New energy power market price limit optimization method and system based on dynamic partitioning

Abstract

The application provides a new energy power market price limit optimization method and system based on dynamic zoning, relates to the field of power market, and comprises the following steps: based on new energy clustering and physical constraint fusion, dynamic zoning is carried out on a given power grid region; taking the upper and lower limits of the price limit of each zoning period as the decision variable, a dynamic zoning price limit model is constructed; a feature vector is constructed for the predicted regional new energy output in the period, and based on the feature vector, the price prediction probability distribution function of each zoning is obtained through LSTM-Transformer; the price prediction probability distribution function of each zoning is used to solve the dynamic zoning price limit model, and the upper and lower limits of the price limit of each zoning period are obtained. The application designs a dynamic zoning method for the power market, and proposes a generation and optimization method for differentiated price limits for the dynamic zoning, so as to realize sufficient connection of the market mechanism and regional differences, and improve the resource allocation efficiency of the market.

Description

Technical Field

[0001] This invention relates to the field of electricity markets, specifically to a method and system for optimizing price limits in the new energy electricity market based on dynamic zoning. Background Technology

[0002] The power system exhibits significant regional characteristics in terms of energy categories and load distribution, but current regulatory boundaries primarily follow administrative divisions, leading to conflicts at both the policy and technical levels.

[0003] At the policy level, "market clearing can be based on factors such as grid topology, congestion characteristics, and resource distribution to divide price zones, reflecting the differences in the spatial and temporal value of electricity," which provides a top-level policy basis for the zoning mechanism that breaks through administrative boundaries in electricity market regulation. Meanwhile, from a practical perspective, spot market pilot programs in Jiangsu and the Southern Region have already implemented "zoning pricing." The Jiangsu spot market scheme requires "establishing a time-of-use zoning pricing mechanism for all electricity volume," with zoning based on grid congestion sections rather than administrative regions. The Southern Region adopts "joint spot market clearing," dynamically identifying congestion sections to form zoning prices, achieving inter-provincial electricity mutual assistance.

[0004] On a technical level, fluctuations in renewable energy sources will render the "administrative zoning" approach to power regulation ineffective. For example, during a midday surge in solar power generation, northern Jiangsu (a renewable energy-rich area) supplies electricity to the load center in southern Jiangsu. However, if northern Anhui simultaneously experiences power shortages, the traditional "province-based" dispatching system cannot automatically utilize surplus power from northern Jiangsu to support Anhui, requiring manual intervention. Dynamic zoning can merge northern Jiangsu and northern Anhui into a "high demand zone" in real time, allowing negative electricity prices to stimulate cross-provincial consumption, while southern Jiangsu, as a "load center area," maintains a price ceiling. Furthermore, fixed and rigid zoning will exacerbate grid congestion and curtailment. For instance, the wind power base in Jiuquan, Gansu, experienced a curtailment rate exceeding 15% in 2023 due to blocked transmission channels. If Jiuquan and the Hexi Corridor energy storage cluster were designated as a "negative price zone," it could stimulate local energy storage charging, while Lanzhou, as an independent "load zone," would avoid being affected by negative electricity prices.

[0005] Aside from the regional zoning mechanism, a current regulatory gap that is out of touch with actual needs lies in fixed price limits. While renewable energy accounts for over 30% of China's installed capacity, its fluctuating output leads to significant price volatility in the spot market (e.g., frequent occurrences of zero or capped electricity prices in Shandong and Shanxi). The existing fixed price cap is ill-suited to the characteristics of renewable energy, easily causing market distortions or insufficient investment incentives. The fixed upper and lower price limits result in renewable energy-rich regions (such as Gansu and Ningxia) frequently experiencing the lower limit being breached (when wind and solar power are booming, supply exceeds demand, requiring negative electricity prices to stimulate consumption, but a fixed lower limit of 0 yuan cannot achieve this), while traditional energy-rich regions (such as Shanxi) face the problem of the upper limit suppressing investment (traditional power sources require high electricity prices to recover capacity costs, but the unified upper limit is too low, such as 1.0 yuan / kWh in Shanxi).

[0006] Current research on dynamic regional pricing for power grids and electricity markets has emerged, but most studies focus on physical grid structures (such as inter-provincial sections) for regionalization, neglecting the impact of renewable energy fluctuations on electrical coupling and real-time power flow. This can lead to mismatches between regional and real-time power flows, exacerbating local congestion and wasting inter-regional regulation capacity. Furthermore, current research rarely addresses dynamic pricing mechanisms for different regions, and most studies fail to adequately consider the relationship between renewable energy fluctuation characteristics and regional characteristics. Summary of the Invention

[0007] To address the aforementioned issues, this invention proposes a method and system for optimizing price limits in the new energy power market based on dynamic zoning. It designs a method for dynamic zoning of the power market and proposes a method for generating and optimizing differentiated price limits for dynamic zoning, so as to achieve full integration of market mechanisms and regional differences and improve the efficiency of market resource allocation.

[0008] According to some embodiments, the present invention adopts the following technical solution: A dynamic partitioning-based method for optimizing the price ceiling in the renewable energy market includes: Based on the fusion of new energy clustering and physical constraints, a given power grid area is dynamically partitioned. A dynamic zonal price limit model is constructed using the upper and lower limits of price limits for each zone and time period as decision variables; To predict regional time-of-use renewable energy output, feature vectors are constructed. Based on these feature vectors, the probability distribution function for electricity price prediction in each region is obtained using LSTM-Transformer. By using the probability distribution function of electricity price prediction for each zone, the dynamic zone price limit model is solved to obtain the upper and lower limits of the price limit for each zone and time period.

[0009] According to some embodiments, the present invention adopts the following technical solution: A new energy power market price limit optimization system based on dynamic partitioning includes: The dynamic partitioning module is configured to dynamically partition a given power grid area based on the fusion of new energy clustering and physical constraints. The model building module is configured to: build a dynamic zonal price limit model using the upper and lower limits of the price limit for each zone and time period as decision variables; The function construction module is configured to: construct feature vectors for predicting regional time-segmented renewable energy output, and based on the feature vectors, use LSTM-Transformer to obtain the probability distribution function of electricity price prediction for each region; The price limit solution module is configured to use the probability distribution function of electricity price prediction for each zone to solve the dynamic zone price limit model and obtain the upper and lower limits of the price limit for each zone and time period.

[0010] According to some embodiments, the present invention adopts the following technical solution: A computer program product includes a computer program that, when executed by a processor, implements the aforementioned dynamic partitioning-based new energy power market price limit optimization method.

[0011] According to some embodiments, the present invention adopts the following technical solution: A non-transitory computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the aforementioned method for optimizing the price limit of the new energy power market based on dynamic partitioning.

[0012] According to some embodiments, the present invention adopts the following technical solution: An electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the dynamic partitioning-based new energy power market price limit optimization method.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention adopts a three-domain coupling mechanism of "new energy fluctuations - power grid physics - market game theory", which solves the problem of traditional models that separate physical constraints and market behavior, leading to a disconnect between price limits and system state. The given price limit range can significantly control the risk of electricity price fluctuations. It can improve the new energy absorption rate, reduce the congestion costs between regions, and thus improve the overall social welfare. Attached Figure Description

[0014] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0015] Figure 1 This is a flowchart of the new energy power market price limit optimization method based on dynamic partitioning in Example 1. Detailed Implementation The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0016] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0017] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0018] Example 1 One embodiment of the present invention provides a method for optimizing the price ceiling of the new energy electricity market based on dynamic partitioning, such as... Figure 1 As shown, it includes: Step S1: Based on the fusion of new energy clustering and physical constraints, dynamically partition the given power grid area; Step S2: Construct a dynamic zonal price limit model using the upper and lower limits of price limits for each zone and time period as decision variables; Step S3: Construct feature vectors for predicting regional time-segmented renewable energy output. Based on the feature vectors, use LSTM-Transformer to obtain the probability distribution function of electricity price prediction for each region. Step S4: Using the probability distribution function of electricity price prediction for each zone, solve the dynamic zone price limit model to obtain the upper and lower limits of the price limit for each zone and time period.

[0019] As one example, this embodiment proposes a dynamic zoning electricity price limit optimization method based on new energy volatility clustering and cross-sectional rigid constraints. The specific implementation process is as follows: I. Dynamic Partitioning of the Electricity Market Based on the Fusion of New Energy Clustering and Physical Constraints As mentioned earlier, with the increase in installed capacity of new energy sources, the fluctuation of new energy output has become a key variable for measuring the regional differences in the power grid. The fluctuation of new energy output can measure the differences in resource abundance in different regions (such as new energy-rich regions and traditional energy-rich regions), and can also indirectly reflect the price fluctuation range of the corresponding regional market.

[0020] Therefore, taking the volatility of renewable energy sources as the core variable for measuring regional similarity, for a given grid region, starting from the substation level, it is divided into several initial candidate regions through grid topology and management level. Then, in a single candidate region i, the normalized variable of its renewable energy volatility is: (1) In the above formula, This represents the predicted renewable energy output of candidate region i during time period t. This represents the actual output of renewable energy in candidate region i during time period t; Represents a time series of a given length The maximum value.

[0021] For two candidate regions i and j, the similarity of their renewable energy volatility is measured by Dynamic Time Warping (DTW), as shown in the following formula: (2) in, This represents the optimal alignment time window. The event segments for the new energy output data of the two regions.

[0022] By performing DTW on all candidate regions, the similarity of renewable energy volatility between each pair of candidate regions can be obtained. A clustering objective function is then constructed with the goal of minimizing the sum of the similarities of renewable energy volatility between all candidate regions and the cluster center. (3) In the above formula, This represents the k-th cluster set; Indicates the total number of clusters; This represents the center of the k-th cluster set to be determined; This indicates the minimum number of partitions set to prevent overfitting; it is typically set to 3.

[0023] By solving (3), an initial partitioning scheme can be obtained, denoted as: (4) in, This represents the total number of initial partitions obtained using the new energy volatility clustering algorithm.

[0024] The above zoning scheme only considers the similarity of renewable energy volatility and does not incorporate physical constraints. Therefore, the relationship between regions and cross-sections needs to be examined. For two candidate regions i and j, if they exhibit drastically different electricity price responses to power changes at the same cross-section m, they should not be classified into the same region. Therefore, a sensitivity index for regional electricity prices and cross-sectional power is constructed, specifically: (5) in, This represents the electricity price in region i; This represents the power at the critical section m; This indicates the rated power of the critical section m.

[0025] Based on the above sensitivity indicators, the following segmentation discrimination conditions can be set. (6) In the above formula, These are the sensitivity index values ​​for candidate regions i and j and the same cross section m, respectively. This represents the cross-sectional sensitivity threshold; a recommended value is 0.05.

[0026] For candidate regions i and j, if they are assigned to... If two elements are in the same partition but satisfy the conditions of equation (6), then they need to be split into their own separate partitions. All candidate regions in the dataset are subjected to the above tests, and the final partition is denoted as: (7) II. Dynamic Zoned Price Limit Model After partitioning a given power grid region, it is necessary to formulate corresponding price-limiting strategies for each partition. To this end, the following dynamic partitioned price-limiting model is established, with the objective of minimizing wind and solar curtailment and the probability (price difference exceeding the threshold event): (8) In the above formula, the decision variable is ,in This represents the lower limit of the price limit for region z during time period t. This represents the upper limit of the price limit for region z during time period t; Indicates the total number of time periods considered in the decision-making process; This represents the clearing price of region z during time period t; This represents the reference electricity price for region z during time period t (which can be represented by the historical average). Indicates the threshold for electricity price fluctuations; This represents the amount of renewable energy wasted in region z during time period t; and This represents the weighting coefficient of the two items; This indicates the probability that electricity price fluctuations exceed a threshold.

[0027] In the above model, It can be used to measure the risk of electricity market price fluctuations. To prevent market speculation or financial risks caused by sharp fluctuations in electricity prices, the following constraints can be set. (9) in, This indicates the risk tolerance level, with a typical value of 5%.

[0028] This represents the loss item for renewable energy consumption, and the corresponding amount of abandoned electricity. It can be represented as (10) in, This represents the available renewable energy output of region z during time period t; This represents the actual renewable energy output of region z during time period t.

[0029] To reduce the amount of electricity wasted from renewable energy sources, a compensation mechanism should be established for the price-limiting policy. (11) in, This represents the compensation coefficient, triggered when the curtailment rate exceeds a threshold (typically 5%). The above formula indicates that when the curtailment rate is too high, the lower limit of the price ceiling should exceed 0, acknowledging negative electricity prices, which can promote energy storage charging.

[0030] At the same time, from Starting from there, it's not difficult to deduce the upper limit of the price limit. have (12) in, This represents the feature vector constructed to predict the renewable energy output of region z during time period t; This represents the inverse function of the cumulative probability distribution function (CDF) of regional electricity prices. Expressing the request Conditional probability Corresponding independent variable .

[0031] In order to find The feature vectors will be given in the next section. And the corresponding probability distribution of electricity price predictions.

[0032] III. Electricity Price Forecast Probability Generation First, we give the feature vector in the following form: (13) In the above formula, This indicates the projected output of new energy sources; Indicates the standard deviation of new energy output; Indicates the peak-to-valley ratio of the load; Indicates the available capacity of flexible resources; Indicates the available power transmission rate across regions; Indicates the severity of line congestion in the area; The ratio of buyer to seller bids in the local electricity market.

[0033] After providing the feature vector form, we use LSTM to extract its temporal features. The internal process of LSTM will not be elaborated here. Let the extracted temporal feature vector be: (14) The Gaussian Mixture Model (GMM) is constructed to determine the probability distribution of electricity prices. (15) In the above formula, M represents the total number of distribution types in the mixed model; Let be the parameters of the m-th type distribution function. This represents the weight of the m-th type distribution function. and The mean and variance of the electricity price distribution function for type m; The parameter is and It follows a normal distribution.

[0034] In order to obtain In its specific form, the Transformer model is adopted, and the attention mechanism is as follows: (16) In the above formula, This refers to the query of a Transformer, which can be understood as an examination. The relationship between medium characteristics and electricity prices; This represents the key in a Transformer, i.e., the transpose of the key-value pair. It can be understood as... Identification attributes of features; express The specific value of the feature; This represents the scaling factor for the key value.

[0035] At the same time, in order to achieve The parameters of each distribution function are used to design the output layer of the Transformer. (17) In the above formula, T still represents the transpose of the vector; the superscript symbol represents the parameter in formula (15). , , The estimate; express Hidden layer weight network coefficients; This indicates the corresponding drift amount.

[0036] The above methods can be obtained The initial parameters belong to the forward propagation process of the neural network; to improve prediction accuracy, the following loss function is given for training the model: (18) loss function It consists of three parts, namely, negative log-likelihood loss. Physical constraint loss function Prior guidance loss function , and These are the corresponding weight parameters.

[0037] The negative log-likelihood loss aims to approximate the predicted distribution to the actual electricity price through data fitting; its specific form is as follows: (19) In the above formula, This represents the actual electricity price.

[0038] Physical constraint loss function The specific form is (20) In the above formula, L represents the total number of lines within the examined zone; This indicates the actual power carried by line l; This indicates the maximum allowable power flow of line l; This represents the overload penalty term; N represents the total number of nodes in the partition under examination; This represents the power imbalance penalty term for node n.

[0039] Prior guidance loss function The specific form is (twenty one) In the above formula, Let be the parameters of the m-th type distribution function. This represents the Nash equilibrium electricity price in the electricity market corresponding to the examined region.

[0040] exist According to the loss function After training, its backpropagation process can be represented as follows: (twenty two) in, Indicates the network parameters at step t; Indicates the learning rate; This represents the gradient operator.

[0041] In the above LSTM- Once the training is complete, the specific form of the probability distribution function for electricity prices can be obtained.

[0042] IV. Solving for Dynamic Zoning Price Limits The previous section derived the GMM model (i.e., the probability distribution of generation) for the regional electricity price. For region z, it is denoted as... At the same time, risk parameters are set. ,in, Indicates electricity price Exceeding the price limit The probability threshold, i.e. the risk tolerance in equation (9); Indicates electricity price Below the lower limit of the price limit The probability threshold is typically 10%.

[0043] Knowing Under the condition of a specific function form and It can be solved by the following formula (twenty three) Furthermore, considering the compensation mechanism for renewable energy consumption, when the curtailment rate of zone z exceeds a threshold, the negative electricity price restriction will be relaxed. The following adjustments will be made: (twenty four) Through the above steps, not only was the electricity market (corresponding to the power grid) dynamically partitioned, but differentiated price limits were also generated for different partitions.

[0044] Example 2 One embodiment of the present invention provides a new energy power market price limit optimization system based on dynamic partitioning, comprising: The dynamic partitioning module is configured to dynamically partition a given power grid area based on the fusion of new energy clustering and physical constraints. The model building module is configured to: build a dynamic zonal price limit model using the upper and lower limits of the price limit for each zone and time period as decision variables; The function construction module is configured to: construct feature vectors for predicting regional time-segmented renewable energy output, and based on the feature vectors, use LSTM-Transformer to obtain the probability distribution function of electricity price prediction for each region; The price limit solution module is configured to use the probability distribution function of electricity price prediction for each zone to solve the dynamic zone price limit model and obtain the upper and lower limits of the price limit for each zone and time period.

[0045] Example 3 One embodiment of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned method for optimizing the price limit of the new energy power market based on dynamic partitioning.

[0046] Example 4 In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided for storing computer instructions. When the computer instructions are executed by a processor, they implement the aforementioned method for optimizing the price limit of the new energy power market based on dynamic partitioning.

[0047] Example 5 One embodiment of the present invention provides an electronic device, including: a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes the method for optimizing the price limit of the new energy power market based on dynamic partitioning.

[0048] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0049] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0050] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for optimizing the price ceiling in the new energy power market based on dynamic partitioning, characterized in that, include: Step S1: Based on the fusion of new energy clustering and physical constraints, dynamically partition the given power grid area; Using the volatility of renewable energy sources as the core variable for measuring regional similarity, for a given grid region, starting from the substation level, it is divided into several initial candidate regions based on grid topology and management level. In a single candidate region i, the normalized variable for the volatility of renewable energy sources is: In the above formula, This represents the predicted renewable energy output of candidate region i during time period t. This represents the actual output of renewable energy in candidate region i during time period t; Represents a time series of a given length The maximum value; A sensitivity index for the relationship between regional electricity prices and cross-sectional power is constructed, specifically as follows: in, This represents the electricity price in region i; This represents the power at the critical section m; Indicates the rated power of the critical section m; Based on the sensitivity index, two candidate partitions in the same initial partition are cut and determined to obtain the final partition; Step S2: Construct a dynamic zonal price limit model using the upper and lower limits of price limits for each zone and time period as decision variables; The dynamic zoning price limit model aims to minimize wind and solar curtailment and its probability. in, This is the final set of partitions obtained from dynamic partitioning. For partitioned sets The partition in the middle, the decision variable is , This represents the lower limit of the price limit for region z during time period t. This represents the upper limit of the price limit for region z during time period t; Indicates the total number of time periods considered in the decision-making process; This represents the clearing price of region z during time period t; This represents the reference electricity price for region z during time period t; Indicates the threshold for electricity price fluctuations; This represents the amount of renewable energy wasted in region z during time period t; and This represents the weighting coefficient of the two items; This indicates the probability that electricity price fluctuations exceed a threshold. To measure the risk of price volatility in the electricity market, the following constraints are set: in, Indicates risk tolerance; This represents the loss item for renewable energy consumption, and the corresponding amount of abandoned electricity. Represented as in, This represents the available renewable energy output of region z during time period t; This represents the actual renewable energy output of region z during time period t; To reduce the amount of electricity wasted from renewable energy sources, a compensation mechanism should be established for the price-limiting policy to determine the lower limit of the price limit. have: in, This represents the compensation coefficient, and the triggering condition is that the curtailment rate exceeds a threshold. The above formula means that when the curtailment rate is too high, the lower limit of the price limit should break through 0, acknowledge negative electricity prices, and promote energy storage charging. from Starting from there, we arrive at the upper limit of the price limit. have: in, This represents the feature vector constructed to predict the renewable energy output of region z during time period t; This represents the inverse function of the cumulative probability distribution function of regional electricity prices. Expressing the request Conditional probability Corresponding independent variable ; Step S3: Construct feature vectors for predicting regional time-segmented renewable energy output. Based on the feature vectors, use LSTM-Transformer to obtain the probability distribution function of electricity price prediction for each region. The feature vector is expressed by the formula: in, This indicates the projected output of new energy sources; Indicates the standard deviation of new energy output; Indicates the peak-to-valley ratio of the load; Indicates the available capacity of flexible resources; Indicates the available power transmission rate across regions; Indicates the severity of line congestion in the area; The ratio of buyer to seller bids in the local electricity market; This represents the clearing price of region z during the time period t-1; Temporal features are extracted from the constructed feature vector using LSTM. : To construct a Gaussian mixture model for the unknown electricity price probability distribution, the following is done: in, Indicated by electricity price A Gaussian mixture model is constructed for the input variables, where M represents the total number of distribution types in the mixture model; Let be the parameters of the m-th type distribution function. This represents the weight of the m-th type distribution function. and The mean and variance of the electricity price distribution function for type m; The parameter is and The normal distribution; Based on the time-series characteristics, the parameters of the electricity price prediction probability distribution function are solved using a Transformer to obtain the electricity price prediction probability distribution function for each region. The output layer of the Transformer is: Here, T still represents the transpose of the vector; the superscript sign indicates the Gaussian mixture model parameters. , , The estimate; express Hidden layer weight network coefficients; This indicates the corresponding drift amount; The loss function of LSTM-Transformer is: in, For negative log-likelihood loss, The physical constraint loss function, The prior guiding loss function, and These are the corresponding weight parameters; Negative log-likelihood loss To make the predicted distribution approximate the actual electricity price through data fitting, its specific form is as follows: in, This represents the total number of distribution types in the mixture model; Let be the parameters of the m-th type distribution function. This represents the weight of the m-th type distribution function. and The mean and variance of the electricity price distribution function for type m; The parameter is and The normal distribution This represents the actual electricity price; Physical constraint loss function The specific form is: Where L represents the total number of lines within the examined zone; This indicates the actual power carried by line l; This indicates the maximum allowable power flow of line l; This represents the overload penalty term; N represents the total number of nodes in the partition under examination; This represents the power imbalance penalty term for node n; Prior guidance loss function The specific form is: in, This represents the total number of distribution types in the mixture model. Let be the parameters of the m-th type distribution function. This represents the Nash equilibrium electricity price in the electricity market corresponding to the examined region; Step S4: Using the probability distribution function of electricity price prediction for each zone, solve the dynamic zone price limit model to obtain the upper and lower limits of the price limit for each zone and time period.

2. The method for optimizing the price limit of new energy power market based on dynamic partitioning as described in claim 1, characterized in that, The specific steps for dynamically partitioning a given power grid area are as follows: Starting from the substation level, a given power grid area is divided into several initial candidate areas based on the power grid topology and management level; Based on the volatility of new energy sources in candidate regions, initial partitioning is obtained through new energy clustering; Based on the relationship between regions and cross sections, physical constraints are incorporated to optimize the initial zoning scheme and obtain the final zoning.

3. The method for optimizing the price limit of new energy power market based on dynamic partitioning as described in claim 2, characterized in that, The new energy clustering is specifically as follows: The similarity of renewable energy volatility between candidate regions is calculated by using dynamic time warping (DTW). Clustering is performed with the objective of minimizing the sum of similarities in new energy volatility between all candidate regions and cluster centers to obtain the initial partitions.

4. A new energy power market price limit optimization system based on dynamic partitioning, characterized in that, The method for optimizing the price ceiling of the new energy power market based on dynamic partitioning as described in any one of claims 1-3 includes: The dynamic partitioning module is configured to dynamically partition a given power grid area based on the fusion of new energy clustering and physical constraints. The model building module is configured to: build a dynamic zonal price limit model using the upper and lower limits of the price limit for each zone and time period as decision variables; The function construction module is configured to: construct feature vectors for predicting regional time-segmented renewable energy output, and based on the feature vectors, use LSTM-Transformer to obtain the probability distribution function of electricity price prediction for each region; The price limit solution module is configured to use the probability distribution function of electricity price prediction for each zone to solve the dynamic zone price limit model and obtain the upper and lower limits of the price limit for each zone and time period.

5. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the new energy power market price limit optimization method based on dynamic partitioning as described in any one of claims 1-3.

6. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the dynamic partitioning-based new energy power market price limit optimization method as described in any one of claims 1-3.

7. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the dynamic partitioning-based new energy power market price limit optimization method as described in any one of claims 1-3.

Images