A renewable energy supply-delivery-demand-storage system space-time coordination configuration method

Abstract

The application provides a renewable energy supply-distribution-demand-storage system space-time coordination configuration method, relates to the technical field of intelligent energy, and comprises the following steps: acquiring data of multiple regions; calculating time mismatch degrees, space mismatch degrees and space-time coupling mismatch degrees of the regions; decomposing an optimization problem into a long-term planning layer, a medium-term scheduling layer and a short-term real-time layer in a time dimension; constructing multi-objective optimization functions of the time layers; decomposing the optimization problems of the time layers in a space dimension; alternately solving local optimization sub-problems and global coordination problems of the regions by using a distributed iteration method; outputting a configuration result of a current time layer when the algorithm converges; executing a decision of a first time period by each time layer and taking the decision as a constraint condition of a next time layer to obtain power generation installed capacities, power transmission line capacities, energy storage capacities and operation strategies of the regions. The application can improve a renewable energy consumption rate and reduce system operation costs.

Description

Technical Field

[0001] This invention relates to the field of smart energy technology, and in particular to a spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system. Background Technology

[0002] The proportion of renewable energy in the power system is constantly increasing. However, there is a significant mismatch between the spatiotemporal distribution characteristics of renewable energy and load demand: in terms of time, peak photovoltaic power generation occurs at noon, while peak load often occurs in the evening; in terms of space, the northwest region has abundant wind and solar resources but low load demand, while the eastern coastal region has concentrated load but limited renewable energy resources. This spatiotemporal mismatch leads to difficulties in renewable energy absorption and severe wind and solar curtailment, hindering the low-carbon transformation of the power system. Existing supply-transmission-demand-storage coordinated configuration methods mainly focus on system economic optimization and equipment capacity configuration, lacking a quantitative characterization of the degree of spatiotemporal mismatch, making it difficult to provide targeted guidance for energy storage configuration and transmission network planning to bridge the spatiotemporal mismatch.

[0003] Chinese patent application CN118100175A discloses an energy management method for energy systems based on distributed optimization. This method proposes a two-stage energy management model of pre- and real-time scheduling. The first stage is pre-economic scheduling, with the optimization objectives of maximizing economic benefits, maximizing efficiency, and minimizing carbon emission costs. The second stage is real-time optimization scheduling, with the optimization objectives of minimizing the costs of interactive power deviation penalties, wind curtailment penalties, and user satisfaction loss. This method uses an improved multi-agent reinforcement learning method to solve the pre-economic scheduling model and the YALMIP toolbox to solve the real-time optimization model. However, this method focuses on economic efficiency and carbon emissions, failing to incorporate the spatiotemporal mismatch between renewable energy and load as an explicit optimization indicator into the objective function. This makes it difficult to achieve optimized coordinated configuration of energy storage and transmission for spatiotemporal mismatch characteristics. Summary of the Invention

[0004] In view of this, the present invention provides a spatiotemporal coordinated configuration method for renewable energy supply-transmission-demand-storage systems. By constructing quantitative models of time mismatch, spatial mismatch, and spatiotemporal coupling mismatch, and taking spatiotemporal mismatch as an explicit optimization objective, a multi-scale optimization framework covering three time scales—long-term planning, medium-term scheduling, and short-term real-time—is established. A distributed iterative solution algorithm based on the alternating direction multiplier method is adopted to achieve coordinated configuration optimization of energy storage to bridge time mismatch and power transmission to alleviate spatial mismatch, thereby improving the renewable energy absorption rate and reducing system operating costs.

[0005] The technical solution of this invention is implemented as follows:

[0006] This invention provides a spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system, comprising:

[0007] S1. Acquire spatial topology data, historical time series data and power system parameters for multiple regions. Spatial topology data includes regional geographic coordinates, inter-regional transmission line connections and inter-regional distances. Historical time series data includes load demand data and renewable energy output data for each region. Power system parameters include existing installed capacity, equipment investment costs and operating costs.

[0008] S2. Based on the renewable energy output data and load demand data of each region, the time mismatch degree of each region is calculated using the similarity calculation method; the energy-load ratio is calculated according to the renewable energy resource potential and load demand data of each region; the spatial weighting coefficient is calculated based on the load demand data of each region and the distance between regions; the spatial mismatch degree is calculated using the weighted deviation method; and the spatial mismatch degree and the load share of each region are obtained.

[0009] S3. The time mismatch is weighted and amplified according to the load ratio of each region. The weighted time mismatch is then fused with the spatial mismatch to obtain the spatiotemporal coupling mismatch.

[0010] S4. Decompose the optimization problem into a long-term planning layer, a medium-term scheduling layer, and a short-term real-time layer in the time dimension, and establish the capacity constraint relationship and power transfer relationship between each time layer.

[0011] S5. Construct multi-objective optimization functions for each time level;

[0012] S6. Decompose the optimization problem of each time layer in the spatial dimension, and decompose the global optimization problem into local optimization sub-problems of each region and global coordination problem;

[0013] S7. Initialize the iteration parameters based on the time mismatch degree and the distance between regions. Use a distributed iteration method to alternately solve the local optimization subproblems and the global coordination problems of each region. During the iteration process, dynamically adjust the iteration parameters based on the optimization residuals. The distributed iteration method includes local variable optimization steps, global variable coordination steps, and dual variable update steps.

[0014] S8. Determine the convergence between the local optimization subproblem and the global coordination problem, as well as the stability of the change in the spatiotemporal coupling mismatch. When both convergence and stability meet the preset conditions, output the configuration result of the current time layer.

[0015] S9. Execute steps S6 to S8 sequentially from long to short time scale for the long-term planning layer, the medium-term scheduling layer, and the short-term real-time layer. Execute the decision for the first time period at each time layer and use it as the constraint condition for the next time layer to obtain the power generation capacity, transmission line capacity, energy storage capacity, and operation strategy for each region.

[0016] Preferably, in step S2, the temporal mismatch degree of each region is calculated using a similarity calculation method, specifically including:

[0017] Normalize the renewable energy output data and load demand data for region s, and calculate the normalized renewable energy output. and load demand ;

[0018] The temporal mismatch degree of region s is calculated using the inverse index of cosine similarity. :

[0019] ;

[0020] In the formula, t is the time index. This represents the total number of historical time periods. The value range is [0, 2].

[0021] Preferably, in step S2, the spatial mismatch degree is calculated using the weighted bias method, specifically including:

[0022] The proportion of renewable energy resource potential in region s to the total system. and the proportion of load demand to the entire system ;

[0023] The spatial weighting coefficients are obtained by calculating the load-weighted geometric center coordinates of the entire system and then normalizing the distance from region s to the load center. ;

[0024] Calculate the spatial mismatch degree (SMD):

[0025] ;

[0026] In the formula, S is the total number of regions, and s is the region index.

[0027] Preferably, in step S3, the calculation formula for the spatiotemporal coupling mismatch degree (STCM) is obtained as follows:

[0028] ;

[0029] In the formula, For time mismatch weights, For spatial mismatch weights, , Let be the load deviation coefficient for region s. The coupling coefficient is... Let be the time mismatch degree of region s. This represents the spatial mismatch degree.

[0030] Preferably, in step S4, the optimization problem is decomposed into a long-term planning layer, a medium-term scheduling layer, and a short-term real-time layer in the time dimension. Specifically, the optimization time domain of the long-term planning layer is 1 year, the optimization granularity is 1 month, and the decision variables are the increase in installed power generation capacity, the increase in transmission line capacity, and the increase in energy storage capacity in each region; the optimization time domain of the medium-term scheduling layer is 7 days, the optimization granularity is 1 hour, and the decision variables are the power generation output, energy storage charging and discharging power, and inter-regional transmission power in each region at each time period; the optimization time domain of the short-term real-time layer is 4 hours, the optimization granularity is 15 minutes, and the decision variables are the power output adjustment, energy storage adjustment, and transmission adjustment.

[0031] Preferably, in step S5, the multi-objective optimization function at each time level includes a spatiotemporal coupling mismatch term, wherein the spatiotemporal coupling mismatch term is:

[0032] ;

[0033] In the formula, For spatiotemporal coupling mismatch degree, Let S be the decision variable for each region, and S be the total number of regions. Let be the total energy storage capacity of region s. This is the energy storage bridging coefficient. To prevent small amounts from being divided by zero, and For the renewable energy output and load demand of region s at time t, Let be the time mismatch degree of region s. For spatial mismatch degree, For time mismatch weights, For spatial mismatch weights, , Let be the load deviation coefficient for region s. is the coupling coefficient.

[0034] Preferably, in step S6, the global optimization problem is decomposed into local optimization subproblems for each region and a global coordination problem, and the augmented Lagrangian function is constructed using the alternating direction multiplier method:

[0035] ;

[0036] In the formula, This represents the augmented Lagrange function. Let S be the decision variable for each region, S be the total number of regions, and z be the global coordination variable, representing the inter-regional power transmission capacity. Let Lagrange multiplier vectors be used. Let be the local cost function of region s. For the global cost function, Let be the transpose of the Lagrange multiplier vector of region s. and It is an incidence matrix. This is the penalty parameter.

[0037] Preferably, in step S7, the iteration parameters are initialized based on the time mismatch degree of each region and the distance between regions. For region pairs (i,j) with transmission line connections, the initial penalty parameter is... for:

[0038] ;

[0039] In the formula, As the baseline penalty parameter, For time mismatch coupling coefficient, Let be the time mismatch degree of region i. Let j be the time mismatch degree of region j. This is the distance coefficient. Let be the distance between region i and region j. This represents the maximum inter-regional distance in the power system.

[0040] During the iteration process, the iteration parameters are dynamically adjusted based on the optimized residual, and the original residual is defined. and dual residual Update the penalty parameter based on the residual ratio. :

[0041] ;

[0042] In the formula, k is the iteration round. The residual ratio threshold, and This is an adjustment factor.

[0043] Preferably, in step S8, determining the convergence between the local optimization subproblem and the global coordination problem, as well as the stability of the change in spatiotemporal coupling mismatch, specifically includes:

[0044] Calculate the original residual norm of the entire system after the k-th iteration. and dual residual norm ;

[0045] Calculate the change in spatiotemporal coupling mismatch between two adjacent iterations. STCM stands for Spatiotemporal Coupling Mismatch.

[0046] Calculate the rate of change of spatiotemporal coupling mismatch and determine its stability;

[0047] when , , The iteration terminates when the rate of change stabilizes, where , and These are the original residual threshold, the dual residual threshold, and the spatiotemporal mismatch change threshold, respectively.

[0048] Preferably, in step S9, rolling optimization is performed on the long-term planning layer, the medium-term scheduling layer, and the short-term real-time layer in descending order of time scale, specifically including:

[0049] The long-term planning layer performs annual optimization at the initial planning stage to obtain the installed capacity deployment plan for each future month. Only the construction decision for the first month is executed, and the plans for the remaining months are used as predictive schemes. The installed capacity increment of the first month is passed to the medium-term scheduling layer as the capacity limit.

[0050] The mid-term scheduling layer performs weekly optimization at the day-ahead time of each day to obtain hourly operation plans for the next 7 days. Only the scheduling plan for the first day is executed, and the plans for the remaining 6 days are used as predictive schemes. The power setpoint for the first day is passed to the short-term real-time layer.

[0051] The short-term real-time layer performs 4-hour optimizations at the real-time point of each hour, and only performs adjustments for the first time period.

[0052] The present invention has the following advantages over the prior art:

[0053] (1) This invention quantifies the temporal mismatch between the renewable energy output curve and the load curve in each region by using the inverse index of cosine similarity, quantifies the spatial mismatch between the distribution of renewable energy resources and the distribution of load by using the weighted deviation method, and constructs a spatiotemporal coupling mismatch model that integrates load ratio, thereby realizing a quantitative characterization of the spatiotemporal mismatch between renewable energy and load, and providing a physical-driven evaluation index for configuration optimization.

[0054] (2) This invention establishes a three-layer time-scale optimization framework covering long-term planning, medium-term scheduling and short-term real-time. Each time layer is coupled through capacity constraint relationship and power transfer relationship. The rolling optimization mechanism is adopted to coordinate long-term capacity configuration decision and short-term operation scheduling decision, avoiding the shortcomings of traditional single time-scale optimization method in dealing with multi-time-scale decision coupling.

[0055] (3) This invention incorporates the spatiotemporal coupling mismatch degree as an explicit optimization objective into the multi-objective optimization function, and introduces an energy storage bridging factor in the calculation of the spatiotemporal coupling mismatch degree, so that the more sufficient the energy storage configuration, the smaller the impact of the time mismatch degree on the objective function. Thus, the optimization process automatically guides the energy storage configuration to tilt towards the region with severe time mismatch, thereby realizing the adaptation of energy storage configuration to time mismatch characteristics.

[0056] (4) The present invention uses the alternating direction multiplier method to decompose the global optimization problem into local optimization subproblems and global coordination problems in each region. Local subproblems in each region can be solved in parallel, and global coordination is achieved through iteration. This avoids the problem of excessive computational dimension in large-scale multi-region systems by centralized optimization methods and improves the solution efficiency.

[0057] (5) The present invention initializes the penalty parameters of distributed iteration based on the time mismatch degree of each region and the distance between regions, making the coordination constraints between regions with severe time mismatch stronger. In the iteration process, the penalty parameters are dynamically adjusted according to the ratio of the original residual to the dual residual. Compared with the traditional alternating direction multiplier method with fixed penalty parameters, the algorithm converges faster and the solution quality is improved.

[0058] (6) The present invention adds the determination conditions of the change in spatiotemporal coupling mismatch degree and the stability of the rate of change to the convergence criterion. In addition to the traditional original residual and dual residual convergence conditions, it ensures that the algorithm not only meets physical constraints such as power balance, but also achieves full optimization of spatiotemporal mismatch degree, thus avoiding premature termination of the algorithm when the spatiotemporal mismatch degree has not decreased sufficiently. Attached Figure Description

[0059] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1 This is a flowchart of the method of the present invention;

[0061] Figure 2 This is a diagram of the two-layer optimization framework of the present invention;

[0062] Figure 3 This is a flowchart of the iterative optimization process of the present invention. Detailed Implementation

[0063] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0064] like Figure 1As shown, this invention provides a spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system, comprising: S1, acquiring spatial topology data, historical time series data, and power system parameters for multiple regions; the spatial topology data includes regional geographic coordinates, inter-regional transmission line connections, and inter-regional distances; the historical time series data includes load demand data and renewable energy output data for each region; and the power system parameters include existing installed capacity, equipment investment costs, and operating costs; S2, calculating the temporal mismatch degree for each region based on the renewable energy output data and load demand data of each region using a similarity calculation method; calculating the energy-load ratio based on the renewable energy resource potential and load demand data of each region; calculating the spatial weight coefficient based on the load demand data and inter-regional distances of each region; and calculating the spatial mismatch degree using a weighted deviation method to obtain the spatial mismatch degree and the load share of each region; S3, weighting and amplifying the temporal mismatch degree based on the load share of each region; and fusing the weighted temporal mismatch degree with the spatial mismatch degree to obtain the spatiotemporal coupling mismatch degree; and S4, decomposing the optimization problem into a long-term planning layer, a medium-term scheduling layer, and a short-term real-time layer in the time dimension. The process involves the following steps: S5. Establishing capacity constraints and power transfer relationships between different time layers; S6. Constructing multi-objective optimization functions for each time layer; S7. Decomposing the optimization problem of each time layer spatially, breaking down the global optimization problem into local optimization sub-problems and global coordination problems for each region; S8. Initializing iteration parameters based on the time mismatch degree and inter-regional distance for each region, and using a distributed iteration method to alternately solve the local optimization sub-problems and global coordination problems for each region. During the iteration process, the iteration parameters are dynamically adjusted based on the optimization residuals. The distributed iteration method includes local variable optimization steps, global variable coordination steps, and dual variable update steps; S9. Determining the convergence between the local optimization sub-problems and the global coordination problem, as well as the stability of the spatiotemporal coupling mismatch degree. When both convergence and stability meet preset conditions, the configuration result of the current time layer is output; S10. Executing steps S6 to S8 sequentially from longest to shortest time scale for the long-term planning layer, the medium-term scheduling layer, and the short-term real-time layer. The decision for the first time period is executed for each time layer and used as the constraint condition for the next time layer, resulting in the installed power generation capacity, transmission line capacity, energy storage capacity, and operation strategy for each region.

[0065] In one embodiment of the present invention, step S1 includes:

[0066] Acquire spatial topology data and time series data of the power system;

[0067] The system contains $S$ region nodes, with region index s, where s = 1, 2, ..., S;

[0068] Get the geographic coordinates of region s ,in Longitude Latitude;

[0069] Obtain the transmission line connection relationships between regions to form edge sets. , where (i,j) indicates that there is a power transmission line between region i and region j;

[0070] Get the geographical distance between region i and region j The unit is kilometers;

[0071] Get historical time period The hourly data is indexed by time t. ;

[0072] For region s, obtain its historical load demand at time t. The unit is megawatts, representing historical power output data from renewable energy sources. The unit is megawatts (MW), and renewable energy includes wind power and solar power.

[0073] Obtain future planning time periods Load forecast data for each region and renewable energy resource potential data The unit is megawatts;

[0074] Obtain the existing installed power generation capacity in each region Transmission line capacity Energy storage capacity The unit is megawatt-hour;

[0075] Obtain the unit investment cost of various types of equipment , , and unit operating cost , , ;

[0076] Obtaining carbon emission constraints and investment budget constraints .

[0077] In one embodiment of the present invention, step S2 includes:

[0078] Based on the historical time series data obtained in step S1, a quantitative assessment is made of the degree of mismatch between renewable energy output and load demand in the time dimension.

[0079] The time series of renewable energy output and load in region s are normalized:

[0080] Calculate the average renewable energy output of region s over a historical period. and average load ;

[0081] Normalize renewable energy output:

[0082] ;

[0083] Normalize the load demand:

[0084] ;

[0085] In the formula, and The normalized time series has had its dimensional effects removed while retaining its temporal variation characteristics.

[0086] The cosine similarity inverse index is used to measure the degree of phase mismatch between the renewable energy output curve and the load curve:

[0087] For region s, calculate the time mismatch degree. :

[0088] ;

[0089] In the formula, the numerator is the inner product of normalized renewable energy output and load, representing the correlation between the two; the denominator is the norm product of the two, used for normalization; t is the time index. This represents the total number of historical time periods. The value range is [0,2];

[0090] when When this occurs, it means that the renewable energy output curve and the load curve are completely synchronized and perfectly matched in the time dimension; when When the two are orthogonal, it indicates that complete energy storage or external power supply is required; when When the two are out of phase, it indicates the most severe time mismatch situation.

[0091] Based on the renewable energy resource potential and load demand data obtained in step S1, a quantitative assessment is made of the spatial mismatch between renewable energy distribution and load distribution:

[0092] For region s, calculate the proportion of its renewable energy resource potential to the entire system. :

[0093] ;

[0094] The proportion of load demand in computing region s to the total system load. :

[0095] ;

[0096] In the formula, S represents the total number of regions, and s represents the region index. Reflecting the renewable energy resource endowment of region s, Reflects the load demand share of region s;

[0097] To reflect the greater impact of spatial mismatch in areas far from the power grid center, distance weighting is introduced; the load-weighted geometric center coordinates of the entire system are calculated:

[0098] ;

[0099] Calculate the distance from region s to the load center The spatial weighting coefficients are obtained by using the great circle distance formula and normalizing it. :

[0100] ;

[0101] Calculate the spatial mismatch degree (SMD):

[0102] ;

[0103] In the formula, This indicates the deviation between the proportion of renewable energy and the proportion of load in region s. The larger the deviation, the more severe the spatial mismatch between supply and demand in that region. The spatial weighting amplifies the mismatch effect in areas far from the load center; the SMD value ranges from [0,2]. The larger the SMD, the more severe the overall spatial mismatch of the system, requiring more cross-regional power transmission.

[0104] In one embodiment of the present invention, step S3 includes:

[0105] Temporal mismatch and spatial mismatch are not independent entities; temporal mismatch in high-load areas has a greater impact on system stability. This invention establishes a spatiotemporal coupling mismatch degree model to quantify the coupling effect between temporal mismatch and spatial load distribution, and integrates spatial mismatch degree to form a unified spatiotemporal mismatch index.

[0106] Calculate the average load of the entire system:

[0107] ;

[0108] For region s, calculate its total load. And calculate the load deviation coefficient. :

[0109] ;

[0110] This coefficient reflects the degree to which the load level of region s deviates from the average level, with high-load regions... Larger;

[0111] Introducing time mismatch weights Spatial mismatch weights and coupling coefficient Calculate the spatiotemporal coupling mismatch degree STCM:

[0112] ;

[0113] The first term represents the spatial weighted distribution of temporal mismatch. The first term is the load deviation amplification factor, which makes the temporal mismatch in high-load areas contribute more to the overall index; the second term is the spatial mismatch degree (SMD), which directly reflects the spatial mismatch between resources and load; weighting coefficients. Ensure the normalization of indicators; in areas with high load demand, the supply and demand gap caused by time mismatch is more difficult to bridge through local adjustment, requiring more energy storage configuration or cross-regional support, thus amplifying its impact. At the same time, spatial mismatch directly determines the demand for cross-regional power transmission.

[0114] This spatiotemporal coupling mismatch model comprehensively considers the time dimension, the spatial dimension, and the coupling effect between the two, providing a quantitative evaluation index for the subsequent coordinated configuration optimization of energy storage to bridge time mismatch and power transmission to alleviate spatial mismatch.

[0115] It should be noted that the time mismatch degree in this invention is calculated separately for each region, and the output is a time mismatch degree vector. The first is used to measure the degree of mismatch between the renewable energy output curve and the load curve in a single region in the time dimension; the second is the spatial mismatch degree (SMD), which is calculated for the entire system, resulting in a scalar value, and is used to measure the degree of mismatch between the distribution of renewable energy resources and the distribution of load demand in the entire system in the spatial dimension; the third is the spatiotemporal coupling mismatch degree (STCM), which is also calculated for the entire system, but takes into account the time mismatch of each region, and outputs a scalar value, which comprehensively considers time mismatch, spatial mismatch, and the coupling effect of the two; when calculating STCM, the time part summarizes the time mismatch of all regions, and the time mismatch of high-load regions is amplified, while the spatial part directly uses the spatial mismatch degree of the entire system.

[0116] like Figure 2 As shown, in one embodiment of the present invention, step S4 includes:

[0117] To address the problem of multi-timescale decision coupling from long-term planning to short-term scheduling, as well as the problem of spatial coordination among multiple regions, a two-layer nested optimization framework is established: the outer layer is a multi-scale rolling optimization framework in the time dimension to handle the decision evolution at different time granularities, and the inner layer is a distributed optimization framework in the spatial dimension to handle coordinated decisions among multiple regions.

[0118] The optimization problem is decomposed into three scale levels along the time dimension:

[0119] Optimization time domain of long-term planning layer Optimize granularity for 1 year For January, a total of For each time period, the decision variable is the increase in installed power generation capacity in each region. Increase in transmission line capacity Incremental energy storage capacity ;

[0120] Optimization of the intermediate scheduling layer in the time domain Optimize granularity for 7 days For 1 hour, in total For each time period, the decision variable is the power generation output of each region during each time period. Energy storage charging and discharging power Inter-regional power transmission ;

[0121] Short-term real-time layer optimization time domain Optimize granularity for 4 hours It lasts for 15 minutes, in total For each time period, the decision variable is the actual output adjustment. Energy storage adjustment amount Transmission adjustment amount ;

[0122] Establish coupling relationships between time layers; output of the long-term planning layer , , As the capacity upper limit constraint of the medium-term dispatch layer; the medium-term dispatch layer performs rolling optimization before each day, only executing the decision for the first time period, and passing the power setpoint for that time period to the short-term real-time layer; the short-term real-time layer makes fine adjustments based on the medium-term dispatch plan according to the real-time load and renewable energy output deviation;

[0123] Define a rolling optimization mechanism based on the time dimension; the long-term planning layer performs annual optimization at the initial planning moment to obtain the installed capacity deployment plan for each future month, only executing the construction decision for the first month, and using the plans for the remaining months as predictive solutions; the medium-term scheduling layer performs weekly optimization at the day-ahead time of each day to obtain the hourly operation plan for the next 7 days, only executing the scheduling plan for the first day, and using the plans for the remaining 6 days as predictive solutions; the short-term real-time layer performs 4-hour optimization at the real-time time of each hour, only executing the adjustment amount for the first period; each time it rolls over, the optimization problem is resolved based on the latest predictive information, but only the decision for the first period is executed, and the solutions for the remaining periods are used as predictive decisions for the next layer to refer to;

[0124] This multi-scale decomposition framework with a time dimension avoids the shortcomings of traditional single-time-scale optimization methods in dealing with the coupling of multi-time-scale decisions, and achieves coordination between long-term capacity configuration decisions and short-term operation scheduling decisions.

[0125] In one embodiment of the present invention, step S5 includes:

[0126] The spatiotemporal coupling mismatch degree calculated in step S3 is incorporated into the optimization objective function to establish a multi-objective optimization model oriented towards reducing spatiotemporal mismatch.

[0127] Taking the intermediate scheduling layer optimization problem defined in step S4 as an example, its objective function is defined as follows:

[0128] ;

[0129] The first item, Cost, represents the system operating cost:

[0130] ;

[0131] The second term is the spatiotemporal coupling mismatch term, which defines the dynamic spatiotemporal mismatch degree by introducing the bridging effect of energy storage configuration:

[0132] ;

[0133] in, For decision variables in each region, The total energy storage capacity of region s is expressed in megawatt-hours (MWH), including both existing and new capacity. The time-series surplus electricity from renewable energy sources is expressed in megawatt-hours (MWh). The energy storage bridging factor is set to 0.8. To prevent small amounts from being divided by zero, and The actual renewable energy output and load demand of region s at time t; the more sufficient the energy storage configuration, the higher the surplus relative to renewable energy, the greater the degree to which the impact of time mismatch is bridged, and the lower the contribution of time mismatch to the objective function; the spatial mismatch degree SMD directly reflects the transmission demand.

[0134] The third item is the inter-regional power transmission power item. For the total transmission power Norms are used to promote local consumption and reduce unnecessary cross-regional transmission;

[0135] Weighting coefficient and Used to balance the importance of the three objectives;

[0136] To guide the flow of transmitted power from areas with severe time mismatch to high-load areas, a time-space mismatch compensation incentive mechanism is established; and the time-space mismatch compensation weights for inter-regional power transmission are defined. :

[0137] ;

[0138] In the formula, For compensation coefficient, and Let i be the load percentage for regions i and j.

[0139] When data is transmitted from region i to region j with a higher load share, and the time mismatch between the two regions is severe, positive compensation is given to incentivize this transmission behavior that bridges the spatiotemporal mismatch.

[0140] This multi-objective optimization function achieves spatiotemporal coordinated optimization through three dimensions: temporal coordination is achieved through a dynamic STCM term; the more sufficient the energy storage capacity, the higher the time-series surplus relative to renewable energy, thus mitigating the impact of time mismatch, reducing STCM, and optimizing the objective function; spatial coordination is achieved through… The project and compensation mechanism promote local consumption while prioritizing incentives for transmission from areas with severe time mismatch to high-load areas, achieving spatial supply and demand matching; spatiotemporal coupling is achieved through the STCM project. This reflects the higher priority of time mismatch in high-load areas, and the coupling guidance of compensation weights on time mismatch and spatial load distribution.

[0141] In one embodiment of the present invention, step S6 includes:

[0142] For the optimization problem at each time scale level defined in step S4, the Alternating Directional Multiplier Method (ADMM) is used for spatial decomposition to achieve distributed parallel solution in each region; taking the mid-term scheduling layer as an example, the global optimization problem is expressed as:

[0143] ;

[0144] ;

[0145] ;

[0146] In the formula, Let be the local decision variables for region s, containing the region's... The power generation output, energy storage charging and discharging power, and local load within a given time period; z is a global coordination variable representing the transmission power vector between all regions. ,in This represents the power transmitted from region i to region j. Let be the local cost function for region s, including generation costs, energy storage operating costs, etc. The global cost function includes transmission loss costs. and The incidence matrix represents the coupling relationship between local and global variables, and the constraints. This represents the power balance condition for region s, which is that the sum of power generation, energy storage discharge, and imported power transmission within the region equals the sum of load, energy storage charging, and imported power transmission. Let be the local constraint set for region s, including upper limits for power generation capacity, constraints on the state of charge of energy storage, etc. This is a global constraint set, including transmission line capacity constraints.

[0147] Introducing Lagrange multiplier vectors Power balance constraints and penalty parameters for region s Construct the augmented Lagrangian function:

[0148] ;

[0149] In the formula, This represents the augmented Lagrange function. Let S be the decision variable for each region, and S be the total number of regions. Let Lagrange multiplier vectors be used. For the Lagrange term, As a penalty item, Let be the transpose of the Lagrange multiplier vector of region s. This is the penalty parameter.

[0150] Define the iterative scheme of the alternating direction multiplier method; the k-th iteration includes a local variable optimization step, a global variable reconciliation step, and a dual variable update step:

[0151] (1) Local variable optimization steps:

[0152] fixed and Each region s solves its local subproblems in parallel:

[0153] ;

[0154] (2) Global variable coordination steps:

[0155] Fix all and Update global variables:

[0156] ;

[0157] This introduces the spatiotemporal mismatch compensation weight defined in step S5. For the compensation incentive term, the negative sign indicates a reduction in the objective function, which is equivalent to a gain;

[0158] (3) Dual variable update steps:

[0159] Update the Lagrange multipliers:

[0160] ;

[0161] This spatially distributed optimization framework enables parallel solutions to local subproblems in different regions, achieving global coordination through iteration. This avoids the problem of excessively high computational dimensionality in large-scale, multi-region systems, which is a problem with centralized optimization methods.

[0162] In one embodiment of the present invention, step S7 includes:

[0163] The penalty parameter for the alternating direction multiplier method in step S6 An adaptive adjustment strategy based on spatiotemporal mismatch is designed; for region s, the adjustment strategy is based on its temporal mismatch degree. Based on distance information from other areas, set the initial penalty parameters;

[0164] For areas with transmission line connections Set the initial penalty parameters between regions. :

[0165] ;

[0166] In the formula, The baseline penalty parameter is set to 10. This is the time mismatch coupling coefficient, with a value of 0.5, used to amplify the coordination needs when both regions have severe time mismatches. This is a distance coefficient, with a value of 0.3, used to amplify transmission constraints between long-distance regions. This represents the maximum inter-region distance in the system.

[0167] Regions with severe time mismatches require stronger inter-regional power mutual assistance, so the penalty parameter is increased to strengthen coordination constraints. Regions with long distances have more stringent transmission losses and constraints, and similarly require a larger penalty parameter.

[0168] During the ADMM iteration process, the penalty parameter is dynamically adjusted based on the original residual and dual residual of the kth round.

[0169] For region pair (i,j), define the original residual. Define dual residual Set residual ratio threshold and adjustment factor The adaptive update rule is:

[0170] ;

[0171] When the original residual is much larger than the dual residual, it indicates that the power balance constraint is seriously violated, and the penalty parameter needs to be increased to strengthen the constraint satisfaction. When the dual residual is much larger than the original residual, it indicates that the algorithm is too conservative, and the penalty parameter can be reduced to accelerate convergence. This adaptive strategy combines physical-driven initialization based on spatiotemporal mismatch and algorithm-driven adjustment based on residual feedback, which improves the efficiency and robustness of spatial distributed solution.

[0172] In one embodiment of the present invention, step S8 includes:

[0173] In the ADMM iteration process defined in step S6, in addition to the traditional convergence conditions of the original residual and dual residual, a convergence criterion of spatiotemporal mismatch degree change is added to ensure that the algorithm not only satisfies the constraints, but also achieves optimized convergence of spatiotemporal mismatch degree.

[0174] After the k-th iteration, calculate the norms of the original residuals and dual residuals of the entire system:

[0175] ;

[0176] ;

[0177] Set the original residual threshold. and dual residual threshold ;

[0178] Based on the configuration scheme obtained in the kth iteration The corresponding spatiotemporal coupling mismatch degree is calculated using the formula in step S5. ; Calculate the absolute change in spatiotemporal mismatch degree between two adjacent iterations. Set the spatiotemporal mismatch convergence threshold ;

[0179] To prevent the algorithm from oscillating near the optimum, we check whether the trend of spatiotemporal mismatch is stable; we define a rate of change index. ,in To prevent small values ​​from being divided by zero; if in 3 consecutive iterations If the sign of the change remains unchanged, the rate of change is considered to be stable.

[0180] The algorithm determines convergence and terminates iteration when the following four conditions are met: , , And the stability condition of the rate of change is satisfied; when the number of iterations k exceeds the maximum number of iterations. Force termination and output of warning message;

[0181] When the convergence condition is met, the optimal configuration scheme for each region is output, including the installed power generation capacity, transmission line capacity, energy storage capacity, and corresponding operation strategy. Simultaneously, the spatiotemporal mismatch indices of the system are output, including the time mismatch degree, spatial mismatch degree, and spatiotemporal coupling mismatch degree for each region, as well as the optimization target value.

[0182] The improved convergence criterion, by adding the determination of the change in spatiotemporal mismatch degree and the stability of the rate of change, ensures that the algorithm not only meets physical constraints such as power balance, but also achieves sufficient optimization of spatiotemporal mismatch degree, thus avoiding premature termination of the algorithm when the spatiotemporal mismatch degree has not decreased sufficiently.

[0183] It should be noted that steps S6-S8 above are illustrated using the intermediate scheduling layer as an example. For the implementation of the long-term planning layer and the short-term real-time layer, the specific algorithm optimization process is similar to that of the intermediate scheduling layer, but the objective function and constraints are different, as follows:

[0184] The multi-objective optimization function for the long-term planning layer is:

[0185] ;

[0186] in, The annual investment cost includes the construction costs of power generation capacity, transmission lines, and energy storage. The calculation formula is as follows: Where m is the month index, The unit investment cost of newly built power generation facilities, The unit investment cost of newly built energy storage facilities, The unit investment cost for newly built transmission lines; This is an annualized operating cost estimate based on a representative daily operating scenario; The spatiotemporal coupling mismatch degree in the planning phase is calculated using the same formula as the mid-term scheduling layer. The difference lies in that the time index t is calculated monthly based on the surplus renewable energy power for each month. In this context, t represents a monthly time point. To balance the weights of investment costs and temporal and spatial mismatch;

[0187] The constraints include: the gradual increase constraint of installed capacity in each region. Ensure that capacity increases monotonically month by month; overall investment budget constraint , Investment budget cap; annual carbon emission constraints , Let be the power generation of region s in month m. The annual carbon emission cap is set; the decision variables are the monthly increases in installed power generation capacity, transmission line capacity, and energy storage capacity in each region.

[0188] The multi-objective optimization function for the short-term real-time layer is:

[0189] ;

[0190] in, The first item is the adjustment cost, used to adjust the cost coefficient. The second term is the tracking bias penalty, which is used to track the weighting coefficients. The power setting value is transmitted to the intermediate scheduling layer. The adjusted actual output For the power generation output planned by the medium-term dispatch layer, This represents the current power generation increment of region s at time t; The weighting coefficient for spatiotemporal mismatch is relatively small. The calculation formula for the real-time spatiotemporal coupling mismatch is the same as that for the intermediate scheduling layer. The difference is that the time index t is counted in 15-minute granularity, and the time range covers 16 time periods in the next 4 hours.

[0191] Constraints include: adjustment range limits Ensure adjustments are within a controllable range; maintain continuity of energy storage state of charge (SOC) at a 15-minute granularity. ,in, This represents the percentage of the energy storage state of charge in region s at time t. Hour, Let be the adjustment amount of energy storage power in region s at time t. The total energy storage capacity of region s; real-time power balance constraints. ,in The actual load demand of region s at time t. Let be the power transmission from region s to region j at time t; the decision variables are the power output adjustment, energy storage adjustment, and transmission adjustment for each region at each time period.

[0192] like Figure 3 As shown, in one embodiment of the present invention, step S9 includes:

[0193] Based on the spatiotemporal two-layer nested framework established in step S4 and the ADMM solution algorithms designed in steps S6 to S8, a complete rolling optimization process is executed. The outer layer is a temporal multi-scale rolling optimization, and the inner layer is a spatially distributed ADMM iterative solution.

[0194] Long-term planning level at the initial stage of planning The annual optimization is performed based on load and renewable energy resource forecasts for the next year, constructing an annual optimization problem. The decision variable is the installed capacity increment of each region, with a time granularity of monthly. The objective function is the net present value of investment cost plus operating cost, with constraints including carbon emission constraints and investment budget constraints. The ADMM algorithm in steps S6 to S8 is invoked. In the k-th round of ADMM iteration, each region solves its local subproblems in parallel, then globally coordinates and updates the transmission power and applies the compensation mechanism in step S5 to update the dual variables. The adaptive parameter strategy in step S7 and the convergence criterion in step S8 are applied, iterating until convergence is achieved. The installed capacity deployment plan for the next 12 months is obtained. Only the construction decision for the first month is executed, and the plans for the remaining 11 months are used as predictive schemes. The installed capacity increment for the first month is added to the existing capacity as the capacity upper limit for the medium-term dispatch layer.

[0195] The mid-term scheduling layer performs weekly optimization at the day-ahead time each day, constructing a weekly optimization problem based on hourly load forecasts and renewable energy output forecasts for the next 7 days. Decision variables include hourly power generation output, energy storage charging and discharging, and inter-regional transmission power in each region, with the capacity limit passed from the long-term planning layer. The objective function includes cost, spatiotemporal mismatch, and transmission power. The ADMM algorithm is invoked for distributed iterative solution in the spatial dimension. Each region optimizes its local operation strategy in parallel, while globally coordinating and updating inter-regional transmission power. Adaptive penalty parameters and spatiotemporal mismatch convergence criteria are applied, iterating until convergence is achieved, resulting in a 168-hour operation plan for the next 7 days. Only the scheduling plan for the first 24 hours is executed, with the remaining 6 days' plans used as predictive schemes. The hourly power setpoint for the first day is passed to the short-term real-time layer. The optimization problem is updated and resolved daily based on the latest forecast information.

[0196] The short-term real-time layer performs 4-hour optimization at real-time times every hour, acquiring 15-minute load and actual renewable energy output data for the next 4 hours, and calculating the deviation from the medium-term scheduling layer's plan. A real-time correction optimization problem is constructed, with the output adjustment amount as the decision variable, and constraints including the adjustment amount not exceeding ±10% of the medium-term plan. The objective function is deviation penalty plus adjustment cost. The ADMM algorithm is used for rapid solution; due to the short time domain and limited adjustment amount, convergence is generally achieved in 10 iterations, coordinating real-time adjustments across regions spatially. Adjustment schemes for 16 time periods in the next 4 hours are obtained. Only the adjustment amount for the first 15 minutes is executed, and the system executes accordingly. The system is rolled over every hour to achieve rapid response to real-time fluctuations.

[0197] Through multiple rounds of rolling optimization and ADMM iterations at the long-term planning, medium-term scheduling, and short-term real-time levels, a spatiotemporal coordinated configuration scheme for the renewable energy supply-transmission-demand-storage system is finally output. The scheme includes the deployment sequence of power generation capacity, transmission line construction, and energy storage capacity for each region in future years, as well as hourly power generation output, energy storage charging and discharging curves, and inter-regional transmission power curves for typical operating days. It also outputs performance indicators, including total system cost, renewable energy absorption rate, carbon emissions, and spatiotemporal mismatch indicators, including the temporal mismatch degree of each region and its reduction percentage compared to before optimization, the reduction in spatial mismatch degree, and the optimization magnitude of spatiotemporal coupling mismatch degree. This scheme can provide a scientific basis for power system planning and operation decisions.

[0198] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system, characterized in that, include: S1. Acquire spatial topology data, historical time series data and power system parameters for multiple regions. Spatial topology data includes regional geographic coordinates, inter-regional transmission line connections and inter-regional distances. Historical time series data includes load demand data and renewable energy output data for each region. Power system parameters include existing installed capacity, equipment investment costs and operating costs. S2. Based on the renewable energy output data and load demand data of each region, the time mismatch degree of each region is calculated using the similarity calculation method. The energy-load ratio is calculated based on the renewable energy resource potential and load demand data of each region. The spatial weighting coefficient is calculated based on the load demand data of each region and the distance between regions. The spatial mismatch is calculated using the weighted deviation method to obtain the spatial mismatch and the load share of each region. S3. The time mismatch is weighted and amplified according to the load ratio of each region. The weighted time mismatch is then fused with the spatial mismatch to obtain the spatiotemporal coupling mismatch. S4. Decompose the optimization problem into a long-term planning layer, a medium-term scheduling layer, and a short-term real-time layer in the time dimension, and establish the capacity constraint relationship and power transfer relationship between each time layer. S5. Construct multi-objective optimization functions for each time level; S6. Decompose the optimization problem of each time layer in the spatial dimension, and decompose the global optimization problem into local optimization sub-problems of each region and global coordination problem; S7. Initialize the iteration parameters based on the time mismatch degree and the distance between regions. Use a distributed iteration method to alternately solve the local optimization subproblems and the global coordination problems of each region. During the iteration process, dynamically adjust the iteration parameters based on the optimization residuals. The distributed iteration method includes local variable optimization steps, global variable coordination steps, and dual variable update steps. S8. Determine the convergence between the local optimization subproblem and the global coordination problem, as well as the stability of the change in the spatiotemporal coupling mismatch. When both convergence and stability meet the preset conditions, output the configuration result of the current time layer. S9. Execute steps S6 to S8 sequentially from long to short time scale for the long-term planning layer, the medium-term scheduling layer, and the short-term real-time layer. Execute the decision for the first time period at each time layer and use it as the constraint condition for the next time layer to obtain the power generation capacity, transmission line capacity, energy storage capacity, and operation strategy for each region. In step S5, the multi-objective optimization function at each time level includes a spatiotemporal coupling mismatch term, which is: ； In the formula, For spatiotemporal coupling mismatch degree, Let S be the decision variable for each region, and S be the total number of regions. Let be the total energy storage capacity of region s. This is the energy storage bridging coefficient. To prevent small amounts from being divided by zero, and For the renewable energy output and load demand of region s at time t, Let be the time mismatch degree of region s. For spatial mismatch degree, For time mismatch weights, For spatial mismatch weights, , Let be the load deviation coefficient for region s. is the coupling coefficient.

2. The spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system according to claim 1, characterized in that, In step S2, the temporal mismatch degree of each region is calculated using a similarity calculation method, specifically including: Normalize the renewable energy output and load demand data for region s, and calculate the normalized renewable energy output. and load demand ; The temporal mismatch degree of region s is calculated using the inverse index of cosine similarity. : ； In the formula, t is the time index. This represents the total number of historical time periods. The value range is [0, 2].

3. The spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system according to claim 1, characterized in that, In step S2, the spatial mismatch degree is calculated using the weighted bias method, specifically including: The proportion of renewable energy resource potential in region s to the total system. and the proportion of load demand to the entire system ; The spatial weighting coefficients are obtained by calculating the load-weighted geometric center coordinates of the entire system and then normalizing the distance from region s to the load center. ; Calculate the spatial mismatch degree (SMD): ； In the formula, S is the total number of regions, and s is the region index.

4. The spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system according to claim 1, characterized in that, In step S3, the calculation formula for the spatiotemporal coupling mismatch degree (STCM) is obtained as follows: ； In the formula, For time mismatch weights, For spatial mismatch weights, , Let be the load deviation coefficient for region s. The coupling coefficient is... Let be the time mismatch degree of region s. This represents the spatial mismatch degree.

5. The spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system according to claim 1, characterized in that, In step S4, the optimization problem is decomposed into a long-term planning layer, a medium-term scheduling layer, and a short-term real-time layer in the time dimension. Specifically, the optimization time domain of the long-term planning layer is 1 year, the optimization granularity is 1 month, and the decision variables are the increase in installed power generation capacity, the increase in transmission line capacity, and the increase in energy storage capacity in each region; the optimization time domain of the medium-term scheduling layer is 7 days, the optimization granularity is 1 hour, and the decision variables are the power generation output, energy storage charging and discharging power, and inter-regional transmission power in each region at each time period; the optimization time domain of the short-term real-time layer is 4 hours, the optimization granularity is 15 minutes, and the decision variables are the power output adjustment, energy storage adjustment, and transmission adjustment.

6. The spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system according to claim 1, characterized in that, In step S6, the global optimization problem is decomposed into local optimization subproblems for each region and a global coordination problem. The augmented Lagrangian function is constructed using the alternating direction multiplier method. ； In the formula, This represents the augmented Lagrange function. Let S be the decision variable for each region, S be the total number of regions, and z be the global coordination variable, representing the inter-regional power transmission capacity. Let Lagrange multiplier vectors be used. Let be the local cost function of region s. For the global cost function, Let be the transpose of the Lagrange multiplier vector of region s. and It is an incidence matrix. This is the penalty parameter.

7. The spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system according to claim 6, characterized in that, In step S7, the iteration parameters are initialized based on the time mismatch degree of each region and the distance between regions. For region pairs (i,j) with transmission line connections, the initial penalty parameters are... for: ； In the formula, As the baseline penalty parameter, For time mismatch coupling coefficient, Let be the time mismatch degree of region i. Let j be the time mismatch degree of region j. This is the distance coefficient. Let be the distance between region i and region j. This represents the maximum inter-regional distance in the power system. During the iteration process, the iteration parameters are dynamically adjusted based on the optimized residual, and the original residual is defined. and dual residual Update the penalty parameter based on the residual ratio. : ； In the formula, k is the iteration round. The residual ratio threshold, and This is an adjustment factor.

8. A spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system according to claim 7, characterized in that, In step S8, the convergence between the local optimization subproblem and the global coordination problem, as well as the stability of the change in spatiotemporal coupling mismatch, are determined, specifically including: Calculate the original residual norm of the entire system after the k-th iteration. and dual residual norm ; Calculate the change in spatiotemporal coupling mismatch between two adjacent iterations. STCM stands for Spatiotemporal Coupling Mismatch. Calculate the rate of change of spatiotemporal coupling mismatch and determine its stability; when , , The iteration terminates when the rate of change stabilizes, where , and These are the original residual threshold, the dual residual threshold, and the spatiotemporal mismatch change threshold, respectively.

9. A spatiotemporal coordinated configuration method for a renewable energy supply-transmission-demand-storage system according to claim 5, characterized in that, In step S9, rolling optimization is performed on the long-term planning layer, the medium-term scheduling layer, and the short-term real-time layer in descending order of time scale. Specifically, this includes: The long-term planning layer performs annual optimization at the initial planning stage to obtain the installed capacity deployment plan for each future month. Only the construction decision for the first month is executed, and the plans for the remaining months are used as predictive schemes. The installed capacity increment of the first month is passed to the medium-term scheduling layer as the capacity limit. The mid-term scheduling layer performs weekly optimization at the day-ahead time of each day to obtain hourly operation plans for the next 7 days. Only the scheduling plan for the first day is executed, and the plans for the remaining 6 days are used as predictive schemes. The power setpoint for the first day is passed to the short-term real-time layer. The short-term real-time layer performs 4-hour optimizations at the real-time point of each hour, and only performs adjustments for the first time period.

Images