An extreme weather-based power transaction auxiliary decision processing method and system

Abstract

The application discloses a kind of based on extreme weather electric power transaction auxiliary decision processing method and system, it is related to electric power transaction decision-making technical field, including obtaining extreme weather early warning message formation original meteorological early warning data set;Space reference alignment is executed to the data set, and the grid chart containing meteorological intensity matrix and power generation facility distribution index is generated;Meteorological intensity matrix is mapped to obtain meteorological sensitivity coefficient matrix, and power generation facility distribution index topological correlation obtains power generation side asset vulnerability correlation table;Coupling deduction is generated in the risk transmission calculation engine by inputting both, and the electric power supply and demand imbalance risk thermodynamic diagram is generated;Based on the thermodynamic diagram, construct transaction price elasticity prediction model, form transaction strategy reference matrix;Through decision rule base matching generation offer interval suggestion, and conversion is standardized declaration instruction stream and is pushed to transaction interface end.This method can accurately connect extreme weather information and electric power transaction decision-making, improve the accuracy and standardization of transaction decision-making.

Description

Technical Field

[0001] This invention belongs to the field of power trading decision-making technology, specifically a power trading auxiliary decision-making processing method and system based on extreme weather. Background Technology

[0002] Currently, power trading decision support under extreme weather conditions largely relies on single meteorological warnings, with decisions made manually in conjunction with historical trading data. Some technologies obtain extreme weather warnings from meteorological monitoring networks while simultaneously collecting information on power generation facilities, attempting to combine both to provide a reference for power trading. However, in these technologies, meteorological warnings originate from monitoring stations in different regions, and the coordinate systems and data formats used by each station vary, resulting in a lack of a unified standard for the spatiotemporal correlation between power generation facility information and meteorological warnings.

[0003] Current technologies fail to align acquired extreme weather warnings with a unified spatiotemporal benchmark, making it impossible to form a standardized grid of extreme weather impact areas. This results in a mismatch between meteorological intensity information and the distribution of power generation facilities, making it difficult to accurately assess the impact of extreme weather on different types of power generation resources. Furthermore, current technologies do not couple meteorological intensity with the vulnerability of power generation assets, nor do they establish a standardized conversion mechanism from risk analysis to transaction declaration instructions. This makes it impossible to accurately quantify the risk of power supply and demand imbalance under extreme weather conditions, lacks a scientific basis for predicting transaction price fluctuations, and makes it difficult to directly adapt decision-making results to the declaration requirements of power trading platforms, leading to low efficiency and insufficient accuracy in transaction decisions. Summary of the Invention

[0004] This invention aims to solve at least one of the technical problems existing in the prior art;

[0005] Therefore, this invention proposes a power trading auxiliary decision processing method based on extreme weather, comprising:

[0006] Extreme weather warning messages containing meteorological element codes, disaster level indicators, and effective time windows are obtained from the meteorological monitoring network to form the original meteorological warning dataset;

[0007] Spatiotemporal benchmark alignment is performed on the original meteorological warning dataset to generate a grid map of extreme weather impact areas in a unified coordinate system. The grid map of extreme weather impact areas includes a gridded meteorological intensity matrix and an index of the distribution of power generation facilities within the coverage area.

[0008] Influence factor mapping is performed on the meteorological intensity matrix to obtain a meteorological sensitivity coefficient matrix for different types of power generation resources. Topological association is performed on the power generation facility distribution index to obtain a power generation asset vulnerability association table.

[0009] The meteorological sensitivity coefficient matrix and the vulnerability correlation table of power generation side assets are input into the risk transmission calculation engine for coupled simulation to generate a heat map of power supply and demand imbalance risk.

[0010] The risk level of each key transmission section is extracted from the heat map of power supply and demand imbalance risk, and the risk level is quantified as the proportion of transmission capacity loss.

[0011] Based on the aforementioned transmission capacity loss ratio, a trading strategy reference matrix is ​​generated using a trading price elasticity prediction model.

[0012] Based on the trading strategy reference matrix, a pricing range suggestion for a specific power trading cycle is generated through a decision rule base matching mechanism, including: the pricing range suggestion includes a benchmark pricing curve and a floating fault tolerance threshold;

[0013] The trading periods in which price fluctuations exceed a preset threshold are identified from the trading strategy reference matrix and marked as key trading windows;

[0014] The decision rule base is retrieved to obtain a pricing strategy template that matches the current extreme weather type. The pricing strategy template specifies the pricing behavior guidelines under scenarios of supply shortage or supply surplus.

[0015] The predicted price fluctuation range and power grid congestion status of the key trading window are filled into the placeholders of the quotation strategy template to calculate the specific quotation base point value and the allowed fluctuation range parameter.

[0016] The base price value is combined with the allowable fluctuation range parameter to form the benchmark price curve and the floating tolerance threshold, which together constitute the price range recommendation.

[0017] The proposed price range is converted into a standardized declaration instruction stream that can be recognized by the power trading platform, and the standardized declaration instruction stream is pushed to the trading interface.

[0018] Furthermore, influence factor mapping is performed on the meteorological intensity matrix to obtain meteorological sensitivity coefficient matrices for different types of power generation resources, including:

[0019] Retrieve a pre-established mapping dictionary between power generation resource types and meteorological elements. The mapping dictionary defines the key meteorological influencing factors corresponding to wind turbine generators, photovoltaic arrays, and conventional thermal power units.

[0020] Traverse each grid cell in the meteorological intensity matrix and extract the intensity value of each meteorological element within the grid cell;

[0021] Based on the mapping dictionary, the extracted meteorological element intensity values ​​are substituted into the sensitivity functions of the corresponding power generation resource types for normalization, and the sensitivity scores of the grid cells for different power generation resource types are output.

[0022] The sensitivity scores of all grid cells are arranged and combined according to the power generation resource type to construct a multi-dimensional meteorological sensitivity coefficient matrix.

[0023] Furthermore, a topological association is performed on the distribution index of the power generation facilities to obtain a vulnerability association table for power generation-side assets, including:

[0024] Obtain the equipment health status file and geographical location coordinates of each power generation unit recorded in the power generation facility distribution index;

[0025] Using geographic information system tools, the spatial intersection relationship between the geographic location coordinates and the boundary of the grid map of the extreme weather impact area is calculated, and a set of power generation units within the impact range is selected.

[0026] By reading the historical fault frequency and current maintenance status from the equipment health status file and combining them with the type of the corresponding power generation unit, a comprehensive vulnerability index is calculated.

[0027] The set of power generation units and their corresponding comprehensive vulnerability indices are bound together as key-value pairs to generate the vulnerability association table of the power generation side assets.

[0028] Furthermore, the meteorological sensitivity coefficient matrix and the vulnerability correlation table of power generation-side assets are input into the risk transmission calculation engine for coupled simulation to generate a heat map of power supply and demand imbalance risk, including:

[0029] In the risk transmission calculation engine, a blank risk accumulation matrix corresponding to the power grid topology is initialized;

[0030] Based on the vulnerability association table of the power generation side assets, the grid node where the high-risk power generation unit is located is located, and the power output attenuation coefficient of the grid node is adjusted according to the meteorological sensitivity coefficient matrix.

[0031] Using the adjusted power output attenuation coefficient as input, the propagation path and amplification effect of the power deficit at the power grid node in the power grid are simulated in the power grid power flow calculation model, and the propagation results are accumulated to the corresponding position of the blank risk accumulation matrix.

[0032] Once the simulation calculations for all affected nodes are completed, the risk accumulation matrix is ​​rendered using color gradation to output a visualized heat map of the power supply and demand imbalance risk.

[0033] Furthermore, the step of generating a trading strategy reference matrix based on the transmission capacity loss ratio and the trading price elasticity prediction model includes:

[0034] The transmission capacity loss ratio is used as an input variable and input into the transaction price elasticity prediction model. The transaction price elasticity prediction model has a pre-set price linkage logic between different regional markets.

[0035] Execute the aforementioned transaction price elasticity prediction model to output the predicted marginal electricity price of each electricity trading node within a series of future trading periods;

[0036] The deviation of the predicted marginal electricity price of the node from the benchmark electricity price is calculated, and the trading period, node identifier and corresponding deviation are integrated into the trading strategy reference matrix.

[0037] Furthermore, the step of filling the predicted price fluctuation range and grid congestion status of the key trading window into the placeholders of the quotation strategy template, and calculating the specific quotation base point value and the allowed fluctuation range parameters, includes:

[0038] Read the risk preference configuration item from the pricing strategy template, which defines three pricing tendencies: aggressive, moderate, and conservative.

[0039] Based on the magnitude of the predicted price fluctuations in the key trading window, the allocation weight of the pricing tendency is dynamically adjusted; the greater the fluctuation, the lower the weight of the aggressive approach.

[0040] Based on the adjusted weights, the pricing calculation formula in the pricing strategy template is instantiated to obtain a set of preliminary pricing base values ​​and fluctuation ranges;

[0041] The grid congestion situation is added as a correction factor to the initial bid base value, and the final bid base value and the allowable fluctuation range parameter are output.

[0042] Furthermore, the proposed price range is transformed into a standardized order flow recognizable by the power trading platform, including:

[0043] The benchmark price curve in the proposed price range is analyzed and converted into time-segmented quantity-price declaration point pairs as specified by the power trading platform.

[0044] The floating fault tolerance threshold is mapped to the upper and lower limits of the bid price allowed by the trading platform;

[0045] In accordance with the data message format specifications of the trading platform, the time-segmented volume and price declaration point pairs and the upper and lower limits of the declaration price are encapsulated into a unified declaration data packet structure;

[0046] The declaration data packet structure is subjected to checksum calculation and signature encryption to generate the standardized declaration instruction stream that meets the requirements for secure transmission.

[0047] Furthermore, after generating the aforementioned heatmap of power supply and demand imbalance risk, the method also includes:

[0048] The areas most severely affected by extreme weather were extracted from the aforementioned heat map of power supply and demand imbalance risk and defined as high-sensitivity load centers.

[0049] Obtain the typical daily load curve of the high-sensitivity load center, and perform morphological analysis on the typical daily load curve to extract the base load component and peak shift potential of the load.

[0050] The base load component and peak shift potential of the load are used as constraints and back-injected into the input of the transaction price elasticity prediction model to perform a second correction on the transaction price elasticity prediction model.

[0051] The trading strategy reference matrix is ​​regenerated using the modified trading price elasticity prediction model to update the quote range recommendations.

[0052] Furthermore, after pushing the standardized declaration instruction stream to the transaction interface, the process also includes:

[0053] Continuously monitor the declaration confirmation receipt and real-time transaction data stream returned by the transaction interface;

[0054] The actual transaction price in the real-time transaction data stream is compared with the predicted price fluctuation range in the previously generated trading strategy reference matrix, and the prediction deviation value is calculated.

[0055] The prediction deviation value is fed back to the transaction price elasticity prediction model as part of the model training samples, and the parameters of the transaction price elasticity prediction model are fine-tuned.

[0056] The adjusted transaction price elasticity prediction model was then applied to the analysis of the next extreme weather cycle.

[0057] Furthermore, the present invention also includes a power trading auxiliary decision processing system based on extreme weather, the system including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, it implements the steps of the power trading auxiliary decision processing method based on extreme weather described above.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] Spatiotemporal alignment is performed on extreme weather warning messages obtained from the meteorological monitoring network, which include meteorological element codes, disaster level identifiers, and effective time windows. This generates a grid map of the extreme weather impact area in a unified coordinate system. This grid map also integrates a gridded meteorological intensity matrix and an index of the distribution of power generation facilities within the coverage area. By aligning the spatiotemporal reference, differences in coordinate systems and data formats between different meteorological monitoring station reports are eliminated. This achieves a precise correspondence between meteorological warning information and the spatial location of power generation facilities, establishing a one-to-one correspondence between the impact range and intensity of extreme weather and specific power generation facilities. It clearly presents the coverage of various power generation facilities by extreme weather of different regions and intensities, avoiding biases in meteorological impact judgments caused by inconsistent spatiotemporal references.

[0060] The meteorological sensitivity coefficient matrix, mapped from the meteorological intensity matrix, and the vulnerability association table of power generation assets, obtained from the topological association of power generation facility distribution index, are jointly input into the risk transmission calculation engine for coupled simulation to generate a heat map of power supply and demand imbalance risk. Based on this heat map, a transaction price elasticity prediction model is constructed to simulate nodal electricity price fluctuations, forming a transaction strategy reference matrix. A bid range suggestion is generated through matching with a decision rule base and then transformed into a standardized declaration command flow recognizable by the power trading platform and pushed to the trading interface. Through the coupled simulation of meteorological sensitivity and power generation asset vulnerability, a deep correlation between the impact of extreme weather and the risk of power generation assets is achieved. This accurately presents the spatial distribution of power supply and demand imbalance, making the electricity price fluctuation simulation more realistic and the generated bid range suggestion more targeted. The standardized declaration command flow can directly connect to the power trading platform without manual secondary conversion, simplifying the transaction declaration process and achieving seamless integration from risk analysis to transaction declaration, thus improving the standardization and timeliness of transaction decisions. Attached Figure Description

[0061] Figure 1 This is a flowchart illustrating the steps of a power trading auxiliary decision processing method based on extreme weather as described in this invention.

[0062] Figure 2 A flowchart for generating the meteorological sensitivity coefficient matrix;

[0063] Figure 3 A flowchart for generating a vulnerability association table for power generation side assets;

[0064] Figure 4 A diagram showing the distribution of standardized declaration command flow price declaration points and price upper and lower limits;

[0065] Figure 5 This is a graph showing the predicted fluctuations in nodal electricity prices. Detailed Implementation

[0066] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0067] See Figure 1 This method begins with the acquisition and standardization of raw meteorological warning data. Extreme weather warning messages issued by the meteorological monitoring network contain key fields such as meteorological element codes, disaster level identifiers, and effective time windows. These messages are collected to form a raw meteorological warning dataset. Subsequently, a spatiotemporal benchmark alignment operation is performed on this dataset, uniformly transforming it to a preset coordinate system to generate a grid map of the extreme weather impact area. This grid map embeds a meteorological intensity matrix organized in grid cells and establishes an index of power generation facility distribution associated with the grid area.

[0068] To assess the impact of weather on the power system, targeted analysis of meteorological data is necessary. An impact factor mapping operation is performed on the generated meteorological intensity matrix, transforming meteorological intensity into a quantitative impact on different types of power generation resources, outputting a meteorological sensitivity coefficient matrix. In parallel, topological correlation analysis is performed on the distribution index of power generation facilities, combining the equipment status and geographical location of power generation units to assess their vulnerability under extreme weather conditions, generating a vulnerability correlation table for power generation assets. Subsequently, the meteorological sensitivity coefficient matrix and the vulnerability correlation table for power generation assets are input into a pre-defined risk transmission calculation engine. This engine simulates the transmission and superposition process of weather-induced power generation capacity attenuation in the power grid, ultimately outputting a heat map of power supply and demand imbalance risk that intuitively reflects the regional power supply and demand tension.

[0069] Based on the risk analysis results, the process moves to the trading strategy simulation stage. Using the aforementioned heatmap of power supply and demand imbalance risk as input, a trading price elasticity prediction model is constructed or invoked. This model can simulate the price fluctuations at power market nodes under a specific risk distribution, outputting a structured trading strategy reference matrix. This matrix clarifies the predicted price fluctuation ranges for different trading periods and nodes in the future. Next, based on the risk and price signals revealed by the trading strategy reference matrix, a pre-set decision rule base is queried to match and instantiate macro strategies into specific price range suggestions for the upcoming specific power trading cycle. Finally, the aforementioned human- or system-readable price range suggestions are transformed into a standardized data format that the power trading platform interface can directly recognize and process, namely a standardized declaration instruction stream, and automatically pushed to the trading interface of the power trading platform, completing the closed-loop operation from weather warning to transaction declaration.

[0070] In one embodiment of the present invention, during the process of generating the meteorological sensitivity coefficient matrix, refer to... Figure 2 The system retrieves a pre-established mapping dictionary, which defines the correspondence between different types of power generation resources and meteorological elements. For example, key meteorological influencing factors for wind turbine generators include wind speed and turbulence intensity, photovoltaic arrays correspond to solar radiation intensity and cloud cover, and conventional thermal power units may be associated with ambient temperature and air pressure. The system traverses each grid cell in the meteorological intensity matrix, reading the intensity values ​​of various meteorological elements within that cell. Guided by the mapping dictionary, the extracted intensity values ​​of specific meteorological elements are substituted into the sensitivity functions of the corresponding types of power generation resources for calculation. The sensitivity functions normalize the original intensity values, converting them into a score representing the degree of sensitivity. After completing the traversal and calculation of all grid cells, the system categorizes the power generation resources by type, arranging and combining the sensitivity scores of each grid cell to construct a multi-dimensional meteorological sensitivity coefficient matrix. Each layer of this matrix corresponds to a type of power generation resource, and each grid point stores the sensitivity score for that type of resource at that location.

[0071] In practical implementation, the mapping dictionary is a structured lookup table stored in the system database. It explicitly defines the key meteorological influencing factors for each of the three main power generation resource types: wind turbine generators, photovoltaic arrays, and conventional thermal power units. Specifically, for wind turbine generators, the mapping dictionary defines key meteorological influencing factors including ten-minute average wind speed, wind speed gust factor, and ambient air density; for photovoltaic arrays, it defines key meteorological influencing factors including total horizontal irradiance, diffuse irradiance, and photovoltaic panel temperature; and for conventional thermal power units, it defines key meteorological influencing factors including ambient dry-bulb temperature, wet-bulb temperature, and atmospheric pressure. In some embodiments, the contents of the mapping dictionary can be updated and maintained through configuration files to adapt to new power generation technologies or the need for more refined meteorological impact analysis.

[0072] In implementation, the system traverses each grid cell in the meteorological intensity matrix. Each grid cell stores standardized intensity values ​​of various meteorological elements after spatiotemporal alignment. The processing module reads the intensity values ​​of each meteorological element within a grid cell, such as wind speed, irradiance, and temperature. The traversal process covers all rows and columns of the meteorological intensity matrix, ensuring that each grid cell corresponding to a geographical location is processed. Guided by the mapping dictionary, the system substitutes the extracted meteorological element intensity values ​​into the sensitivity function for the corresponding power generation resource type. In implementation, the sensitivity function is a normalization function that maps the original meteorological element intensity values ​​to a sensitivity score between 0 and 1. The closer the score is to 1, the more sensitive the meteorological conditions are to the type of power generation resource. For the sensitivity function of wind turbine generators, the input is the wind speed intensity value, and the output is the sensitivity score for the wind turbine generator. In practical implementation, the sensitivity function for photovoltaic arrays takes irradiance intensity and photovoltaic panel surface temperature intensity as inputs, and outputs a sensitivity score for the photovoltaic array. Similarly, the sensitivity function for conventional thermal power units takes ambient dry-bulb temperature intensity as input and outputs a sensitivity score for the conventional thermal power unit. Optionally, the sensitivity function for wind turbine generators can be calculated using the following form:

[0073] ;

[0074] in: This indicates a sensitivity score for wind turbine generators. This represents the ten-minute average wind speed intensity value extracted from the meteorological intensity matrix. This represents the center point parameter of the wind speed sensitivity curve. This parameter represents the steepness of the curve. In practical implementation, the parameter... and The values ​​are predefined based on the technical characteristics of the wind turbine generator set, such as the cut-in wind speed, rated wind speed, and cut-out wind speed.

[0075] In practical implementation, after calculating the sensitivity scores for all grid cells, the system arranges and combines the scores according to the power generation resource type. Specifically, the system creates a two-dimensional numerical layer for wind turbine generators, photovoltaic arrays, and conventional thermal power units. The system then fills the sensitivity scores of wind turbine generators into the corresponding grid cell positions, forming a wind power sensitivity layer. Similarly, the system fills the sensitivity scores of photovoltaic arrays into the corresponding grid cell positions, forming a photovoltaic power sensitivity layer. Finally, the system fills the sensitivity scores of conventional thermal power units into the corresponding grid cell positions, forming a thermal power sensitivity layer. These three sensitivity layers are logically stacked to form a three-dimensional meteorological sensitivity coefficient matrix. Each grid point in the meteorological sensitivity coefficient matrix corresponds to a vector containing the three types of scores.

[0076] In one embodiment of the invention, the process of generating a vulnerability association table for power generation-side assets begins with the distribution index of power generation facilities. See also... Figure 3 The system retrieves detailed information for each power generation unit recorded in the index, including its equipment health status profile and precise geographic coordinates. Using the spatial analysis capabilities of Geographic Information System (GIS) tools, it calculates the positional relationship between the geographic coordinates of each power generation unit and the spatial range represented by the grid map of extreme weather impact areas. By calculating the spatial intersection, it filters out all power generation units located within the extreme weather impact area, forming a set of affected power generation units. Next, the system reads data recorded in the equipment health status profile of each power generation unit in this set, such as historical failure frequency records and the current maintenance status identifier. Combining this with the type of power generation unit, it calculates a comprehensive vulnerability index using a pre-defined weighted calculation model. Finally, the system uses the unique identifier of the power generation unit as the key and its calculated comprehensive vulnerability index as the value to perform key-value pair binding, generating a power generation asset vulnerability association table.

[0077] In practical implementation, the processing begins with the power generation facility distribution index, a data structure that records the basic attributes and spatial locations of all registered power generation units within a region. In some embodiments, the power generation facility distribution index exists in the form of a database table, where each record corresponds to a power generation unit. Record fields include at least the unique identifier of the power generation unit, the power generation unit type, rated capacity, grid connection node number, equipment health status file storage path, and geographical coordinates expressed in latitude and longitude. The system retrieves all records in the power generation facility distribution index through a database query interface, parsing the equipment health status file storage path and geographical coordinate fields in each record. The equipment health status file is an independently stored, time-updated file or data table that records the historical operation, faults, and maintenance information of the power generation unit. After obtaining the geographical coordinates, the system calls integrated geographic information system (GIS) tools for spatial analysis calculations. The GIS tools load the boundary polygon data of a pre-generated grid map of extreme weather impact areas. This grid map consists of a series of continuous grid cells, each with its own defined geographical boundary. Geographic Information System (GIS) tools calculate the spatial relationship between the geographic coordinates of each power generation unit and the polygonal area covered by the grid map of the extreme weather impact area. By performing a point-polygon spatial inclusion relationship determination, all power generation units whose coordinates lie within the boundary polygon of the grid map of the extreme weather impact area are selected. These power generation units constitute a set of affected power generation units. Optionally, the GIS tools can use the ray casting method or the wrap-around number method from the computational geometry library to accurately determine the point-polygon inclusion relationship.

[0078] In practical implementation, for each power generation unit in the affected power generation unit set, the system reads the corresponding equipment health status file according to the equipment health status file storage path. In this implementation, the equipment health status file includes a historical fault frequency field and a current maintenance status field. The historical fault frequency field records the number or frequency of unplanned outages that occurred in the power generation unit within a specific past period. The current maintenance status field indicates whether the power generation unit is in a state of "normal operation," "planned maintenance," "temporary maintenance," or "fault outage." The system combines the read historical fault frequency value, current maintenance status identifier, and power generation unit type identifier, inputting them into a preset calculation model to calculate a comprehensive vulnerability index. It can be understood that the calculation model may assign different basic vulnerability weights to different types of power generation units. In some embodiments, the formula for calculating the comprehensive vulnerability index is as follows:

[0079] ;

[0080] in: This represents the calculated comprehensive vulnerability index. This represents the normalized historical fault frequency value. This represents the state coefficients obtained based on the current maintenance status mapping. This represents the basic coefficient for the type determined based on the type of power generation unit. , , These are the weight parameters for the corresponding items. In practical implementation, the weight parameters... , , The value is set by domain experts based on equipment reliability statistics and can be configured and adjusted.

[0081] In implementation, after calculating the comprehensive vulnerability index for all affected power generation units, the system performs data binding and table generation. Specifically, the system uses the unique identifier of each power generation unit as the primary key and the calculated comprehensive vulnerability index as the associated value to create a key-value pair. The system then aggregates all key-value pairs from all affected power generation units into a structured data table, defined as the Power Generation Asset Vulnerability Association Table. Each row in the Power Generation Asset Vulnerability Association Table represents a power generation unit affected by extreme weather, and each row must contain at least two core fields: the unique identifier of the power generation unit and the comprehensive vulnerability index. Optionally, the Power Generation Asset Vulnerability Association Table may also include auxiliary information fields such as power generation unit type, affiliated plant, and grid connection node to facilitate rapid association queries in subsequent processing stages.

[0082] In one embodiment of the present invention, after receiving the meteorological sensitivity coefficient matrix and the vulnerability association table of generation-side assets, the risk transmission calculation engine initiates coupled simulation to generate a heat map of power supply and demand imbalance risk. Internally, the engine first initializes a blank risk accumulation matrix that perfectly corresponds to the node-branch topology of the target power grid, with all elements of this matrix having an initial value of zero. Based on the vulnerability association table of generation-side assets, the engine locates the power grid nodes connected to generation units with high vulnerability indices, and dynamically adjusts the power output attenuation coefficients of these nodes in subsequent simulations according to the values ​​in the meteorological sensitivity coefficient matrix at the locations of these nodes. Using the adjusted power output attenuation coefficients as input, the engine simulates the power deficit caused by the decline in generation capacity at these nodes in a built-in power flow calculation model, and calculates the power flow redistribution caused by this deficit on power grid transmission lines and transformers, capturing the propagation path of the power deficit along the power grid topology and its potential amplification effect due to network congestion. The power imbalance risk value accumulated at each network node during the propagation process is added to the corresponding node position in the blank risk accumulation matrix.

[0083] After all marked affected nodes have completed simulation calculations, the final risk accumulation matrix, whose values ​​are no longer zero, undergoes color-gradient rendering, mapping different levels of risk values ​​to different colors, and outputting a visualized heatmap of power supply and demand imbalance risk. After generating the heatmap, the method may include a feedback correction step. The system extracts the areas with the highest risk values ​​from the generated heatmap, i.e., those most severely affected by extreme weather, defining these areas as high-sensitivity load centers. It acquires historical typical daily load curve data for these high-sensitivity load centers, performs morphological analysis on these curves, decomposes and extracts the base load component, and assesses the potential for load shifting during peak hours. The analyzed base load component and peak shifting potential information are used as new constraints, back-injected into the input of the trading price elasticity prediction model, and the model's internal parameters or logic are corrected a second time. Using this corrected trading price elasticity prediction model, the node electricity price fluctuation simulation is rerun, thereby updating and generating a new round of trading strategy reference matrix, which can then be used to further update the suggested price range.

[0084] In practical implementation, the risk transmission calculation engine initializes a blank risk accumulation matrix. The number of rows and columns in the blank risk accumulation matrix is ​​consistent with the number of nodes in the target power grid. Each element of the blank risk accumulation matrix corresponds to a physical node in the power grid, and the initial value of all elements is set to zero. In practical implementation, the risk transmission calculation engine queries the generation-side asset vulnerability association table. Records in the generation-side asset vulnerability association table with a comprehensive vulnerability index higher than a preset threshold are identified as high-risk generation units. The risk transmission calculation engine locates the power grid nodes connected to these high-risk generation units through the power grid topology connection table. After locating the power grid nodes, the risk transmission calculation engine adjusts the power output attenuation coefficient of the power grid nodes according to the meteorological sensitivity coefficient matrix. The risk transmission calculation engine reads the sensitivity score of the grid unit corresponding to the geographical location of the high-risk generation unit in the meteorological sensitivity coefficient matrix. In practical implementation, for a node containing a hybrid power station that includes both wind and solar power generation, the risk transmission calculation engine needs to comprehensively calculate a weighted average sensitivity score by combining the wind power sensitivity score and the solar power sensitivity score, along with the installed capacity ratio of the power station. In practice, the adjustment of the power output attenuation coefficient is calculated based on the weighted average sensitivity score and the comprehensive vulnerability index in the generator-side asset vulnerability correlation table. The power output attenuation coefficient represents the proportion of the expected maximum power generation output relative to the rated output at this node under extreme weather conditions. Optionally, the formula for calculating the power output attenuation coefficient can be expressed as:

[0085] ;

[0086] in: Represents a node The power output attenuation coefficient, Represents a node The overall weighted average sensitivity score of the associated power generation units, Represents a node The normalized value of the comprehensive vulnerability index of the associated power generation unit.

[0087] In practical implementation, the risk transmission calculation engine uses the adjusted power output attenuation coefficient as input and calls the built-in power flow calculation model for simulation. The power flow calculation model is constructed based on DC or AC power flow algorithms. The risk transmission calculation engine sets the injected power reduction for one or more nodes in the power flow calculation model. The injected power reduction is determined by the node's rated output and the power output attenuation coefficient. The power flow calculation model solves for the network power flow distribution under power deficit conditions. In practical implementation, the simulation captures the propagation path of the power deficit in the power grid. The propagation path is reflected through power changes in lines and transformers. The risk transmission calculation engine quantifies the unsafe factors caused by the power deficit, such as line overload and node voltage exceedance, into a local risk value. The risk transmission calculation engine accumulates the calculated node-related local risk values ​​to the corresponding node position in the blank risk accumulation matrix. When multiple affected nodes simultaneously have power deficits, their resulting local risk values ​​are algebraically superimposed at the middle node position of the blank risk accumulation matrix. In practice, after the simulation calculations for all marked affected nodes are completed, the risk transmission calculation engine performs post-processing on the accumulated risk matrix, where each element represents the comprehensive risk level of the corresponding power grid node. In the implementation, the color-gradient rendering process maps the risk level values ​​to a predefined color spectrum; for example, values ​​from low to high correspond to a color gradient from green and yellow to red. The final output is a heat map of power supply and demand imbalance risk with a geographic background and color depth representing the level of risk.

[0088] In some embodiments, after generating a heat map of power supply and demand imbalance risk, the system performs a feedback correction step. In a specific implementation, the system extracts several consecutive regions with the highest risk values ​​from the heat map; these regions are defined as highly sensitive load centers. The extraction process can be achieved by setting a risk threshold or selecting the top N regions by risk ranking. In another specific implementation, the system obtains typical daily load curves for highly sensitive load centers from a historical load database. These typical daily load curves are time-series data reflecting the electricity consumption habits of the region under specific seasons or workday types. In yet another specific implementation, morphological analysis is performed on the typical daily load curves. This morphological analysis includes, but is not limited to, wavelet decomposition, Fourier analysis, or moving average filtering. The aim is to separate the relatively stable base load component and the peak component with obvious peak-valley characteristics from the load curves, and to assess the potential for the peak component to be excited and transferred over time.

[0089] In practical implementation, the system uses the analyzed load base load component values ​​and peak shift potential quantification indicators as a new set of constraint vectors, which are then injected back into the input of the transaction price elasticity prediction model. After receiving these new constraints, the input of the transaction price elasticity prediction model performs a secondary correction on the parameters or functional relationships related to load elasticity within the model. This secondary correction aims to make the model's price fluctuation simulation more closely resemble the actual response behavior of highly sensitive load centers under extreme weather conditions. The nodal price fluctuation simulation program is then rerun using the corrected transaction price elasticity prediction model. The model outputs updated nodal marginal price predictions and price deviations, and the system uses these updated results to regenerate the transaction strategy reference matrix. Based on the newly generated transaction strategy reference matrix, the system can rerun the decision rule base matching and calculation process to update the previously generated price range recommendations.

[0090] In one embodiment of the present invention, the construction and operation of the transaction price elasticity prediction model are carried out as follows: From the heat map of power supply and demand imbalance risk, key transmission sections affecting power transmission are identified, and the risk levels borne by these sections are extracted, quantifying the risk level into a specific transmission capacity loss ratio. This transmission capacity loss ratio is used as the main input variable and input into the transaction price elasticity prediction model, which has a pre-built linkage logic reflecting the price transmission relationship between different regional power markets. The model's calculation program is executed, and the model outputs the predicted marginal electricity price of each power trading node for a series of consecutive future trading periods. The deviation of the predicted marginal electricity price of each node in each period from the benchmark electricity price under the risk-free scenario is calculated. The three dimensions of information—trading period, power trading node identification, and corresponding price deviation—are integrated to form a trading strategy reference matrix. The decision rule base matching mechanism generates a suggested price range based on the trading strategy reference matrix. This suggested price range includes a benchmark price curve and a floating fault tolerance threshold. The mechanism identifies trading periods from the trading strategy reference matrix where predicted price fluctuations exceed preset thresholds, marking these periods as critical trading windows requiring special attention. It then retrieves a decision rule base and, based on the disaster type in current extreme weather warnings, obtains a matching pricing strategy template. This template specifies the pricing behavior guidelines to be followed in different market scenarios, such as supply shortages or oversupply. Finally, the predicted price fluctuation range of the critical trading windows and real-time congestion information obtained from the power grid's operational status are filled into the corresponding placeholder variables in the obtained pricing strategy template.

[0091] When calculating specific parameters, the system reads the risk preference configuration item from the pricing strategy template, which defines three different pricing tendencies: aggressive, moderate, and conservative. The system dynamically adjusts the weights of these pricing tendency configuration items based on the predicted price volatility of key trading windows; generally, the larger the predicted volatility, the lower the weight of the aggressive tendency. Based on this dynamically adjusted weight, the pricing calculation formula in the pricing strategy template is instantiated to obtain a preliminary set of pricing base point values ​​and fluctuation range parameters. Next, grid congestion is used as a correction factor, superimposed on the initially calculated pricing base point values ​​for final fine-tuning, outputting the final suggested pricing base point values ​​and allowable fluctuation range parameters. Combining the calculated pricing base point value sequence with the allowable fluctuation range parameters forms a clear benchmark pricing curve and a fluctuation tolerance threshold, together constituting a complete pricing range suggestion. After generating the pricing range suggestion, it needs to be converted into a standardized order flow recognizable by the power trading platform. The system parses the benchmark pricing curve in the pricing range suggestion, decomposes it according to the format specified by the power trading platform, and converts it into a series of time-segmented electricity-price order point pairs. Simultaneously, the floating fault tolerance threshold is mapped to the upper and lower limits of the bid price allowed by the trading platform's interface protocol. Following the data message format specifications published by the power trading platform, the aforementioned time-segmented quantity-price bid pairs and the upper and lower limits of the bid price are encapsulated into a unified bid data packet structure. An integrity check code is calculated for this bid data packet structure, and a security key is used for digital signature and encryption to generate a standardized bid instruction stream that meets secure transmission requirements.

[0092] In practical implementation, the process of constructing a transaction price elasticity prediction model based on the heat map of power supply and demand imbalance risk begins with the extraction of risk information from key transmission sections. The system overlays a power grid topology layer onto the heat map of power supply and demand imbalance risk, identifies lines or sections with high risk values ​​in the map, and defines these lines or sections as key transmission sections. In practical implementation, the system reads the risk level value corresponding to the location of the key transmission section. The risk level value is a normalized scalar. The system quantifies the risk level value into a specific transmission capacity loss ratio according to a preset mapping table. For example, a risk level of 0.8 may correspond to a transmission capacity loss ratio of 40%. In practical implementation, the transmission capacity loss ratio is transmitted as the core input variable to the input end of the transaction price elasticity prediction model. The transaction price elasticity prediction model is a simulation model that incorporates market clearing logic and regional price transmission relationships. The transaction price elasticity prediction model internally pre-defines the price linkage logic between different regional markets. The price linkage logic is defined through the transmission capacity between regions, historical price correlation coefficients, and market rules.

[0093] In practical implementation, the price elasticity prediction model performs simulation calculations. The model uses the transmission capacity loss ratio as a condition for modifying grid transmission constraints. Combined with market-based data such as load forecasts and generator price curves, it runs optimal power flow or market clearing algorithms to output the predicted marginal electricity price for each power trading node within a series of consecutive trading periods. In practice, the system obtains benchmark electricity price data for the same period from an external database. The benchmark price is typically the predicted price under a scenario without extreme weather risks or the historical average price for the same period. The system calculates the deviation between the predicted marginal electricity price for each node in each trading period and the corresponding benchmark price. The deviation can be expressed as an absolute price difference or a relative percentage. In practice, the system integrates the trading period, power trading node identifier, and the calculated deviation to form a trading strategy reference matrix. The trading strategy reference matrix is ​​a structured data table, as shown in Table 1.

[0094] Table 1: Trading Strategy Reference Matrix

[0095]

[0096] In implementation, the decision rule base matching mechanism generates suggested price ranges based on the trading strategy reference matrix. These suggested ranges include a benchmark price curve and a floating tolerance threshold. The mechanism first traverses the trading strategy reference matrix, identifying trading periods and node combinations where the absolute value of the deviation exceeds a preset threshold, marking these combinations as key trading windows. The system then searches the decision rule base, which stores price strategy templates for different extreme weather types. The system matches and calls the corresponding price strategy template based on the current weather type warning code. These templates exist as structured text or executable code snippets, specifying the pricing behavior guidelines and calculation formulas under different supply and demand scenarios. In implementation, when calculating the price base value and floating range parameters, the system reads the risk preference configuration item defined in the price strategy template. This risk preference configuration item clearly distinguishes the mathematical expressions of three pricing tendencies: aggressive, moderate, and conservative. The system dynamically adjusts the configuration weights of the pricing tendencies based on the predicted price fluctuation range of the key trading windows; the greater the price fluctuation range, the higher the proportion of the configuration weights adjusted towards the moderate and conservative tendencies. Based on the dynamically adjusted weights, the system instantiates the calculation formula in the pricing strategy template to obtain a preliminary set of pricing base point values ​​and fluctuation range parameters. In some embodiments, the calculation formula for the pricing base point value integrates risk appetite and volatility:

[0097] ;

[0098] in: This represents the calculated base point value for the quote. This represents a benchmark price determined by historical market data or cost analysis. This represents the predicted price fluctuation range obtained from the trading strategy reference matrix. This represents the risk weighting factor dynamically determined by the risk preference allocation item. This represents a correction factor indicating the severity of congestion on the transmission path within the trading window, obtained from real-time grid operation data. The preliminarily calculated base price value is combined with the allowable fluctuation range parameter to form a complete price range recommendation. The benchmark price curve in the recommendation consists of a series of time-series arranged base price values. The fluctuation tolerance threshold specifies the allowed percentage or absolute amount of upward and downward fluctuation in the price for each time period.

[0099] In practical implementation, converting the proposed price range into a standardized declaration instruction stream recognizable by the power trading platform involves format encapsulation and security processing. The system parses the benchmark price curve in the proposed price range and, according to the segmented or continuous curve price format specified by the power trading platform, converts the curve into a series of time-segmented quantity-price declaration point pairs. Each point pair contains a time tag, a power consumption value, and a price value. Simultaneously, the system maps the floating fault tolerance threshold to the upper and lower limit parameter fields of the declaration price defined in the trading platform's application programming interface protocol. Following the publicly released data message format specifications of the power trading platform, the time-segmented quantity-price declaration point pair sequence and the upper and lower limit parameters of the declaration price are encapsulated into a unified declaration data packet structure. This declaration data packet structure typically includes a message header, a main data field, and a message body. In practical implementation, the system calculates a checksum for the complete declaration data packet structure, employing cyclic redundancy check or message digest algorithms, and uses a pre-installed security certificate to digitally sign and encrypt the data packet, generating a standardized declaration instruction stream that ultimately meets secure transmission requirements. Optionally, the encryption algorithm can use the national cryptographic standard algorithm or the RSA algorithm.

[0100] See Figure 4This is a chart showing the distribution of standardized order volume-price bid points and their upper and lower price limits. The price trend at the bid points exhibits a single-peak pattern, rising initially and then falling, reaching a peak of 980 yuan / MWh at 18:00, before gradually declining. From 14:00 to 18:00, the price climbed continuously from 850 yuan / MWh, reflecting an intensified supply-demand tension in the market; after 18:00, the price fell back to 860 yuan / MWh at 21:00, easing market pressure somewhat. The bid price reached its highest point of the day at 18:00, but exceeded the upper limit, which is not allowed in actual trading, indicating that the bidding strategy during this period needs close attention and adjustment. From 14:00 to 17:00, the price continued to rise, reaching 950 yuan / MWh at 17:00, hitting the upper price limit, reflecting the most prominent supply-demand imbalance during this period. By monitoring the relationship between the declared price and the upper and lower limits, periods of abnormal price fluctuations can be quickly identified. The comparison between the price trend and the upper and lower limits can be used to analyze the impact of factors such as extreme weather on the supply and demand relationship in the electricity market.

[0101] In one embodiment of the present invention, after the standardized declaration instruction stream is pushed to the trading interface, the system enters the post-execution monitoring and learning phase. The system continuously monitors the data stream returned by the trading interface, including acknowledgment messages confirming receipt of the declaration instructions and real-time transaction data streams generated after market clearing. The system parses the real-time transaction data stream, extracts the actual transaction price information, and compares this actual transaction price with the predicted price of the corresponding time period and node when the trading strategy reference matrix was previously generated, calculating the prediction deviation value between the two. This prediction deviation value, along with the input conditions on which the prediction was based, is used as a new set of training samples and fed back to the trading price elasticity prediction model. The trading price elasticity prediction model uses this feedback sample to initiate its model parameter fine-tuning program to reduce prediction deviations in similar scenarios in the future. After the parameter fine-tuning is completed, the performance of the trading price elasticity prediction model is iteratively updated and will be applied to the analysis and processing of the next round of extreme weather warning cycles, thereby achieving self-optimization of the decision-making model.

[0102] In practice, after the standardized declaration command stream is pushed to the trading interface, the system initiates a continuous monitoring process. This process maintains communication with the power trading platform's trading interface through a pre-defined data interface. The system continuously monitors and receives two types of key data streams returned from the trading interface: declaration confirmation receipts, which are confirmation messages returned by the trading platform after verifying the format and legality of the received standardized declaration command stream; and real-time transaction data streams, released after the market clearing cycle, containing detailed transaction records of each market participant. The system parses the received real-time transaction data streams, extracting the actual transaction price information corresponding to the previous declaration period and power trading node. This actual transaction price information includes the transaction volume and unit price. In practice, the system retrieves the trading strategy reference matrix generated in the previous analysis period from local storage or cache. This matrix stores the predicted price fluctuation range and the predicted marginal electricity price for the corresponding trading period and power trading node. The system compares the extracted actual transaction price with the predicted marginal electricity price of the same trading period and the same electricity trading node in the trading strategy reference matrix point by point, calculating the prediction deviation value for each data point. The prediction deviation value quantifies the difference between the predicted price and the actual market price. In some embodiments, the prediction deviation value... The calculation can be performed in the form of relative error, expressed by the formula:

[0103] ;

[0104] in: This represents the actual transaction price parsed from the real-time transaction data stream. This represents the predicted marginal electricity price of the corresponding node obtained from the trading strategy reference matrix. This represents the benchmark electricity price corresponding to that node. It can be understood as the prediction deviation value. It can be positive, negative, or zero, representing that the actual price is higher than predicted, lower than predicted, or consistent with predicted, respectively.

[0105] In practical implementation, the system will calculate the prediction deviation value. Together with the input dataset used to generate this prediction, these data are packaged into a new set of training samples. The input dataset includes, but is not limited to, the power supply and demand imbalance risk heatmap data, load forecast data, and network topology status used when generating the original trading strategy reference matrix. This set of training samples is fed back into the training module of the trading price elasticity prediction model. The training module uses the feedback samples to fine-tune the model's internal parameters. The goal of parameter fine-tuning is to make the model's output more closely approximate the actually observed market price when encountering similar input conditions. Optionally, parameter fine-tuning can employ gradient descent or its variants, iteratively updating the model parameters by minimizing the loss function composed of prediction deviations. After parameter fine-tuning, the internal price formation logic and risk transmission coefficients of the trading price elasticity prediction model are updated, and the system saves and registers this updated model version. In some embodiments, when the meteorological monitoring network issues a new round of extreme weather warnings and the system initiates the next analysis cycle, it will automatically invoke the parameter-fine-tuned transaction price elasticity prediction model and apply it to the new analysis process. The parameter-fine-tuned transaction price elasticity prediction model will conduct a new round of nodal electricity price fluctuation simulation based on the electricity supply and demand imbalance risk heat map derived from the new original meteorological warning dataset, thus forming a closed-loop learning process from decision execution to effect feedback and model self-optimization. This process allows the transaction price elasticity prediction model to continuously adapt and adjust as market operation data accumulates, improving its accuracy in price prediction during subsequent extreme weather events.

[0106] See Figure 5 This is a chart showing the predicted fluctuations in nodal electricity prices. The predicted electricity price is the marginal electricity price forecast output by the transaction price elasticity prediction model, reflecting the anticipation of market prices. The actual electricity price, the price at which transactions take place, is a true reflection of market supply and demand. The benchmark electricity price, a reference price under the scenario of no extreme weather risk, is usually the historical average price for the same period or the risk-free forecast value, and is stable at 500 yuan / MWh in the chart. The predicted electricity price shows a significant peak at the 8th hour, far exceeding the benchmark price, and then generally declines, falling below the benchmark price after the 17th hour. Two significant peaks occur at the 4th hour and the 8th-9th hour, with overall fluctuations smaller than the predicted electricity price, and also falling below the benchmark price after the 17th hour. For most of the period, both the predicted and actual electricity prices are higher than the benchmark price, reflecting the pressure of supply falling short of demand in the market; after 15 hours, the price falls back below the benchmark, easing the supply-demand imbalance.

[0107] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A power trading auxiliary decision processing method based on extreme weather, characterized in that, include: Extreme weather warning messages containing meteorological element codes, disaster level indicators, and effective time windows are obtained from the meteorological monitoring network to form the original meteorological warning dataset; Spatiotemporal benchmark alignment is performed on the original meteorological warning dataset to generate a grid map of extreme weather impact areas in a unified coordinate system. The grid map of extreme weather impact areas includes a gridded meteorological intensity matrix and an index of the distribution of power generation facilities within the coverage area. Influence factor mapping is performed on the meteorological intensity matrix to obtain a meteorological sensitivity coefficient matrix for different types of power generation resources. Topological association is performed on the power generation facility distribution index to obtain a power generation asset vulnerability association table. The meteorological sensitivity coefficient matrix and the vulnerability correlation table of power generation side assets are input into the risk transmission calculation engine for coupled simulation to generate a heat map of power supply and demand imbalance risk. The risk level of each key transmission section is extracted from the heat map of power supply and demand imbalance risk, and the risk level is quantified as the proportion of transmission capacity loss. Based on the aforementioned transmission capacity loss ratio, a trading strategy reference matrix is ​​generated using a trading price elasticity prediction model. Based on the trading strategy reference matrix, a pricing range suggestion for a specific power trading cycle is generated through a decision rule base matching mechanism, including: the pricing range suggestion includes a benchmark pricing curve and a floating fault tolerance threshold; The trading periods in which price fluctuations exceed a preset threshold are identified from the trading strategy reference matrix and marked as key trading windows; The decision rule base is retrieved to obtain a pricing strategy template that matches the current extreme weather type. The pricing strategy template specifies the pricing behavior guidelines under scenarios of supply shortage or supply surplus. The predicted price fluctuation range and power grid congestion status of the key trading window are filled into the placeholders of the quotation strategy template to calculate the specific quotation base point value and the allowed fluctuation range parameter. The base price value is combined with the allowable fluctuation range parameter to form the benchmark price curve and the floating tolerance threshold, which together constitute the price range recommendation. The proposed price range is converted into a standardized declaration instruction stream that can be recognized by the power trading platform, and the standardized declaration instruction stream is pushed to the trading interface.

2. The power trading auxiliary decision processing method based on extreme weather as described in claim 1, characterized in that, Performing influence factor mapping on the meteorological intensity matrix yields meteorological sensitivity coefficient matrices for different types of power generation resources, including: Retrieve a pre-established mapping dictionary between power generation resource types and meteorological elements. The mapping dictionary defines the key meteorological influencing factors corresponding to wind turbine generators, photovoltaic arrays, and conventional thermal power units. Traverse each grid cell in the meteorological intensity matrix and extract the intensity value of each meteorological element within the grid cell; Based on the mapping dictionary, the extracted meteorological element intensity values ​​are substituted into the sensitivity functions of the corresponding power generation resource types for normalization, and the sensitivity scores of the grid cells for different power generation resource types are output. The sensitivity scores of all grid cells are arranged and combined according to the power generation resource type to construct a multi-dimensional meteorological sensitivity coefficient matrix.

3. The power trading auxiliary decision processing method based on extreme weather as described in claim 2, characterized in that, Perform a topological association on the distributed index of the power generation facilities to obtain a vulnerability association table for power generation side assets, including: Obtain the equipment health status file and geographical location coordinates of each power generation unit recorded in the power generation facility distribution index; Using geographic information system tools, the spatial intersection relationship between the geographic location coordinates and the boundary of the grid map of the extreme weather impact area is calculated, and a set of power generation units within the impact range is selected. By reading the historical fault frequency and current maintenance status from the equipment health status file and combining them with the type of the corresponding power generation unit, a comprehensive vulnerability index is calculated. The set of power generation units and their corresponding comprehensive vulnerability indices are bound together as key-value pairs to generate the vulnerability association table of the power generation side assets.

4. The power trading auxiliary decision processing method based on extreme weather as described in claim 3, characterized in that, The meteorological sensitivity coefficient matrix and the vulnerability correlation table of power generation assets are input into the risk transmission calculation engine for coupled simulation to generate a heat map of power supply and demand imbalance risk, including: In the risk transmission calculation engine, a blank risk accumulation matrix corresponding to the power grid topology is initialized; Based on the vulnerability association table of the power generation side assets, the grid node where the high-risk power generation unit is located is located, and the power output attenuation coefficient of the grid node is adjusted according to the meteorological sensitivity coefficient matrix. Using the adjusted power output attenuation coefficient as input, the propagation path and amplification effect of the power deficit at the power grid node in the power grid are simulated in the power grid power flow calculation model, and the propagation results are accumulated to the corresponding position of the blank risk accumulation matrix. Once the simulation calculations for all affected nodes are completed, the risk accumulation matrix is ​​rendered using color gradation to output a visualized heat map of the power supply and demand imbalance risk.

5. The power trading auxiliary decision processing method based on extreme weather as described in claim 4, characterized in that, The step of generating a trading strategy reference matrix based on the transmission capacity loss ratio and the trading price elasticity prediction model includes: The transmission capacity loss ratio is used as an input variable and input into the transaction price elasticity prediction model. The transaction price elasticity prediction model has a pre-set price linkage logic between different regional markets. Execute the aforementioned transaction price elasticity prediction model to output the predicted marginal electricity price of each electricity trading node within a series of future trading periods; The deviation of the predicted marginal electricity price of the node from the benchmark electricity price is calculated, and the trading period, node identifier and corresponding deviation are integrated into the trading strategy reference matrix.

6. The power trading auxiliary decision processing method based on extreme weather as described in claim 5, characterized in that, The process involves filling the predicted price fluctuation range and grid congestion status of the key trading window into the placeholders of the pricing strategy template, and calculating the specific pricing base point value and the allowed fluctuation range parameters, including: Read the risk preference configuration item from the pricing strategy template, which defines three pricing tendencies: aggressive, moderate, and conservative. Based on the magnitude of the predicted price fluctuations in the key trading window, the allocation weight of the pricing tendency is dynamically adjusted; the greater the fluctuation, the lower the weight of the aggressive approach. Based on the adjusted weights, the pricing calculation formula in the pricing strategy template is instantiated to obtain a set of preliminary pricing base values ​​and fluctuation ranges; The grid congestion situation is added as a correction factor to the initial bid base value, and the final bid base value and the allowable fluctuation range parameter are output.

7. The power trading auxiliary decision processing method based on extreme weather as described in claim 6, characterized in that, The proposed price range is converted into a standardized order flow that can be recognized by the power trading platform, including: The benchmark price curve in the proposed price range is analyzed and converted into time-segmented quantity-price declaration point pairs as specified by the power trading platform. The floating fault tolerance threshold is mapped to the upper and lower limits of the bid price allowed by the trading platform; In accordance with the data message format specifications of the trading platform, the time-segmented volume and price declaration point pairs and the upper and lower limits of the declaration price are encapsulated into a unified declaration data packet structure; The declaration data packet structure is subjected to checksum calculation and signature encryption to generate the standardized declaration instruction stream that meets the requirements for secure transmission.

8. The power trading auxiliary decision processing method based on extreme weather as described in claim 7, characterized in that, After generating the aforementioned heatmap of power supply and demand imbalance risk, the following steps are also included: The areas most severely affected by extreme weather were extracted from the aforementioned heat map of power supply and demand imbalance risk and defined as high-sensitivity load centers. Obtain the typical daily load curve of the high-sensitivity load center, and perform morphological analysis on the typical daily load curve to extract the base load component and peak shift potential of the load. The base load component and peak shift potential of the load are used as constraints and back-injected into the input of the transaction price elasticity prediction model to perform a second correction on the transaction price elasticity prediction model. The trading strategy reference matrix is ​​regenerated using the modified trading price elasticity prediction model to update the quote range recommendations.

9. The power trading auxiliary decision processing method based on extreme weather as described in claim 8, characterized in that, After pushing the standardized declaration instruction stream to the transaction interface, the process also includes: Continuously monitor the declaration confirmation receipt and real-time transaction data stream returned by the transaction interface; The actual transaction price in the real-time transaction data stream is compared with the predicted price fluctuation range in the previously generated trading strategy reference matrix, and the prediction deviation value is calculated. The prediction deviation value is fed back to the transaction price elasticity prediction model as part of the model training samples, and the parameters of the transaction price elasticity prediction model are fine-tuned. The adjusted transaction price elasticity prediction model was then applied to the analysis of the next extreme weather cycle.

10. A power trading auxiliary decision processing system based on extreme weather, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the power trading auxiliary decision processing method based on extreme weather as described in any one of claims 1 to 9.

Images