Cross-provincial power grid resource dynamic planning method, device, equipment and medium

Abstract

The present application relates to the technical field of power grid resource planning, and provides a cross-provincial power grid resource dynamic planning method, device, equipment and medium. The implementation scheme is: the uncertainty state of new energy output and load demand driven by the cross-regional coordination state vector of each first time is dynamically scene constructed, and a time scale scene set and a scene probability set are obtained; a multi-agent game model is constructed based on the cross-regional coordination state vector of each first time, the time scale scene set and the scene probability set, and the multi-agent game model is game optimized to obtain the resource planning scheme of each power grid agent; the first income in the resource planning scheme of each power grid agent is reconstructed to obtain the second income of each power grid agent; and the power grid resources of each power grid agent are rolled based on the resource planning scheme and the second income of each power grid agent. The embodiment of the present application can improve the power grid resource configuration efficiency.

Description

Technical Field

[0001] This invention relates to the field of power grid resource planning technology, and in particular to a method, apparatus, equipment and medium for dynamic planning of cross-provincial power grid resources. Background Technology

[0002] With the increasing demand for inter-provincial power trading, the coupling relationships between transmission channels, regulation resources, load demand, and distributed energy sources involved in power grid operation are becoming increasingly complex. Existing technologies typically employ a one-time planning approach based on fixed scenarios for inter-regional power grid resource allocation, using historical operating data to statically optimize transmission capacity, intra-provincial resource dispatch, and renewable energy consumption. In these methods, the allocation results for various resources are mostly calculated uniformly based on preset operating boundaries.

[0003] However, in actual operation, due to the random fluctuations in the output of new energy sources and the time-varying characteristics of load demand, coupled with the resource competition and profit game relationship between different power grid entities, the existing static resource allocation method results in low power grid resource allocation efficiency, making it difficult to meet the application requirements for dynamic collaborative optimization of resources in complex power grid environments. Summary of the Invention

[0004] This invention provides a method, apparatus, equipment, and medium for dynamic planning of cross-provincial power grid resources, which can solve at least one of the above-mentioned technical problems.

[0005] In a first aspect, embodiments of the present invention provide a method for dynamic planning of cross-provincial power grid resources, including: The multi-source power grid resource data of multiple power grid entities at each first moment are structured and constructed to obtain the cross-regional collaborative state vector at each first moment; Dynamic scenarios are constructed for the uncertain states of new energy output and load demand driven by the cross-regional collaborative state vector at each first moment, resulting in a time-scale scenario set and a scenario probability set. A multi-agent game model is constructed based on the cross-regional collaborative state vectors at each first moment, the time-scale scenario set, and the scenario probability set. The multi-agent game model is then optimized through game theory to obtain resource planning schemes for each of the power grid entities. The first benefit in the resource planning scheme of each of the power grid entities is reconstructed to obtain the second benefit of each of the power grid entities; Based on the resource planning schemes and secondary benefits of each of the aforementioned power grid entities, rolling planning is carried out on the power grid resources of each of the aforementioned power grid entities.

[0006] Secondly, embodiments of the present invention provide a cross-provincial power grid resource dynamic planning device, comprising: The structured construction module is used to construct the multi-source power grid resource data of multiple power grid entities at each first moment in a structured manner, so as to obtain the cross-regional collaborative state vector at each first moment. The dynamic scenario construction module is used to dynamically construct scenarios for the uncertainty of new energy output and load demand driven by the cross-regional collaborative state vector at each first moment, and obtain a time-scale scenario set and a scenario probability set. The game optimization module is used to construct a multi-agent game model based on the cross-regional collaborative state vectors at each first moment, the time scale scenario set, and the scenario probability set, and to perform game optimization on the multi-agent game model to obtain the resource planning schemes of each of the power grid entities. The reconstruction module is used to reconstruct the first benefit in the resource planning scheme of each of the power grid entities to obtain the second benefit of each of the power grid entities; The rolling planning module is used to perform rolling planning of the power grid resources of each of the power grid entities based on the resource planning scheme and the second benefit of each of the power grid entities.

[0007] Thirdly, embodiments of the present invention also provide a computer device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method described in any one of the embodiments of the present invention.

[0008] Fourthly, embodiments of the present invention also provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the method described in any one of the embodiments of the present invention.

[0009] The technical solution of this invention firstly constructs a structured representation of multi-source power grid resource data from multiple power grid entities at various first-moment intervals, obtaining cross-regional collaborative state vectors for each first-moment interval. Secondly, it dynamically constructs scenarios for the uncertainties in renewable energy output and load demand driven by these cross-regional collaborative state vectors, obtaining time-scale scenario sets and scenario probability sets. This enables the system to dynamically predict and probabilistically describe renewable energy fluctuations, load changes, and future operating trends, thereby improving the adaptability of resource scheduling to complex operating environments. Then, based on the cross-regional collaborative state vectors, time-scale scenario sets, and scenario probability sets, a multi-entity game model is constructed. Through game optimization, resource planning schemes for each power grid entity are obtained, enabling dynamic collaborative decision-making among different power grid entities. This reduces resource competition conflicts and local resource redundancy issues present in traditional independent planning methods, improving cross-regional resource collaborative allocation capabilities. Furthermore, the first benefit in the resource planning scheme is reconstructed to obtain the second benefit. That is, by coordinating and optimizing the benefit relationships between different power grid entities, the problem of interest imbalance in the multi-entity collaboration process can be reduced. Finally, based on resource planning schemes and secondary benefits, rolling planning is implemented for the power grid resources of each power grid entity, enabling the system to continuously update resource allocation results according to the latest operating status within each decision-making cycle. By dynamically optimizing and adjusting transmission channel resources, resource idleness and scheduling conflicts are reduced, thereby improving the efficiency of power grid resource allocation.

[0010] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0011] The accompanying drawings are provided for a better understanding of this solution and do not constitute a limitation of the invention. Wherein: Figure 1 This is a flowchart of a cross-provincial power grid resource dynamic planning method according to an embodiment of the present invention; Figure 2 This is a structural block diagram of a cross-provincial power grid resource dynamic planning device according to an embodiment of the present invention; Figure 3 This is a schematic block diagram of a computer device used to implement the methods of the embodiments of the present invention. Detailed Implementation

[0012] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0013] This invention provides a method, apparatus, device, and medium for dynamic planning of cross-provincial power grid resources. The executing entity of this method can be the cross-provincial power grid resource dynamic planning apparatus provided in this invention, or a computer device integrating the apparatus. The apparatus can be implemented in hardware or software, and the computer device can be a terminal or a server.

[0014] Figure 1 This is a flowchart of a cross-provincial power grid resource dynamic planning method according to an embodiment of the present invention.

[0015] like Figure 1 As shown, this cross-provincial power grid resource dynamic planning method may include: S110: The multi-source power grid resource data of multiple power grid entities at each first moment are structured and constructed to obtain the cross-regional collaborative state vector at each first moment. S120, dynamically constructs the uncertain state of new energy output and load demand driven by the cross-regional collaborative state vector at each first moment, and obtains the time-scale scenario set and scenario probability set; S130: Based on the cross-regional collaborative state vectors, time-scale scenario sets, and scenario probability sets of each first moment, a multi-agent game model is constructed, and the multi-agent game model is optimized through game theory to obtain the resource planning schemes of each power grid entity. S140, reconstruct the first benefit in the resource planning scheme of each power grid entity to obtain the second benefit of each power grid entity; S150, based on the resource planning schemes and secondary benefits of each power grid entity, conducts rolling planning of the power grid resources of each power grid entity.

[0016] For example, a power grid entity refers to a power grid operating unit with independent decision-making power and revenue accounting capabilities in an inter-provincial electricity market. Examples include inter-regional transmission operators (responsible for the operation, maintenance, and capacity allocation of inter-provincial transmission lines), provincial power grid entities (responsible for load balancing and safe dispatching within the province), and distributed resource aggregation entities (responsible for source-load resource interaction and local optimization).

[0017] For example, multi-source power grid resource data refers to a heterogeneous data set encompassing physical transmission channel operation characteristic data and provincial regional node supply and demand status data. The transmission channel operation characteristic data includes topology parameters, line impedance parameters, thermal stability limits, and dynamic capacity information; the provincial regional node supply and demand status data includes load demand information, renewable energy output information, and real-time electricity price signals.

[0018] For example, the unit of the first moment can be an hour, 15 minutes, or 5 minutes, depending on the frequency of electricity market transactions.

[0019] For example, the cross-regional collaborative state vector refers to a unified state variable that reflects the operating capacity of transmission channels and the supply and demand status of regional nodes, constructed by performing feature extraction and time-series alignment processing on multi-source power grid data. It is used to establish a unified situational awareness foundation in multi-agent games.

[0020] In this example, the process of constructing the cross-regional collaborative state vector is as follows: obtain power grid operation data (such as power and voltage), meteorological information (such as wind speed and sunshine) and market transaction data, and map the physical structure of the transmission channel (topology, impedance, thermal stability limit) and the regional node operation information (load demand, renewable energy output) to a unified state space through a preset feature mapping function.

[0021] In this example, the expression for the cross-regional collaborative state vector is as follows: ; In the formula, is the cross-regional collaborative state vector at the first time t; T is the set of original data time indices (such as a timestamp sequence sampled every 5 minutes); The alignment weight for mapping from the original time k to the target time t is determined by both the time interval and data reliability; L is the set of transmission channels (such as the set of inter-provincial interconnection lines); N is the set of regional nodes (such as the set of hub substations in a provincial power grid). This is the raw data set of channel l at time k (such as line topology, impedance, thermal stability limit, and current transmission power). This is the raw data set of node n at time k (such as real-time load power, wind power output, photovoltaic power output, and node voltage). , These are the corresponding feature mapping functions; , These are the corresponding weighting coefficients (reflecting the importance of channel operation safety and node supply and demand balance to the system status). The mean of the cross-regional state sequence; The standard deviation of the cross-regional state sequence.

[0022] For example, the study area is set to include two core power transmission channels between Province A and Province B. At a certain moment... , obtained through calculation An example numerical vector (the dimension of which is determined by the number of channels and the number of nodes): Where T is the transpose; the numerical vector represents the following from top to bottom: Channel 1 open circuit status (1.0 indicates normal connection), Channel 1 load rate (0.8 indicates 80% load), Channel 2 open circuit status (0.8 indicates minor fault warning), Channel 2 power flow (0.0 indicates no power flow at present), Channel 2 fault repair progress (1.5 indicates maintenance and optimization status), A province node voltage amplitude (220.5), B province renewable energy output ratio (50.2%), regional grid access fee coefficient (0.65), Channel 1 geographical correlation (0.42), Channel 2 geographical correlation (0.42), power coupling coefficient between Channel 1 and Channel 2 (0.38), and overall system stability index (0.90, the closer the value is to 1, the better the system stability).

[0023] By standardizing and persistently storing the aforementioned vectors, the decision-making system can capture real-time operating conditions where Channel 1 is running normally and under high load, while Channel 2 currently has low power flow but exhibits geographical correlation and the risk of localized failures. This quantitative expression method allows the decision-making system to directly access the overall system state rather than scattered raw data, providing a unique, reliable, and traceable system state input basis for multi-agent game decision-making.

[0024] It should be noted that feature mapping functions are used to transform high-dimensional, multi-source heterogeneous data into unified feature values ​​that reflect the key operational capabilities of a system. For channel data, the feature mapping function... The original data set of transmission channel l Using this as input, through feature extraction, normalization, and dimensionality compression, the output is a unified feature quantity that directly reflects the transmission capacity, safety status, and available capacity of the channel, making the physical attributes of different channels comparable and computable. The feature mapping function for node-side data... The original data set of region node n Using the input as input, through feature extraction, time-series smoothing and standardization processing, the output is a unified feature quantity that can accurately characterize the power supply and demand surplus and shortage, power fluctuation level and system regulation support capability of the node, thereby realizing a refined characterization of the real-time supply and demand status of the node.

[0025] For example, the time-scale scenario set refers to the set of evolutionary paths with temporal correlation formed after the uncertainties are hierarchically expanded under different planning periods such as years, months, and days. The scenario probability set refers to the probability distribution of the occurrence of the above-mentioned scenario paths, calculated based on a dynamic evolution probability model, and is used to quantify the risks of uncertain environments.

[0026] For example, a resource planning scheme refers to the inter-provincial channel capacity allocation strategy, intra-provincial power generation and consumption plan, and distributed resource dispatch instructions obtained through game theory optimization under the premise of satisfying inter-provincial power transmission constraints, provincial safe operation constraints, and distributed resource regulation constraints.

[0027] For example, inter-regional power transmission operators determine the capacity allocation ratio and grid access fee pricing strategy for different inter-provincial transmission channels at various time periods based on the operational status of inter-provincial transmission channels and the demand for renewable energy transmission; provincial power grid operators determine the output plan of thermal power units, the energy storage charging and discharging plan, and the provincial power flow dispatch scheme based on the allocated inter-provincial transmission capacity and the load demand within the province; and distributed resource aggregation entities determine the distributed photovoltaic output control strategy, the adjustable load response strategy, and the user-side energy storage operation strategy based on real-time electricity price signals and dispatch instructions. This results in a resource planning scheme that includes inter-provincial transmission resources, provincial regulation resources, and distributed source-load resources.

[0028] For example, the first benefit refers to the original economic benefit obtained by each power grid entity in solving the game model based solely on its own interest maximization objective before utility coordination is carried out.

[0029] For example, the second benefit refers to the equilibrium economic benefits reconstructed through benefit redistribution and compensation rules after the introduction of a utility coordination mechanism, which is used to balance the individual rationality of each subject with the overall optimization of the system and to achieve cross-regional resource synergy incentives.

[0030] According to the above implementation method, by structuring multi-source power grid resource data from multiple power grid entities, a cross-regional collaborative state vector is formed, achieving a unified representation of the cross-regional power grid operation status. Furthermore, based on the cross-regional collaborative state vector, dynamic scenarios are constructed for the uncertainties in renewable energy output and load demand, resulting in a time-scale scenario set and a scenario probability set. On this basis, a multi-entity game model is constructed and optimization is performed to obtain resource planning schemes for each power grid entity, enabling different power grid entities to make collaborative decisions while considering their own interests and system operation requirements, thereby reducing resource conflicts and redundancy. Further, by reconstructing the first benefit in the resource planning scheme, a second benefit is obtained, achieving benefit coordination and optimization among different power grid entities, thereby improving the stability of resource collaborative scheduling among various entities. Finally, rolling planning of power grid resources is performed based on the resource planning scheme and the second benefit, enabling the system to continuously and dynamically adjust resource allocation results according to real-time operating status, thereby improving power grid resource allocation efficiency.

[0031] In one implementation, dynamic scenario construction is performed on the uncertainty state of new energy output and load demand driven by the cross-regional collaborative state vector at each first moment, resulting in a time-scale scenario set and a scenario probability set. This includes: for the cross-regional collaborative state vector at each first moment, summing the first product of the cross-regional collaborative state vector at the first moment and a preset influence coefficient, the second product of the historical uncertainty variable at the second moment and a preset inertia coefficient, and a preset random disturbance term at each time scale to obtain the joint uncertainty variable at each time scale at the first moment; coupling the joint uncertainty variable at each time scale at each first moment with multiple time scales to obtain the target uncertainty variable at each time scale at each first moment; associating and recombining the target uncertainty variable at each time scale at each first moment according to the time sequence to obtain the time-scale scenario set; for each time-scale scenario in the time-scale scenario set, determining the conditional probability of the time-scale scenario at each first moment based on the target uncertainty variable at each first moment in the time-scale scenario; and multiplying the conditional probabilities of each time-scale scenario at each first moment to obtain the scenario probability set.

[0032] For example, the joint uncertainty variable refers to the comprehensive uncertainty state quantity that characterizes the fluctuations in new energy output and load demand. This variable is no longer an isolated random value, but a system state variable with a time evolution structure driven by a cooperative state vector and taking into account historical inertia and random disturbance terms.

[0033] For example, in intraday scheduling scenarios, this variable reflects the combined effect of fluctuations in photovoltaic and wind power output and deviations in user-side electricity demand.

[0034] In this example, for the cross-regional collaborative state vector at each first moment, at each preset time scale, the product of the current collaborative state vector and the influence coefficient, the product of the historical uncertainty variable and the inertia coefficient at the previous moment (i.e., the second moment), and the random disturbance term are linearly weighted and summed to characterize the dynamic evolution of uncertainty.

[0035] In this example, the calculation expression for the joint uncertainty variables is as follows: ; In the formula, Let be the joint uncertainty variables at the first time point t; This is the cooperative state vector; The coefficient representing the influence of the current power grid state on uncertain variables (range 0-1). This represents the uncertain state of the previous moment; The inertia coefficient for time evolution (characterizing the continuity of different time periods); This is a random disturbance term (used to characterize unpredictable short-term fluctuations, which follow a zero-mean distribution).

[0036] For example, in a scenario of stable operation within the day, set , random disturbance term If the state at the previous moment... Current cooperative state vector Then the calculation yields Through this calculation, the model can capture the dynamic trends of continuous changes in new energy sources and load over time.

[0037] It should be noted that the values ​​of various parameters required in the calculation of joint uncertainty variables can be determined through historical sample regression, grid search, or cross-validation. For example, the influence coefficient is preferably 0.45; the inertia coefficient can be 0.40-0.80, preferably 0.65; the random disturbance term can be set as a zero-mean Gaussian disturbance; and the standard deviation is 0.5-1.0 times the standard deviation of the historical prediction error.

[0038] For example, a time-scale scenario set refers to a set of scenario paths with temporal correlations formed by incorporating different planning cycles such as years, months, and days into the same evolutionary framework and coupling them across scales. This set achieves a synergistic mapping between short-term fluctuations and long-term trends.

[0039] In this example, the uncertainty at each time scale is defined as follows: By using a pre-defined cross-scale coupling function, the uncertainty representations at different time dimensions are weighted and summed according to the scale influence weights to obtain the target uncertainty variable. Then, based on the time sequence, the target uncertainty variables at each time point are correlated and recombined to form a scenario path covering the decision-making cycle.

[0040] In this example, the computational expression for multi-timescale coupling is as follows: ; In the formula, For the target uncertainty variable under the time scale T; For the set of all time scales; The weight of the influence of time scale T' on time scale T (reflecting the constraint of long-term trends on short-term fluctuations and the feedback of short-term fluctuations on long-term trends).

[0041] It should be noted that the cross-scale influence weights can be determined by normalizing the inverse of the historical forecast error, and the sum of the influence weights under the same target scale is 1. For example, the influence weight of the monthly scale on the intraday scale is 0.30, and the influence weight of the day-ahead scale on the intraday scale is 0.70.

[0042] For example, if you need to build an intraday scheduling scenario, It includes both monthly and daily scales. The system will apply long-term constraint weights to intraday fluctuations based on monthly trends. This method integrates the monthly forecast deviation of electricity demand with the actual daily fluctuations. This mapping ensures that short-term decisions do not deviate from long-term planning goals.

[0043] For example, suppose the decision-making cycle consists of four periods (T=4), and the set of timescales includes intraday scales ( ) and day-ahead scale ( At each first moment t ( The corresponding target uncertainty variables have been obtained through the aforementioned coupling calculations. and .

[0044] During the recombination process, the system concatenates state variables at different time scales according to their time series: for the first possible evolutionary path... : At time t=1, select the state variable value. , ; At time t=2, select the state variable values. , ; At time t=3, select the state variable values. , ; At time t=4, select the state variable values. , .

[0045] Through the above association and recombination, the scene sequence corresponding to this path is obtained as follows: .

[0046] Similarly, according to the random disturbance term With different values, the system can construct multiple similar scene sequences. Ultimately, all scene sequences that satisfy the evolutionary logic are aggregated to form a time-scale scene set. Each one Each represents a complete time path from the initial moment to the end of the decision cycle, thus reorganizing the originally discrete, multi-scale data into an evolutionary trajectory with temporal correlation, providing a clear and evolvable scenario input for subsequent game optimization.

[0047] In this example, the expression for calculating the scene probability set is as follows: ; In the formula, For the scene The probability of occurrence; The conditional probability is determined by the evolutionary process; For the scene The conditional probability at time t (i.e., the scenario) (The target uncertainty variable at time t). During calculation, the complex continuous random process is discretized into a finite number of computable scenario paths by multiplying the conditional probabilities at each time step in the scenario, providing deterministic probabilistic inputs for subsequent multi-agent game decisions.

[0048] Among them, the conditional probability can be obtained by normalizing the target uncertainty variable through Softmax or estimating the Gaussian kernel density, and the probability of all candidate scenarios is normalized at each first time step so that the sum of the probabilities of each scenario is 1, thus avoiding the ambiguity of the units caused by equating the "target uncertainty variable" with the probability.

[0049] According to the above implementation method, for the cross-regional collaborative state vector at each first moment, the current operating state, historical uncertainty variables, and random disturbance terms are fused at different time scales to obtain joint uncertainty variables. This enables unified modeling of the dynamic characteristics of new energy fluctuations and load changes, improving the ability to characterize random changes in complex operating environments. Furthermore, the joint uncertainty variables at each time scale are coupled across multiple time scales to obtain target uncertainty variables, enabling the synergistic fusion of relationships between different time scales and improving the continuity and consistency of the uncertainty state description. Subsequently, the target uncertainty variables are correlated and reorganized according to time sequence to obtain a set of time-scale scenarios, forming a multi-scenario operating path with temporal evolution relationships, thereby improving the dynamic prediction capability of future operating states. Based on this, the corresponding conditional probabilities are determined based on the target uncertainty variables in each time-scale scenario, and the conditional probabilities at each first moment are multiplied to obtain a set of scenario probabilities. This enables a quantitative assessment of the probability of different evolutionary scenarios, improving the system's ability to characterize uncertain operating risks.

[0050] In one implementation, a multi-agent game model is constructed based on the cross-regional collaborative state vectors, time-scale scenario sets, and scenario probability sets at each first moment. The multi-agent game model is then optimized through game theory to obtain resource planning schemes for each power grid entity. The multi-agent game model includes a first decision sub-model, a second decision sub-model, and a third decision sub-model. The first decision sub-model includes: determining the external operating conditions of the cross-regional transmission operator under different time-scale scenarios based on the cross-regional collaborative state vectors, time-scale scenario sets, and scenario probability sets at each first moment; constructing a global revenue function for the cross-regional transmission operator based on the channel operation revenue, grid access fee revenue, and resource scheduling cost under each time-scale scenario in the time-scale scenario set; constructing a first decision sub-model with the goal of maximizing the global revenue function; solving the first decision sub-model through game theory based on the constraints of the first decision sub-model to obtain the channel capacity allocation strategy and price signal; and optimizing the channel capacity allocation strategy, price signal, and the operating states of each power grid entity through hierarchical linkage game theory to obtain resource planning schemes for each power grid entity.

[0051] For example, a multi-agent game model refers to a dynamic interactive model in a hierarchical decision-making system composed of inter-regional power transmission operators, provincial power grid operators, and distributed resource aggregation operators, in which each entity pursues its own profit maximization while achieving global utility equilibrium through price signals, capacity allocation strategies, and operational constraints.

[0052] It should be noted that this model essentially maps the problem of inter-provincial channel resource allocation in the power system into a sequential game process involving multiple agents in an uncertain scenario. By achieving Nash equilibrium through strategic responses between levels, it ensures that each agent achieves the global optimality of system operation while satisfying its own operational constraints.

[0053] In this example, the multi-agent game model can be defined as a three-layer master-slave game structure: the upper-level decision-making sub-model (leader): with the inter-regional power transmission operator as the core, decides on the inter-provincial channel capacity allocation and grid access fee pricing signals based on the goal of maximizing global revenue; the middle-level decision-making sub-model (follower / coordinator): with the provincial power grid operator as the core, with the goal of minimizing the operating cost within the province, receives the upper-level pricing signals and optimizes the output of generating units, load regulation, and reserve resource allocation within the province; the lower-level decision-making sub-model (subordinate responder): with the distributed resource aggregation operator as the core, with the goal of maximizing its own revenue, responds to the middle-level dispatch instructions and real-time electricity prices, and adjusts the output of distributed power sources and the electricity consumption behavior of controllable loads.

[0054] For example, consider a game-theoretic scenario: a high grid access fee signal from an upper-level entity will prompt the mid-level provincial power grid entity to reduce its demand for inter-provincial electricity, thereby increasing the output of its domestic generating units or guiding the lower-level aggregation entity to respond on the load side. In responding to electricity price fluctuations, the load-cutting behavior of the lower-level entities will generate a "peak shaving and valley filling" effect, reducing the overall congestion penalty cost of the system. This iterative and feedback process through pricing and output behavior enables the system to automatically filter out unreasonable resource allocation schemes, ultimately converging to a set of equilibrium strategies that satisfy the physical operational constraints and profit maximization conditions of each entity.

[0055] For example, external operating conditions refer to the boundary environment constraints faced by cross-regional power transmission operators during cross-provincial power grid dispatching, including channel physical carrying capacity constraints mapped from cross-regional collaborative state vectors, uncertainty risk exposure embodied by scenario sets, and safety margin requirements determined by power grid topology.

[0056] For example, the first decision sub-model refers to a higher-level decision-making model that aims to maximize the overall expected revenue of inter-regional power transmission operators and coordinates the allocation of transmission capacity with the formulation of pricing mechanisms. This model serves as the guiding layer of a three-layered interconnected game, transmitting economic signals to middle and lower-level entities by optimizing the allocation of transmission resources.

[0057] In this example, the global revenue function consists of the sum of basic transmission revenue from inter-regional power transmission, electricity market transaction price difference revenue, and renewable energy consumption revenue, minus system operation and dispatch costs. Its calculation expression is as follows: ; In the formula, The upper-level entity expects overall benefits; For the scene The probability of; This is the function for the total revenue of the entire network in the given scenario; The function is the cost function for channel operation and scheduling; This is the decision vector for channel capacity allocation and network access fee signals.

[0058] In this example, the cost function This includes fixed operating costs, variable transmission costs, congestion penalty costs, and safety breach penalty costs. Among these, congestion penalty costs... Defined as: ; In the formula, For the scene The actual transmission power of the lower channel l; This represents the thermal stability limit of channel l; This is the congestion penalty coefficient. The congestion penalty coefficient can be determined based on the scheduling reconfiguration cost corresponding to the unit over-limit power, with an example of 1000-5000 yuan / megawatt (MW), preferably 3000 yuan / MW; if the channel is a critical section, this coefficient can be increased to 5000 yuan / MW.

[0059] For example, the constraints that the first decision sub-model needs to satisfy include: First, power balance constraints: ensuring dynamic balance between load demand at each node and renewable energy supply and inter-provincial transmission power under various scenarios; second, channel capacity constraints: limiting the power of each channel within the rated thermal stability limit, i.e. Third, node voltage safety constraints: Ensure that the voltage fluctuation range of each hub node is within the allowable safe operating range, i.e. Fourth, non-negative price constraint: requires that the grid access fee and transmission price signals always be non-negative in order to comply with market regulations.

[0060] For example, It is the cross-regional power transmission operator in the scenario The following refers to all operational and scheduling costs incurred in executing channel capacity configuration and price signal formulation.

[0061] ; These are fixed costs, specifically channel depreciation, operation and maintenance, and management costs that do not change with scheduling. ; This is a variable cost, meaning the cost varies depending on channel power and capacity allocation. ; For the scene Lower passage Transmission power, For channel Variable cost per unit power It sums up all transmission channels; Safety costs are the penalties for exceeding voltage limits and failing to meet stability standards. ; It is the voltage at node n. Rated voltage, Voltage deviation penalty factor; the voltage deviation penalty factor can be determined according to the safety correction cost corresponding to each 1% increase in voltage deviation beyond the rated voltage, for example, 500-2000 yuan. The allowable range for node voltage is 0.95-1.05 per unit.

[0062] It is a scene The total revenue generated by channel resource allocation and price signals across the entire network includes transmission revenue, market transaction revenue, and revenue from renewable energy consumption: ; The basic benefits generated from inter-regional power transmission: ; For the scene Lower passage Grid access fee / transmission price; For the profit from the price difference in electricity market transactions: ; Node n's transaction volume; It sums the sums of all region nodes. It is a scene Below, the electricity price at node n, For the scene Below, the electricity purchase price at node n, Given a set of regional nodes, summing all nodes yields the total transaction revenue of the entire network market.

[0063] To absorb the subsidies / carbon revenue generated by wind and solar power: ; Contribute to new energy sources at node n The unit consumption revenue coefficient; Maximize the upper layer At that time, the physical constraints of the power grid, operational safety constraints, and market constraints must be met, among which the power balance constraint is: In the formula, It is a scene Below, the load power of node n, i.e., the total electricity demand, It is a scene Below, the renewable energy power generation capacity of node n, For the scene The total transmission power of all inter-provincial power transmission channels. This formula represents the net load demand after deducting local renewable energy consumption from the total load of all network nodes. It also represents the power balance constraint, ensuring the scenario... The system power supply and demand are balanced. Channel capacity constraint is In the case where the channel power does not exceed the thermal stability limit, the node voltage constraint is: Price nonnegativity constraint The system stability constraint is , Refers to system stability indicators.

[0064] For example, the solution can be obtained using nonlinear programming strategies such as Sequential Quadratic Programming (SQP), interior-point methods, or particle swarm optimization. Through iterative optimization, the channel capacity allocation strategy and price signal that satisfy the optimal conditions in each scenario are output as boundary inputs for the mid-level decision-making process.

[0065] In this example, the solution process for the multi-scenario nonlinear programming of the first decision sub-model is illustrated by using a sequential quadratic programming algorithm for iterative optimization. The specific solution implementation process is as follows: First, in each scenario... Under probability constraints, for the objective function of maximizing global profit... We construct the corresponding Lagrange function. By introducing Lagrange multipliers, we explicitly integrate various power balance constraints, channel capacity constraints, node voltage security constraints, and market nonnegativity constraints into the Lagrange function, thereby achieving dimensionality reduction for the constrained optimization problem.

[0066] Secondly, in each iteration step At this point, use the current state decision variables A second-order Taylor expansion of the Lagrange function is performed, and the nonlinear constraints are linearized, thus constructing a quadratic programming (QP) subproblem. Solving this subproblem determines the search direction vector. This vector represents the factors contributing to improving the overall system benefit. Channel capacity allocation strategy and optimal correction increment of price signals.

[0067] Then, the step size factor is determined using a linear search technique. And update the decision variables: During this process, the system strictly performs constraint checks to ensure that the updated channel allocation strategy aligns with price signals. Meet the physical safety limits of the power grid (such as the thermal stability limit of the channel and the range of node voltage fluctuations) to prevent the power grid system from collapsing due to excessive optimization.

[0068] Finally, monitor the global profit function. The improved gradient and the optimality deviation under the Karush-Kuhn-Tucker Conditions (KKT) conditions. When both the iteration step size and the profit increment are lower than the preset convergence threshold (e.g. When the algorithm converges, it is determined that the algorithm has converged.

[0069] For example, the algorithm convergence condition can also be set to reach the maximum number of iterations, which can be 200.

[0070] For example, if the system measures that the remaining available capacity of inter-provincial channel 1 is large within a certain decision-making cycle, but the market electricity price difference revenue is low, the SQP algorithm will automatically guide the price signal by calculating the gradient of the objective function during the solution process. The algorithm shifts towards a more economically efficient adjustment range while increasing the capacity allocation weight of channel 1. After several iterations, the algorithm finally outputs a channel capacity allocation strategy and price signal that satisfy the optimality condition. This serves as the boundary condition for the mid-level provincial power grid entities to redistribute resources within the province, thereby achieving the globally optimal allocation of cross-regional power transmission resources in a multi-entity game environment.

[0071] In one implementation, a hierarchical linkage game optimization is performed on the channel capacity allocation strategy, price signals, and the operating status of each power grid entity to obtain the resource planning scheme corresponding to each power grid entity. This includes: constructing the operating cost function of the provincial power grid entity under different time scale scenarios based on the channel capacity allocation strategy, price signals, the provincial load demand, renewable energy output, and the operating status of regulation resources corresponding to the provincial power grid entity; constructing a second decision sub-model with the goal of minimizing the operating cost function; solving the second decision sub-model through game theory based on the constraints of the second decision sub-model to obtain the provincial resource allocation strategy; and performing a hierarchical linkage game optimization on the provincial resource allocation strategy, price signals, and the operating status of the power grid entity to obtain the resource planning scheme corresponding to each power grid entity.

[0072] For example, the second decision sub-model refers to a resource allocation model that coordinates unit power generation, power flow regulation, and demand response within the region with the goal of minimizing the main operating cost of the provincial power grid. As the intermediate layer of the three-level linkage decision-making, this model is mainly responsible for receiving inter-regional dispatch signals from the upper level and transforming them into resource allocation strategies that can be executed by the provincial power grid.

[0073] In this example, the provincial-level power grid operation cost function not only covers conventional power generation costs but also includes grid losses and system reserve costs. Its calculation expression is as follows: ; In the formula, For provincial power grid entities in time-scale scenarios The power grid operation cost function under the following conditions; Let n be the generator cost coefficient for node n. For the scene The active power generation of the conventional generator set connected to the next node n (this power value is dynamically driven by the load demand in the province and the output of new energy). Network loss cost (calculated by combining unit output, load demand, and channel switching power); System backup costs (determined by adjusting resource operating status).

[0074] For example, the active power generation of a conventional generator set The driving mechanism: This power is driven in real time by the provincial power balance equation.

[0075] For node n, its power generation must satisfy: ; In the formula, Let n be the load demand of node n. Contribute to distributed renewable energy; This represents the net power exchanged between node n and the inter-provincial power transmission channel. This is due to constraints imposed by the channel capacity allocation strategy issued by the upper layer. The adjustable range requires dynamic adjustment of each unit in the provincial decision-making model. This is to fill the supply and demand gap between load and new energy sources, so that the unit's power generation cost can evolve in real time with changes in channel strategy and load fluctuations.

[0076] Network loss cost Based on power flow distribution and line resistance calculations, it is found that there is a nonlinear mapping relationship between the power output, load demand, and channel switching power. The calculation expression is as follows: ; In the formula, The price is for compensation for network damage. Let be the electrical conductance of the line between nodes i and j; , Scenes Lower node voltage; This refers to the phase angle difference. Because the switching power of the transmission channels (controlled by upper-level capacity strategies) alters the overall power flow direction and intensity, network loss costs change accordingly, reflecting the impact of inter-regional power transmission on intra-provincial network losses.

[0077] System backup cost This is used to quantify the system's resilience to fluctuations in renewable energy sources and load deviations. Its calculation expression is as follows: ; In the formula, , These are the unit cost coefficients for upper and lower reserves, respectively; , These are the reserve capacities configured for node n. The operational status of regulating resources (such as the depth of energy storage charging and discharging, and the unit regulation rate) limits the upper limit of the reserve capacity, making the reserve cost directly dependent on the regulatory redundancy reserved by the provincial power grid main body in response to cross-regional strategy changes.

[0078] In this example, by modeling the functions of the aforementioned cost items, the upper-level channel capacity allocation strategy and price signals are explicitly embedded into the cost structure of the provincial power grid, ensuring the resource allocation strategy within the province. In the process of seeking the best in a game, it can automatically converge towards the direction that satisfies the optimal global benefit.

[0079] It should be noted that the upper-level channel capacity allocation strategy determines the "maximum power received or transmitted" allowed for each regional node in this scenario, while the price signal directly affects the marginal procurement cost of the provincial power grid, influencing unit output plans and load regulation decisions, and thus transmitting to the operating cost function.

[0080] For example, the provincial resource allocation strategy refers to the set of optimization decisions made by the provincial power grid main body in order to respond to inter-regional dispatch signals and balance local supply and demand, under the premise of meeting the power system operation safety constraints, regarding various types of power generation resources, adjustable load resources and inter-provincial power exchange behaviors within the province.

[0081] For example, during periods of high renewable energy generation, provincial resource allocation strategies may include: reducing the planned output of conventional thermal power units, increasing the charging power of pumped storage power stations, and mobilizing interruptible industrial loads to increase electricity consumption, thereby achieving local consumption of renewable energy within the province and ensuring that the voltage and power flow levels at provincial nodes are within safe limits.

[0082] In this example, the construction expression for the second decision sub-model is as follows: ; In the formula, For decision-making variables in the allocation and power flow optimization of resources within the province (such as unit output plan, load regulation, new energy quota, etc.); This represents the weighting coefficient for the degree of response to the upper-level strategy; For deviation measurement function; For upper-layer channel capacity and price signals The mapping results of resource allocation within the region.

[0083] For example, the conventional safe operation constraints that the second decision sub-model needs to meet include: First, power flow safety constraints: ensuring that the transmission power of each line and transformer does not exceed the thermal stability limit; Second, node voltage constraints: requiring the voltage amplitude of each regional node to operate within the allowable deviation range; Third, channel capacity constraints: strictly adhering to the upper limit of the inter-provincial interconnection line transmission capacity issued by the upper level; Fourth, power balance constraints: maintaining the real-time balance between power generation, power reception, load and loss within the province; Fifth, equipment operation constraints: the unit output and adjustable resources operate within the upper and lower limits of physical output.

[0084] In this example, the solution can be obtained iteratively using optimization algorithms such as interior-point method and sequential quadratic programming. Through this mechanism, the system not only realizes the quantitative transmission of upper-level channel strategies with provincial generator output plans, load regulation, and new energy consumption quotas, but also ensures the precise implementation of inter-provincial channel strategies at the provincial execution level, achieving optimal cross-regional collaboration and global resource utilization efficiency.

[0085] For example, the solution process for the second decision sub-model can be iteratively optimized using the interior-point method. The specific solution process is as follows: First, construct a logarithmic barrier function: For various constraints of the second decision sub-model (such as power flow safety constraints, power balance constraints, etc.), transform them into a set of nonlinear equality or inequality constraints; by introducing a logarithmic barrier function to handle inequality constraints, the original constrained optimization problem is transformed into a sequence of unconstrained optimization sub-problems, so as to ensure that the decision variables always remain within the feasible region.

[0086] Second, iterative calculation and solution: based on the current decision variables. The gradient information is used to calculate the search direction using Newton's method, and the decision variables are updated to adjust the running cost function. Under the premise of satisfying the consistency constraint, the search converges rapidly along the central path; in each iteration, the approach accuracy of the search direction to each constraint condition is dynamically controlled by adjusting the size of the obstacle factor.

[0087] Third, convergence judgment: If the decrease in the objective function and the change in the decision variables are both less than the preset convergence threshold during the iteration process, the model is considered to have reached the optimal solution, the iteration stops, and the decision variables at this time are output. .

[0088] It should be noted that the threshold for the relative decrease in the objective function of the second decision sub-model can be taken as follows: The threshold for the magnitude of change in decision variables can be taken as follows: The power balance residual threshold can be set to 0.1MW, and the maximum number of iterations can be set to 200.

[0089] For example, in a specific scenario Initially, the load demand of the provincial power grid is high, while the output of wind and solar renewable energy is insufficient. The interior-point method solution process is as follows: First, the system initializes decision variables such as unit power output and load regulation. In the first iteration, the algorithm detects that the current unit output has not reached the power balance constraint, and then calculates the search step size in the direction of increasing unit output. As the iteration progresses, the algorithm dynamically adjusts the output ratio by adjusting the marginal generation cost coefficient of each unit while ensuring that the output of each unit does not exceed the thermal stability limit. After several iterations, the algorithm outputs the provincial resource allocation strategy, which includes the target active power output of each gas turbine unit in the province at time t. This is in conjunction with the adjustment of adjustable loads and the power flow distribution results that meet the requirements for safe and stable operation of the power grid.

[0090] For example, the resource planning scheme is the equilibrium result after game convergence, which includes the aforementioned provincial resource allocation strategy. inter-provincial connection line trend This solution includes the settlement of collaborative benefits among various entities. It achieves optimal coordination among these entities across multiple time scales, ensuring not only the maximization of economic benefits for the regional power grid under uncertain conditions but also, through real-time updates and coupling of signals at various levels, enabling the quantitative implementation of the upper-level global plan at the provincial execution level.

[0091] In one implementation, a hierarchical linkage game optimization is performed on the provincial resource allocation strategy, price signals, and the operating status of the power grid entities to obtain resource planning schemes corresponding to each power grid entity. This includes: constructing aggregation revenue functions for distributed resource aggregation entities under different time scale scenarios based on price signals, provincial resource allocation strategies, adjustable load status, energy storage operation status, and distributed power output status corresponding to distributed resource aggregation entities; constructing a third decision sub-model with the goal of maximizing the aggregation revenue function; solving the third decision sub-model through game theory based on the constraints of the third decision sub-model to obtain power generation and consumption strategies; and determining resource planning schemes corresponding to each power grid entity based on channel capacity allocation strategies, provincial resource allocation strategies, and power generation and consumption strategies.

[0092] For example, the third decision sub-model refers to a lower-level decision-making model that aims to maximize the self-interest of the distributed resource aggregation entity and coordinates the response capabilities of distributed photovoltaic, energy storage systems, and controllable loads within the region. As the terminal response layer of the three-layer linkage mechanism, this model receives pricing signals from the upper layer and resource allocation instructions from the middle layer to achieve refined regulation of distributed resources and optimal power generation and consumption behavior decisions.

[0093] In this example, the aggregate revenue function consists of the revenue from electricity sales / consumption resulting from market electricity prices, minus resource operation and maintenance costs and development costs. Its calculation expression is as follows: ; In the formula, For the revenue function of the distributed resource aggregation entity; As a market price signal; For the scene The power generation / consumption strategy is as follows (positive indicates power generation and supply, negative indicates power consumption). This is the operating and maintenance cost coefficient per unit power. To compensate for the costs of resource development.

[0094] In this example, the construction expression for the third decision sub-model is as follows: ; In the formula, The variables representing the power generation and consumption strategies to be decided upon; Resource dispatch instructions allocated to mid-level provincial power grids; Capacity and price signals are established for upper-level cross-regional entities; This represents the current state of an uncertain scenario.

[0095] For example, the constraints that the third decision sub-model needs to meet include: First, distributed resource output constraints: limiting the upper limit of power output of photovoltaic, energy storage and other equipment within the allowable range; Second, adjustable capacity constraints: limiting the adjustment range of interruptible or transferable loads according to the load-side response capability; Third, local operation constraints: meeting the state of charge balance and power charging and discharging limits of the energy storage system; Fourth, power balance constraints: ensuring the balance requirements of distributed resources with the local microgrid or distribution side in various scenarios.

[0096] For example, in a summer evening peak scenario, market price signals rise rapidly, and the provincial power grid issues peak-shaving instructions. At this time, the distributed resource aggregation entity coordinates resource adjustments within the region based on the current energy storage load status, the adjustable load status on the user side, and the real-time output of distributed photovoltaic power. Specifically, during the high-price period from 18:00 to 20:00, the energy storage system is controlled to discharge, reducing the power consumption of some transferable industrial and charging loads; during the low-price period after 22:00, the energy storage system is controlled to recharge, restoring the aforementioned adjustable load operation; simultaneously, during the midday peak photovoltaic output phase, priority is given to consuming local distributed photovoltaic power generation, reducing external power purchase demand. This forms the corresponding power generation and consumption strategy for the current scenario. .

[0097] For example, to solve the third decision sub-model, since its objective function and constraints are mostly linear or convex functions, methods such as convex optimization, linear programming, or first-order optimality conditions can be used to obtain the optimal power generation / consumption strategy. .

[0098] In this example, we will use a convex optimization algorithm to solve the third decision sub-model. First, we will construct the convex optimization model: This involves using the aggregated reward function... The problem is converted to a negative number and then standardized into a convex optimization problem by combining the upper and lower limits of distributed resource output, energy storage charging and discharging power constraints, state of charge (SOC) balance constraints, and local power balance constraints. In this process, a penalty term is introduced to handle nonlinear characteristics such as energy storage lifetime loss, further transforming the model into a quadratic programming problem to ensure global convergence of the solution.

[0099] Second, determine the optimal search path: The primal-dual interior-point method is used to iteratively calculate the QP problem. First, the current Newton step size is calculated using a predictor-corrector mechanism to determine the resource output strategy. The update direction; in each iteration, by adjusting the barrier parameters under complementary relaxation conditions, ensure that the decision variables, while satisfying the physical constraints of distributed device operation, move towards the aggregation benefit function. The direction of maximization is approximated.

[0100] Third, convergence criterion verification: Monitor the incremental changes in the aggregate revenue function and the residual values ​​under various constraints (such as power imbalance and output exceeding limits). Convergence criterion verification is performed if and only if the residual value is lower than a preset tolerance limit (e.g., ...). If the incremental revenue from two consecutive iterations is less than the convergence threshold, the optimization process stops, and the optimal power generation and consumption strategy for distributed resources in this scenario at the current moment is output. .

[0101] For example, the tolerance limit can be specifically set as follows: power balance residual error is less than 0.1MW, output over-limit residual error is less than 0.5% of the equipment's rated power; the relative revenue increment between two consecutive iterations is less than Stop the optimization process when the time comes.

[0102] For example, in a specific scenario Under the condition that distributed photovoltaic power output is hindered and the electricity price signal is... At its peak. During execution, the convex optimization solver first identifies that the energy storage system is currently in a low-power state. As iterations proceed, the algorithm automatically calculates the optimal charging and discharging power while satisfying the energy storage charging and discharging rate constraints. This means generating revenue through power outages during periods of highest electricity prices and encouraging user participation in peak shaving during periods of lower electricity prices. Ultimately, the solver outputs the optimal strategy. It can balance the revenue from curtailment of photovoltaic power with the operation and maintenance costs of energy storage, maximize the revenue of the entity under the guidance of current price signals, and upload this strategy to the provincial dispatch center as the basis for the execution of the linkage decision chain.

[0103] For example, based on the channel capacity allocation strategy, the provincial resource allocation strategy, and the optimized power generation and consumption strategy, a resource planning scheme covering all entities in the entire network can be integrated, which can realize the systematic evolution of cross-provincial power grid dispatch from single planning to multi-entity game collaboration.

[0104] In one implementation, the first benefit in the resource planning scheme of each power grid entity is reconstructed to obtain the second benefit of each power grid entity. This includes: for each time scale scenario, performing the following operations: constructing a benefit compensation term corresponding to each power grid entity based on a preset utility transfer coefficient and the benefit difference between each power grid entity; determining the coordination benefit corresponding to each power grid entity based on the benefit compensation term and the first benefit, wherein the coordination benefit is regulated by transfer constraints; constructing a multi-objective collaborative optimization model based on each coordination benefit, a preset new energy absorption rate, and a safety margin index; and iteratively solving the multi-objective collaborative optimization model based on utility conservation constraints to obtain the second benefit corresponding to each power grid entity.

[0105] For example, the utility conservation constraint is a mechanism used to balance the intensity of the redistribution of benefits among agents, ensuring that the utility coordination process only changes the benefit distribution structure of each agent, without changing the total benefit scale of the system in a specific scenario.

[0106] In this example, the process of determining the coordination benefit is as follows: Obtain the first benefit (i.e., the original benefit) of each power grid entity under the game equilibrium strategy, and define the utility transfer coefficient obtained by entity i from entity j. By calculating the difference between the original benefit and the benchmark benefit of each entity, construct the corresponding benefit compensation term, and thus determine the coordination benefit. The calculation expression is as follows: ; In the formula, For the scene The coordinated benefits of subject i; This is the original income (i.e., the first income); The utility transfer coefficient is used to characterize the intensity of redistribution. The benchmark return for subject j under independent decision-making conditions; That is, the income compensation item.

[0107] For example, the weights of the absorption rate and the safety margin should be set in a normalized manner, and the sum of their values ​​should be 1; for the scenario of prioritizing the absorption of new energy, the weights can be 0.6 and 0.4 respectively, and for the scenario of prioritizing safety constraints, the weights can be 0.4 and 0.6 respectively.

[0108] In this example, the construction process of the multi-objective collaborative optimization model is as follows: With fairness, global benefit, and system performance as optimization objectives, a multi-objective hybrid utility function is constructed based on the aforementioned coordination benefits and system indicators. The calculation expression is as follows: ; In the formula, For scene probabilities; This is a fairness evaluation function used for revenue compression; , These are the weighting coefficients for the absorption rate and the safety margin, respectively. The utility conservation constraint is defined as follows: That is, the sum of the benefits of all parties after coordination is always equal to the total benefits of the system. .

[0109] In this example, the solution process for the multi-objective collaborative optimization model is as follows: a hybrid optimization strategy combining the Non-dominated Sorting Genetic Algorithm II (NSGA-II) algorithm with the multi-objective interior point method is adopted. First, the global search capability of NSGA-II is used to traverse the feasible solution space, generating a population of initial benefit allocation schemes that satisfy the utility conservation constraint, and obtaining the Pareto optimality (Pareto) front. Then, the local convergence property of the multi-objective interior point method is used to refine the candidate solutions on the Pareto front with high precision, eliminating non-dominated solutions that are prone to getting trapped in local optima, and finally outputting a set of second benefit schemes that satisfy the balance of multiple objectives of fairness, efficiency and security.

[0110] For example, taking a system that includes three types of entities—inter-regional power transmission entities, provincial power grid entities, and distributed resource aggregation entities—as an example, in the scenario probability... Total System Revenue Given a capital of 10,000 yuan, the initial original return is Significant uneven distribution exists. After solving using the above algorithm, the system can output multiple Pareto compromise solutions, as shown in Table 1: Table 1 The system scheduler selected solution 2 (38, 42, 20) as the final execution plan. This solution significantly improved the fairness index while ensuring the total revenue remained unchanged, and ensured optimal coordination between system safety margin and renewable energy absorption rate. Based on preset preference weights, the system scheduler selected this compromise plan as the final second revenue distribution result, achieving a smooth convergence of multiple stakeholders from local interest games to the unified optimal state of the system.

[0111] It should be noted that the preference weights can be determined using the analytic hierarchy process or the entropy weight method; when there are no special preferences, the weights for fairness, efficiency and security can be set to 0.3, 0.4 and 0.3 respectively.

[0112] In one implementation, rolling planning of power grid resources for each power grid entity is performed based on their resource planning schemes and second benefits. This includes: determining the target resource scheduling strategy and target benefit allocation structure for the current decision period based on the resource planning schemes and second benefits of each power grid entity; dynamically adjusting the capacity of inter-provincial transmission channels, intra-provincial resource scheduling schemes, and distributed power generation and consumption behavior based on the target resource scheduling strategy to obtain the rolling planning result for the current decision period; updating the cross-regional collaborative state vector based on the rolling planning result to obtain the initial collaborative state for the next decision period; dynamically correcting the time-scale scenario set and scenario probability set for the next decision period based on the new energy operation results, load operation results, and resource scheduling results for the current decision period; updating the second benefits for each power grid entity based on the target benefit allocation structure for the current decision period to obtain the benefit coordination structure for the next decision period; and executing the rolling planning for the next decision period based on the initial collaborative state, time-scale scenario set, scenario probability set, and benefit coordination structure for the next decision period.

[0113] For example, the target resource scheduling strategy refers to the optimal set of behavioral instructions for each power grid entity obtained after game optimization convergence within the current decision-making cycle, including the capacity allocation of inter-provincial channels, power generation / load regulation instructions of units within the province, and distributed resource power generation and consumption behavior plans.

[0114] For example, when there is a risk of power flow congestion in the system, the target resource scheduling strategy will automatically reduce the capacity configuration of the blocked channels, while increasing the response priority of the adjustable load within the province.

[0115] For example, the target benefit distribution structure refers to the second benefit that each subject should obtain in the current scenario and its corresponding compensation rules after being reconstructed through a utility coordination mechanism.

[0116] For example, the structure includes the basic revenue share of each entity and the risk compensation ratio for high / low revenue entities, reflecting the system's revenue orientation of "fairness and efficiency".

[0117] In this example, the rolling planning and dynamic correction process is as follows: Taking the current decision cycle (e.g., 15 minutes) as the unit, the system uses the previous round's equilibrium strategy (channel capacity allocation, intra-provincial scheduling plan, distributed resource strategy) and coordination benefits as the initial benchmark, and substitutes them into the three-layer linkage decision model; each subject conducts multiple rounds of game iteration based on the updated real-time operation data (cross-regional collaborative state vector) and scenario evolution path, and finally outputs the resource rolling planning result that meets the latest working conditions.

[0118] In this example, the dynamic correction process for the state and scene is as follows: First, by updating the cross-regional collaborative state vector As the initial state for the next cycle, it characterizes the real-time changes in the power grid topology and power flow distribution; secondly, based on the actual operating results of the previous cycle (fluctuations in renewable energy output and load deviations), it sets up time-scale scenarios. With probability distribution The system is modified by employing a Bayesian update method to optimize the conditional probabilities of each scenario. Finally, the subject utility transfer coefficients are adjusted using the recurring updated benefit coordination structure. Make corrections.

[0119] In this example, the logical chain for dynamic updates is as follows: First, state update: First, to achieve closed-loop tracking of system state; second, scenario correction: Third, revenue restructuring: Ensure the predicted scenarios closely reflect current operating conditions. , This ensures fairness in the benefits during the ongoing game.

[0120] For example, within a 15-minute decision-making cycle, if a sudden extreme weather event causes the actual transmission power of a cross-provincial channel to be far lower than expected, the model will immediately trigger rolling planning after identifying the deviation: first, correcting the current collaborative state. This mechanism reflects the reduced load conditions of transmission channels; then, it dynamically resamples the predicted scenarios of new energy and load for the next hour to reduce the probability of high-risk scenarios; finally, it updates the utility transfer coefficients of each entity to guide the inter-regional transmission operators and provincial power grid entities to redistribute capacity, ensuring that the system operating path approaches the optimal stability point again in the next cycle. This rolling mechanism realizes the transformation from "static predictive planning" to "dynamic adaptive evolution," significantly improving the operability and robustness of the system in complex and ever-changing game environments.

[0121] For example, the subject utility transfer coefficient can be calculated as follows: Update, in which, To adjust the rate (e.g., 0.05). This means that the utility transfer coefficient is limited to the range of 0-0.3.

[0122] In the formula, This is the updated subject utility transfer coefficient between subject i and subject j in the next decision cycle (time t+1). It is used for revenue coordination and restructuring in the next cycle. The initial / current subject utility transfer coefficient between subject i and subject j in the current decision-making cycle (time t). This is used to adjust the rate (or learning rate / step size). It controls how quickly the utility transfer coefficient changes; the example value in the image is 0.05. This refers to the current utility (or benefit) of subject j. In the context of the discussion, this typically refers to subject j's "secondary benefit" within the current decision-making cycle. This refers to the current utility (or benefit) of subject i. Similarly, it refers to the "secondary benefit" of subject i. In the formula... This indicates the difference in returns between the two, reflecting the fair allocation orientation of "risk compensation ratio for high / low-return entities" in the background. This represents the total system utility (or total system revenue). It serves as a benchmark for normalizing the revenue differences between agents. This makes the adjustment range relatively smooth and unaffected by the overall absolute return of the system. This is a truncation / limiting function. It means that the calculation result within the parentheses will be forcibly restricted to a specified upper and lower limit range. In this formula, it means that no matter how large the payoff difference is, the calculated utility transfer coefficient cannot exceed 0.3 at most and cannot be lower than 0 at least, in order to prevent overcompensation from causing game divergence in the system or creating new unfairness.

[0123] It should be noted that in the simulation scenario involving three types of power grid entities, two inter-provincial transmission channels, and four 15-minute decision cycles, the renewable energy absorption rate was 92%, the average channel utilization rate was 81%, and the number of over-limit incidents was 3 when using static planning. After adopting the rolling planning method in this embodiment, the renewable energy absorption rate can be increased to 96%, the average channel utilization rate can be increased to 88%, and the number of over-limit incidents can be reduced to 0.

[0124] Figure 2 This is a structural block diagram of a cross-provincial power grid resource dynamic planning device according to an embodiment of the present invention.

[0125] like Figure 2 As shown, the cross-provincial power grid resource dynamic planning device may include: The structured construction module 510 is used to construct the multi-source power grid resource data of multiple power grid entities at each first moment in a structured manner, so as to obtain the cross-regional collaborative state vector at each first moment. The dynamic scenario construction module 520 is used to construct dynamic scenarios for the uncertain states of new energy output and load demand driven by the cross-regional collaborative state vector at each first moment, and obtain a time-scale scenario set and a scenario probability set. The game optimization module 530 is used to construct a multi-agent game model based on the cross-regional collaborative state vectors at each first moment, the time scale scenario set, and the scenario probability set, and to perform game optimization on the multi-agent game model to obtain the resource planning schemes of each of the power grid entities. The reconstruction module 540 is used to reconstruct the first benefit in the resource planning scheme of each of the power grid entities to obtain the second benefit of each of the power grid entities; The rolling planning module 550 is used to perform rolling planning of the power grid resources of each of the power grid entities based on the resource planning scheme and the second benefit of each of the power grid entities.

[0126] In one implementation, the dynamic scene construction module includes: The joint uncertainty variable calculation unit is used to calculate the joint uncertainty variable of each time scale at each time scale by summing the first product of the cross-regional cooperative state vector at each time scale and the preset influence coefficient, the second product of the historical uncertainty variable at the second time scale and the preset inertia coefficient, and the preset random disturbance term, for the cross-regional cooperative state vector at each time scale at each time scale at each time scale. The target uncertainty variable calculation unit is used to perform multi-time-scale coupling on the joint uncertainty variables of each time scale at each first moment to obtain the target uncertainty variables of each time scale at each first moment. The association and recombination unit is used to associate and recombine the target uncertainty variables of each time scale at each first moment according to the time sequence, so as to obtain the time scale scene set. The conditional probability determination unit is used to determine the conditional probability of the time scale scene at each first moment based on the target uncertainty variable at each first moment in the time scale scene set. The multiplication calculation unit is used to perform multiplication calculations on the conditional probabilities of each of the time scale scenarios at each first moment to obtain the scenario probability set.

[0127] In one implementation, the multi-agent game model includes a first decision sub-model, a second decision sub-model, and a third decision sub-model, and the game optimization module includes: The external operating condition determination unit is used to determine the external operating conditions of the cross-regional power transmission operator under different time scale scenarios based on the cross-regional collaborative state vector at each first moment, the time scale scenario set, and the scenario probability set. The global revenue function construction unit is used to construct the global revenue function corresponding to the inter-regional power transmission operator based on the channel operation revenue, network access fee revenue and resource scheduling cost under each time scale scenario in the time scale scenario set. The first decision sub-model construction unit is used to construct the first decision sub-model with the goal of maximizing the global benefit function. The game-solving unit is used to solve the first decision sub-model based on the constraints of the first decision sub-model to obtain the channel capacity allocation strategy and price signal. The hierarchical linkage game optimization unit is used to perform hierarchical linkage game optimization on the channel capacity allocation strategy, the price signal and the operating status of each of the power grid entities to obtain the resource planning scheme corresponding to each of the power grid entities.

[0128] In one embodiment, the hierarchical linkage game optimization unit includes: The sub-unit for constructing the operating cost function is used to construct the operating cost function of the provincial power grid under different time scale scenarios based on the channel capacity allocation strategy, the price signal, the load demand within the province corresponding to the provincial power grid, the output of new energy sources, and the operating status of regulation resources. The second decision sub-model construction unit is used to construct the second decision sub-model with the goal of minimizing the operating cost function; The game-solving subunit is used to solve the second decision-making sub-model based on the constraints of the second decision-making sub-model to obtain the resource allocation strategy within the province. The hierarchical linkage game optimization subunit is used to perform hierarchical linkage game optimization on the provincial resource allocation strategy, the price signal and the operating status of the power grid entity to obtain the resource planning scheme corresponding to each of the power grid entities.

[0129] In one implementation, the hierarchical linkage game optimization subunit is specifically used for: Based on the price signal, the provincial resource allocation strategy, the adjustable load status, energy storage operation status, and distributed power output status of the distributed resource aggregation entity, an aggregation revenue function of the distributed resource aggregation entity under different time scale scenarios is constructed. The third decision sub-model is constructed with the goal of maximizing the aggregated return function. Based on the constraints of the third decision sub-model, a game-theoretic solution is performed on the third decision sub-model to obtain the power generation and consumption strategy; Based on the channel capacity allocation strategy, the provincial resource allocation strategy, and the power generation and consumption strategy, resource planning schemes corresponding to each of the power grid entities are determined respectively.

[0130] In one implementation, the reconstruction module is specifically used for: For each timescale scenario, the following operations are performed: Based on the preset utility transfer coefficient and the revenue difference between each of the power grid entities, a revenue compensation item corresponding to each of the power grid entities is constructed. Based on the revenue compensation items and first revenue corresponding to each of the power grid entities, the coordination revenue corresponding to each of the power grid entities is determined, wherein the coordination revenue is regulated by transfer constraints; Based on the aforementioned coordination benefits, the preset new energy consumption rate, and the safety margin indicators, a multi-objective collaborative optimization model is constructed. The multi-objective collaborative optimization model is iteratively solved based on the utility conservation constraint to obtain the second benefit corresponding to each of the power grid entities.

[0131] In one implementation, the rolling planning module includes: The determining unit is used to determine the target resource scheduling strategy and target revenue allocation structure for the current decision-making cycle based on the resource planning scheme and second revenue of each of the power grid entities. The dynamic adjustment unit is used to dynamically adjust the capacity of inter-provincial power transmission channels, intra-provincial resource scheduling schemes, and distributed resource power generation and consumption behavior based on the target resource scheduling strategy, so as to obtain the resource rolling planning results corresponding to the current decision cycle. The first update unit is used to update the cross-regional collaborative state vector based on the resource rolling planning results to obtain the initial collaborative state corresponding to the next decision cycle. The dynamic correction unit is used to dynamically correct the time-scale scenario set and scenario probability set corresponding to the next decision cycle based on the new energy operation results, load operation results and resource scheduling results corresponding to the current decision cycle. The second update unit is used to update the second revenue corresponding to each of the power grid entities based on the target revenue allocation structure corresponding to the current decision cycle, so as to obtain the revenue coordination structure corresponding to the next decision cycle. The execution unit is used to execute the rolling plan for the next decision cycle based on the initial collaborative state, time scale scenario set, scenario probability set, and benefit coordination structure corresponding to the next decision cycle.

[0132] The specific functions and examples of each module and submodule of the system in this embodiment of the invention can be found in the relevant descriptions of the corresponding steps in the above method embodiments, and will not be repeated here.

[0133] The acquisition, storage, and application of user personal information involved in the technical solution of this invention all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0134] This invention also provides a computer device, comprising: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method described in any one of the embodiments of the present invention.

[0135] The beneficial effects of the computer equipment in this embodiment of the invention are equivalent to the beneficial effects of the above-described cross-provincial power grid resource dynamic planning method, and will not be repeated here.

[0136] This invention also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the method described in any one of the embodiments of this invention.

[0137] The beneficial effects of the storage medium of the present invention are equivalent to the beneficial effects of the above-mentioned cross-provincial power grid resource dynamic planning method, and will not be elaborated here.

[0138] Figure 3 A schematic block diagram of an example computer device 800 that can be used to implement embodiments of the present invention is shown. Computer device 800 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Computer device 800 may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0139] like Figure 3 As shown, the computer device 800 includes a computing unit 801, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. The RAM 803 may also store various programs and data required for the operation of the computer device 800. The computing unit 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0140] Multiple components in computer device 800 are connected to I / O interface 805, including: input unit 806, such as keyboard, mouse, etc.; output unit 807, such as various types of monitors, speakers, etc.; storage unit 808, such as disk, optical disk, etc.; and communication unit 809, such as network card, modem, wireless transceiver, etc. Communication unit 809 allows computer device 800 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0141] The computing unit 801 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as the cross-provincial power grid resource dynamic planning method. For example, in some embodiments, the cross-provincial power grid resource dynamic planning method can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as storage unit 808. In some embodiments, part or all of the computer program can be loaded and / or installed on the computer device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of the cross-provincial power grid resource dynamic planning method described above can be performed. Alternatively, in other embodiments, the computing unit 801 may be configured to perform a cross-provincial power grid resource dynamic planning method by any other suitable means (e.g., by means of firmware).

[0142] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0143] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0144] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0145] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0146] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0147] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0148] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0149] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of this invention should be included within the scope of protection of this invention.

Claims

1. A dynamic planning method for cross-provincial power grid resources, characterized in that, include: The multi-source power grid resource data of multiple power grid entities at each first moment are structured and constructed to obtain the cross-regional collaborative state vector at each first moment; Dynamic scenarios are constructed for the uncertain states of new energy output and load demand driven by the cross-regional collaborative state vector at each first moment, resulting in a time-scale scenario set and a scenario probability set. A multi-agent game model is constructed based on the cross-regional collaborative state vectors at each first moment, the time-scale scenario set, and the scenario probability set. The multi-agent game model is then optimized through game theory to obtain resource planning schemes for each of the power grid entities. The first benefit in the resource planning scheme of each of the power grid entities is reconstructed to obtain the second benefit of each of the power grid entities; Based on the resource planning schemes and secondary benefits of each of the aforementioned power grid entities, rolling planning is carried out on the power grid resources of each of the aforementioned power grid entities.

2. The method according to claim 1, characterized in that, The dynamic scenario construction, driven by the cross-regional collaborative state vector at each first moment, generates a time-scale scenario set and a scenario probability set, including: For each cross-regional cooperative state vector at the first moment, the first product of the cross-regional cooperative state vector at the first moment and the preset influence coefficient, the second product of the historical uncertainty variable at the second moment and the preset inertia coefficient, and the preset random disturbance term are summed at each time scale to obtain the joint uncertainty variable at the first moment of each time scale. Multi-time-scale coupling is performed on the joint uncertainty variables of each time scale at each first moment to obtain the target uncertainty variables of each time scale at each first moment. Based on the time sequence, the target uncertainty variables of each time scale at each first moment are respectively correlated and reorganized to obtain the set of time scale scenarios. For each time-scale scenario in the set of time-scale scenarios, the conditional probability of the time-scale scenario at each first moment is determined based on the target uncertainty variable at each first moment in the time-scale scenario. The conditional probabilities of each timescale scene at each first moment are multiplied together to obtain the scene probability set.

3. The method according to claim 1, characterized in that, The process involves constructing a multi-agent game model based on the cross-regional collaborative state vectors at each first moment, the time-scale scenario set, and the scenario probability set. The multi-agent game model is then optimized through game theory to obtain resource planning schemes for each of the power grid entities. The multi-agent game model includes a first decision sub-model, a second decision sub-model, and a third decision sub-model, comprising: Based on the cross-regional collaborative state vectors at each first moment, the time-scale scenario set, and the scenario probability set, the external operating conditions of the cross-regional power transmission operator under different time-scale scenarios are determined. Based on the channel operation revenue, grid access fee revenue, and resource scheduling cost under each time scale scenario in the time scale scenario set, a global revenue function corresponding to the cross-regional power transmission operator is constructed. The first decision sub-model is constructed with the goal of maximizing the global benefit function; Based on the constraints of the first decision sub-model, a game theory solution is performed on the first decision sub-model to obtain the channel capacity allocation strategy and price signal; The channel capacity allocation strategy, the price signal, and the operating status of each power grid entity are optimized through hierarchical linkage game theory to obtain the resource planning scheme corresponding to each power grid entity.

4. The method according to claim 3, characterized in that, The step of performing hierarchical linkage game optimization on the channel capacity allocation strategy, the price signal, and the operating status of each of the power grid entities to obtain the resource planning scheme corresponding to each of the power grid entities includes: Based on the channel capacity allocation strategy, the price signal, the load demand within the province corresponding to the provincial power grid, the output of new energy sources, and the operating status of regulation resources, an operating cost function of the provincial power grid under different time scale scenarios is constructed. The second decision sub-model is constructed with the goal of minimizing the aforementioned operating cost function; Based on the constraints of the second decision sub-model, a game-theoretic solution is performed on the second decision sub-model to obtain the resource allocation strategy within the province; The resource allocation strategy within the province, the price signal, and the operating status of the power grid entity are optimized through hierarchical linkage game theory to obtain the resource planning scheme corresponding to each power grid entity.

5. The method according to claim 4, characterized in that, The step of performing hierarchical linkage game optimization on the provincial resource allocation strategy, the price signal, and the operating status of the power grid entity to obtain the resource planning scheme corresponding to each of the power grid entities includes: Based on the price signal, the provincial resource allocation strategy, the adjustable load status, energy storage operation status, and distributed power output status of the distributed resource aggregation entity, an aggregation revenue function of the distributed resource aggregation entity under different time scale scenarios is constructed. The third decision sub-model is constructed with the goal of maximizing the aggregated return function. Based on the constraints of the third decision sub-model, a game-theoretic solution is performed on the third decision sub-model to obtain the power generation and consumption strategy; Based on the channel capacity allocation strategy, the provincial resource allocation strategy, and the power generation and consumption strategy, resource planning schemes corresponding to each of the power grid entities are determined respectively.

6. The method according to claim 1, characterized in that, The process of reconstructing the first benefit in the resource planning schemes of each of the power grid entities to obtain the second benefit for each of the power grid entities includes: For each timescale scenario, the following operations are performed: Based on the preset utility transfer coefficient and the revenue difference between each of the power grid entities, a revenue compensation item corresponding to each of the power grid entities is constructed. Based on the revenue compensation items and first revenue corresponding to each of the power grid entities, the coordination revenue corresponding to each of the power grid entities is determined, wherein the coordination revenue is regulated by transfer constraints; Based on the aforementioned coordination benefits, the preset new energy consumption rate, and the safety margin indicators, a multi-objective collaborative optimization model is constructed. The multi-objective collaborative optimization model is iteratively solved based on the utility conservation constraint to obtain the second benefit corresponding to each of the power grid entities.

7. The method according to claim 1, characterized in that, The rolling planning of power grid resources for each of the power grid entities, based on their respective resource planning schemes and second benefits, includes: Based on the resource planning schemes and secondary benefits of each of the aforementioned power grid entities, the target resource scheduling strategy and target benefit allocation structure for the current decision-making cycle are determined. Based on the target resource scheduling strategy, the capacity of inter-provincial power transmission channels, intra-provincial resource scheduling schemes, and distributed resource power generation and consumption behavior are dynamically adjusted to obtain the resource rolling planning results corresponding to the current decision cycle. Based on the resource rolling planning results, the cross-regional collaborative state vector is updated to obtain the initial collaborative state corresponding to the next decision cycle. Based on the new energy operation results, load operation results and resource scheduling results corresponding to the current decision cycle, the time scale scenario set and scenario probability set corresponding to the next decision cycle are dynamically corrected. Based on the target revenue allocation structure corresponding to the current decision cycle, the second revenue corresponding to each of the power grid entities is updated to obtain the revenue coordination structure corresponding to the next decision cycle. Based on the initial collaborative state, timescale scenario set, scenario probability set, and benefit coordination structure corresponding to the next decision cycle, the rolling planning for the next decision cycle is executed.

8. A cross-provincial power grid resource dynamic planning device, characterized in that, include: The structured construction module is used to construct the multi-source power grid resource data of multiple power grid entities at each first moment in a structured manner, so as to obtain the cross-regional collaborative state vector at each first moment. The dynamic scenario construction module is used to dynamically construct scenarios for the uncertainty of new energy output and load demand driven by the cross-regional collaborative state vector at each first moment, and obtain a time-scale scenario set and a scenario probability set. The game optimization module is used to construct a multi-agent game model based on the cross-regional collaborative state vectors at each first moment, the time scale scenario set, and the scenario probability set, and to perform game optimization on the multi-agent game model to obtain the resource planning schemes of each of the power grid entities. The reconstruction module is used to reconstruct the first benefit in the resource planning scheme of each of the power grid entities to obtain the second benefit of each of the power grid entities; The rolling planning module is used to perform rolling planning of the power grid resources of each of the power grid entities based on the resource planning scheme and the second benefit of each of the power grid entities.

9. A computer device, characterized in that, include: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-7.

Images