This indicates the actual power flow of the tie line. The power was cleared for the connecting line ij a day ago. As a penalty coefficient (a constant K>1), when the actual power flow direction is opposite to the day-ahead clearing power direction, a fixed high penalty value is applied, significantly increasing the potential barrier. For the weighting coefficients, satisfying .

[0045] The transmission resistance calculated from the risk transmission barrier function The mapping is converted into the dynamic weights of the corresponding connection edges. :

[0046]

[0047] in, Scale adjustment parameters It is a natural exponential function.

[0048] By integrating the network nodes, the real-time risk intensity values ​​of each network node, the connecting edges, and the dynamic weights of the connecting edges, a dynamically weighted risk transmission network is generated. The dynamic weights of the connecting edges are functions that change over time and are dynamically updated (e.g., every 5 minutes) as real-time data (power flow, electricity price, congestion status) changes.

[0049] In this embodiment, the calculation steps for the conduction resistance are as follows:

[0050] The day-ahead market clearing results of the inter-regional power system are obtained to determine the power transmission direction between nodes; simultaneously, real-time dispatch operation status is collected to obtain the actual power flow of each inter-provincial channel. Based on the day-ahead results and real-time operation status, channels in the dynamic weighted risk transmission network whose power transmission ratio exceeds a preset threshold are identified and defined as main power transmission paths. For each connection edge on the main power transmission path, its actual power flow is recorded, and combined with the day-ahead clearing power direction, a dynamic risk transmission barrier function is used to determine the power flow direction. For bidirectional asymmetric barrier calculations, the calculation focuses on the transmission margin of the physical channel for a positive barrier where the actual power flow direction is aligned with the day-ahead clearing power direction, yielding the positive barrier value. For a negative barrier where the actual power flow direction is opposite to the day-ahead clearing power direction, the calculation introduces an additional penalty coefficient on top of the blockage level of the physical channel. The reverse potential barrier value is obtained, and the forward and reverse potential barrier values ​​are weighted and combined to generate the transmission resistance. .

[0051] S3: Input any risk indicator from the risk indicator dataset into the risk transmission network for simulation and prediction of the probability distribution of risk intensity at each node in the future.

[0052] In this embodiment, any risk indicator from the risk indicator dataset is input into the risk transmission network for simulation and extrapolation to predict the probability distribution of risk intensity at each node in the future. Specifically:

[0053] Based on the weights of each connection edge in the dynamic risk transmission network, a time-varying system matrix is ​​constructed. Time-varying system matrix It is The square formation, For the total number of nodes, the matrix of the time-varying system is... Singular value decomposition is performed to obtain multiple propagation modes and their corresponding mode weights, specifically:

[0054]

[0055] in, and These are the left singular matrix and the right singular matrix, representing the directions of different risk patterns. It is a diagonal matrix containing a series of singular values. , This represents the intensity of the k-th conduction mode at time t.

[0056] For singular values Normalization is performed to obtain the mode weights of each transmission mode. .

[0057] Based on real-time system state data from multiple sources, the activation factors for each transmission mode are calculated:

[0058]

[0059] in, To construct a system state vector based on real-time data from multiple sources at the same time t; Let k be the left singular vector corresponding to the k-th pattern. The inner product of the system state vector and the mode vector. Let be the Euclidean norm of the system state vector. Let be the Euclidean norm of the left singular vector corresponding to the k-th pattern. is the activator of the k-th conduction mode.

[0060] The process of quantifying the risk indicator into initial risk energy and injecting it into the source nodes directly affected by it, followed by several rounds of Monte Carlo simulation, is as follows:

[0061] The Monte Carlo simulation is specifically as follows:

[0062] In a dynamically weighted risk propagation network, the dynamic weights of each connection edge are used to calculate the probability of risk propagation from the current node to any of its neighboring nodes.

[0063]

[0064] in, Let be the set of neighboring nodes of node i. It is the sum of the propagation weights of all neighboring nodes at time t.

[0065] The transition probability is adjusted based on the mode weight and the corresponding activation factor, specifically as follows:

[0066]

[0067] in, The adjusted transition probability. This is an adjustment factor.

[0068] Based on the adjusted transition probability, the transmission path and direction of risk are determined by random sampling;

[0069] During the transmission process, risk energy is dissipated according to the dynamic risk transmission barrier function until the simulation time reaches the preset future period or the risk energy falls below the preset energy threshold, thus completing one round of simulation. Specifically:

[0070]

[0071] in For node j at time... The risk energy gained Let i be the risk energy at time t. Let be the risk transmission barrier function at time t.

[0072] Repeat the above process for 1000 rounds of simulation, statistically analyze the simulation results of all rounds, and generate the probability distribution of risk intensity for future time periods at each node.

[0073] S4, based on the probability distribution of risk intensity, outputs strategic early warning instructions by optimizing the decision-making model;

[0074] In this embodiment, the step of outputting strategic early warning instructions based on the risk intensity probability distribution and by optimizing the decision model specifically includes:

[0075] The statistical characteristic value can be the expected value or the quantile; in this embodiment, the expected value is used as an example.

[0076] The risk debt equivalent is obtained by subtracting the expected value of the risk intensity probability distribution of each node from the real-time available transmission capacity margin. When the expected value exceeds the available capacity margin, it indicates that the node needs intervention, and the set of nodes that need to be forcibly intervened is determined accordingly.

[0077] Based on the physical operational constraints of the nodes requiring mandatory intervention, specifically through the outgoing thermal stability limit of the nodes requiring mandatory intervention and the real-time power flow... Calculate its physical sensitivity weight using the following formula:

[0078]

[0079] in, Let be the outgoing line thermal stability limit of node i, representing the maximum power that the node can withstand without exceeding the equipment safety limits. The real-time power flow at node i at time t is obtained in real time by the SCADA or EMS system. Let be the physical sensitivity weight of node i at time t, reflecting the proportion of its power flow to the limit capacity.

[0080] With the aforementioned risk debt equivalent For the product, the physical sensitivity weight Using cost coefficients and real-time electricity price differences between nodes as revenue coefficients, an equilibrium model is constructed, specifically as follows:

[0081]

[0082] in, The amount of control task for node i, For the control workload of nodes other than node i, This is the total control quantity of the system. For the profit of node i, The real-time electricity price difference between node i and other nodes is obtained from market clearing data.

[0083] The equilibrium price is obtained by solving the Nash-Cournot equilibrium equations. Specifically:

[0084] Introducing multipliers through the Lagrange method Define the Lagrange function as :

[0085]

[0086] Find the partial derivative:

[0087]

[0088] Obtain the optimal control amount for each node. :

[0089]

[0090] Reuse the total control amount of the system Apply constraints:

[0091]

[0092] Solving the equations yields the Lagrange method, which introduces multipliers. :

[0093]

[0094] The Lagrange method obtained from the solution is introduced into the multipliers. Defined as equilibrium price ,Right now:

[0095]

[0096] The specific steps for generating executable, strategy-based early warning instructions based on equilibrium prices are as follows:

[0097] The risk hedging coefficient is calculated by comparing the equilibrium price with the real-time marginal electricity price at each node.

[0098]

[0099] in, The risk hedging coefficient represents the relative deviation between the current marginal electricity price and the equilibrium price of a node. Its value can be positive or negative. A positive value indicates that the equilibrium price is higher than the node price (the node can obtain positive revenue / hedging opportunities through regulation), while a negative value indicates that the equilibrium price is lower than the node price (the node may bear relative costs through regulation). The real-time marginal electricity price of node i at time t is obtained from market clearing or node-level market data.

[0100] Multiplying the node's risk hedging coefficient by its physical sensitivity weight yields a node-level risk adjustment priority score. A higher score indicates a higher priority for the node when allocating limited regulatory resources. All nodes requiring intervention are sorted in descending order based on their risk adjustment priority scores, with the node having the highest score receiving the highest regulatory priority. The total regulatory amount is based on the risk-debt equivalent, and regulatory tasks are allocated to each node according to their risk adjustment priority scores. The allocated regulatory task for each node must not exceed its adjustable physical capacity or be lower than its minimum acceptable adjustment amount. Each node's regulatory task and its symbol are encapsulated into an executable, strategy-based early warning instruction. For example, if a node's equilibrium price is greater than its real-time marginal electricity price, and its allocated regulatory task is greater than zero, the instruction would be "reduce power purchases / increase external power output / reduce load," because the node has hedging benefits at the equilibrium price and should prioritize power output or reduce power purchases. If a node's equilibrium price is less than its real-time marginal electricity price, it might suggest "reduce external power transmission / increase power purchases" or execute in the negative direction.

[0101] Example 2, Figure 2 The present invention provides an intelligent early warning device for cross-regional electricity spot market purchase risks, characterized in that it comprises:

[0102] The data extraction module is used to extract risk indicator data that affect inter-provincial spot electricity purchases and form a risk indicator dataset.

[0103] The network building module is used to construct a dynamically weighted risk transmission network based on a risk indicator dataset;

[0104] The risk simulation module is used to input any risk indicator from the risk indicator dataset into the risk transmission network for simulation and to predict the probability distribution of risk intensity at each node in the future.

[0105] The optimization decision module is used to output strategic early warning instructions based on the probability distribution of risk intensity through an optimized decision model.

[0106] In detail, the modules in the intelligent early warning device for cross-regional electricity spot market purchase risks described in this embodiment of the invention employ the same technical means as the intelligent early warning method for cross-regional electricity spot market purchase risks described in the accompanying drawings, and can produce the same technical effects, which will not be repeated here.

[0107] This embodiment includes an intelligent early warning device for electricity purchase risks in cross-regional electricity spot markets, and the device may include:

[0108] Memory and processor;

[0109] The memory is used to store programs;

[0110] The processor is used to execute programs stored in the memory.

[0111] The processor is the control unit of the device. It connects to various components of the electronic device through various interfaces and lines. It executes programs or modules stored in the memory and calls data stored in the memory to perform various functions and process data of the intelligent early warning device for cross-regional electricity spot market purchase risks.

[0112] The figure only shows an electronic device with components. Those skilled in the art will understand that the structure shown in the figure does not constitute a limitation on the electronic device and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0113] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0114] The program stored in the memory of the electronic device is a combination of multiple instructions. When run in the processor, it can implement the steps in the above-mentioned intelligent early warning method for electricity purchase risks in the cross-regional electricity spot market.

[0115] Specifically, the processor's specific implementation system of the above instructions can be found in the description of the relevant steps in the corresponding embodiments of the accompanying drawings, which will not be repeated here.

[0116] 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.

[0117] 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.

[0118] 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.

[0119] 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.

[0120] 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.

[0121] 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 intelligent early warning of electricity purchase risks in cross-regional electricity spot markets, characterized in that, Includes the following steps: Extract risk indicator data affecting inter-provincial spot electricity purchases, form and construct a dynamic weighted risk transmission network based on the risk indicator dataset. Specifically, divide the power system into regions, define each region as a network node, and define the connection edges between nodes based on the grid topology and transaction coupling relationship between regions. Based on the risk indicator dataset, calculate the real-time risk intensity value of each network node; The transmission resistance of each connecting edge is dynamically calculated through the risk transmission barrier function, and the transmission resistance is mapped to the dynamic weight of each connecting edge. Based on the network nodes, real-time risk intensity values, connection edges, and dynamic weights, a dynamically weighted risk transmission network is generated. Input any risk indicator from the risk indicator dataset into the dynamic weighted risk transmission network for simulation and prediction of the probability distribution of risk intensity at each node in the future. Based on the probability distribution of risk intensity, a strategic early warning instruction is output by optimizing the decision-making model, specifically: Based on the probability distribution of risk intensity of each node and the real-time available transmission capacity margin, the risk debt equivalent is calculated, and the set of nodes that require mandatory intervention is determined. Calculate the physical sensitivity weight of nodes that require mandatory intervention based on their physical operation constraints. Using risk debt equivalent as commodity quantity, physical sensitivity weight and real-time electricity price difference between nodes as market parameters, an equilibrium model is constructed and solved to obtain the equilibrium price. The early warning instruction is generated based on the equilibrium price, specifically as follows: The equilibrium price is compared with the real-time node marginal electricity price to calculate the risk hedging coefficient, and then combined with the physical sensitivity weight to generate a risk adjustment priority score. The nodes requiring intervention are sorted according to their priority scores to determine the priority sequence of regulation. The total amount of regulation is based on the risk-debt equivalent, and the amount of regulation task is allocated to each node according to the priority score ratio of the nodes in the regulation priority sequence. The control tasks of each node are encapsulated into early warning instructions, including: if the equilibrium price of a node is greater than the real-time marginal electricity price of the node, and the control task allocated to the node is greater than 0, the instruction is to increase external power transmission; otherwise, the instruction is to reduce external power transmission or execute in the negative direction.

2. The intelligent early warning method for cross-regional electricity spot market purchase risks according to claim 1, characterized in that, The extraction of risk indicator data affecting inter-provincial spot electricity purchases to form a risk indicator dataset specifically involves: Acquire data from multiple transaction events, power grid operation, external environment, and risk indicators to construct a multi-source integrated database; Based on the power grid topology and power flow data in the power grid operation data, calculate the transmission congestion margin and power deviation of each inter-provincial tie line; Based on transmission congestion margin and power deviation, calculate the congestion surplus index of each inter-provincial connecting line; The overall operating status of the inter-provincial market is dynamically determined based on the blockage surplus index. Based on the operating status, the corresponding risk indicator data is extracted from the multi-source fusion database and organized according to time series to form a risk indicator dataset.

3. The intelligent early warning method for cross-regional electricity spot market purchase risks according to claim 2, characterized in that, The calculation steps for the transmission resistance are as follows: Based on the day-ahead clearing results and real-time operating status of the spot market, identify the main power transmission paths and directions in the dynamic weighted risk transmission network; For the connecting edges on the main power transmission path, distinguish between the forward and reverse directions of power transmission, and calculate the asymmetric dynamic risk conduction barrier value separately; The positive barrier value aligned with the direction of the clearing power is determined based on the transmission margin of the physical channel; The reverse barrier value, which is opposite to the direction of the clearing power, is calculated by introducing a penalty coefficient based on the transmission margin. The forward and reverse potential barrier values ​​are weighted and combined to generate the transmission resistance.

4. The intelligent early warning method for cross-regional electricity spot market purchase risks according to claim 3, characterized in that, The step of inputting any risk indicator from the risk indicator dataset into the risk transmission network for simulation and prediction of the probability distribution of risk intensity at each node in the future is as follows: A time-varying system matrix is ​​constructed based on the weight values ​​of each connection edge in the dynamic risk transmission network; Singular value decomposition is performed on the time-varying system matrix to obtain multiple transmission modes and the mode weights corresponding to each transmission mode; Activation factors for each transmission mode are calculated based on real-time system state data from multiple sources. The risk indicators are quantified into initial risk energy and injected into the source nodes directly affected by them, and several rounds of Monte Carlo simulation are performed. The simulation results of all rounds are statistically analyzed to generate the probability distribution of risk intensity for future time periods at each node.

5. The intelligent early warning method for cross-regional electricity spot market purchase risks according to claim 4, characterized in that, The Monte Carlo simulation method is as follows: The probability of risk propagating from the current node to its neighboring nodes is calculated based on dynamic weights. The transition probability is adjusted based on the mode weight and the corresponding activation factor; Based on the adjusted transition probability, the transmission path and direction of risk are determined by random sampling; During the transmission process, the initial risk energy is dissipated according to the dynamic risk transmission barrier function until the simulation time reaches the preset future period or the risk energy is lower than the preset energy threshold, thus completing one round of simulation.

6. An apparatus using the intelligent early warning method for cross-regional electricity spot market purchase risks as described in any one of claims 1-5, characterized in that, include: The data extraction module is used to extract risk indicator data that affect inter-provincial spot electricity purchases and form a risk indicator dataset. The network building module is used to construct a dynamically weighted risk transmission network based on a risk indicator dataset; The risk simulation module is used to input any risk indicator from the risk indicator dataset into the risk transmission network for simulation and to predict the probability distribution of risk intensity at each node in the future. The optimization decision module is used to output strategic early warning instructions based on the probability distribution of risk intensity through an optimized decision model.

7. A smart early warning device for electricity purchase risks in a cross-regional electricity spot market, characterized in that, include: Memory and processor; The memory is used to store programs; The processor is used to execute a program to implement the intelligent early warning method for cross-regional electricity spot market purchase risks as described in any one of claims 1-5.

Images