A method, device and medium for optimizing the clearing of an out-of-province power supply equivalent market

Abstract

The present invention discloses an optimization method, device and medium for clearing out-of-province power supply equivalent market, the method comprising: optimizing the tie line power function of each provincial power grid to obtain the optimal tie line output power; establishing an equivalent quantity-price curve according to the power supply out of the province; introducing the optimal tie line output power into the provincial optimal unit output model for calculation to obtain the average network access cost of the virtual node; calculating the unit output arrangement according to the optimal tie line output power, the equivalent quantity-price curve and the average network access cost, and configuring the output of each provincial power grid. The present invention proposes an optimization method, device and medium for clearing out-of-province power supply equivalent market, which ensures the balance of power transmission and exchange by calculating the optimal tie line output power; by calculating the network access cost and the equivalent quantity-price curve, the optimal configuration of power resources can be achieved, which can solve the problem that the optimal configuration level of power resources is low and it is difficult to ensure the rationality of inter-provincial power transactions.

Description

Technical Field

[0001] The present invention relates to the field of electric power market-oriented transactions, and in particular to an out-of-province power source equivalent market clearing optimization method, device and medium. Background Art

[0002] At present, the process of sharing, mutual assistance and optimal allocation of power resources in the inter-provincial and intra-provincial markets has been continuously promoted, which has enabled more and more market players to access the grid nodes to participate in the balance of power supply and demand. In this context, the continuous expansion of the scale of the grid has brought great potential optimization benefits to the dispatching operation and market clearing. The realization of safe dispatching of power and efficient market optimization clearing across the entire network has become the main goal of the next development of grid dispatching. For large-scale power resource allocation schemes, domestic and foreign scholars mostly use linear programming, mixed integer programming or heuristic algorithms to deal with complex nonlinear and non-convex power resource allocation problems.

[0003] However, due to the continuous expansion of the scale of the power grid, the number of power facilities, lines and units involved has increased dramatically. These complex factors make it difficult for dispatchers to make comprehensive and accurate decisions through traditional power resource allocation methods when faced with multiple constraints and goals; at the same time, power supply is also constrained by multiple factors such as unit operating status, maintenance plans, fuel supply, etc. This uncertainty in supply and demand increases the difficulty of resource scheduling, resulting in a low level of optimal allocation of power resources, making it difficult to ensure the fairness, efficiency and stability of cross-provincial power transactions. Summary of the invention

[0004] The present invention provides an out-of-province power source equivalent market clearing optimization method, device and medium to solve the problem that the optimization configuration level of power resources is low and it is difficult to ensure the rationality of inter-provincial power transactions.

[0005] Obtain network data of provincial power grids;

[0006] Establishing a provincial optimal unit output model based on the network data;

[0007] Optimizing the tie line power function of each provincial power grid to obtain the optimal tie line output power;

[0008] By simplifying and equalizing the power sources outside the province, an equivalent quantity-price curve of the virtual node is established;

[0009] The optimal tie line output power is introduced into the provincial optimal unit output model for calculation to obtain the average network cost of the virtual node;

[0010] The unit output arrangement is calculated based on the optimal interconnection line output power, the equivalent quantity-price curve and the average grid access fee, and the output of each provincial power grid is configured based on the unit output arrangement.

[0011] The present invention establishes a provincial optimal unit output model by collecting and analyzing network data, which fully considers the actual situation and needs within each provincial power grid. By simplifying and equivalence processing of power sources outside the province, an equivalent quantity-price curve of a virtual node is established, which simplifies the complex inter-provincial power trading problem and improves the accuracy and efficiency of the calculation. The optimal tie line output power is introduced into the provincial optimal unit output model for calculation, which can accurately predict and evaluate the impact of different scheduling schemes on the power system, and provide strong data support for making scientific scheduling decisions. For the unit output arrangement, the introduction of the optimal tie line output power ensures the coordination and balance of power transmission and exchange between different provincial power grids; the economic factor of the network access fee is considered, so that the unit output arrangement is more in line with market laws, which helps to reduce the cost of power transmission and trading; by comparing the equivalent quantity-price curves of different virtual nodes, it can be identified which power sources are more competitive under the current market conditions, so as to give priority to the output of these power sources and achieve the optimal allocation of power resources.

[0012] Compared with the prior art, the present invention ensures the balance of power transmission and exchange by calculating the optimal interconnection line output power; helps to reduce the cost of power transmission and transaction by calculating the network transmission fee; and can achieve optimal allocation of power resources by calculating the equivalent quantity-price curve, so that effective and reasonable unit output arrangements can be generated to carry out resource allocation, thereby solving the problem of low level of optimal allocation of power resources and difficulty in ensuring the rationality of inter-provincial power transactions.

[0013] As a preferred solution, a provincial optimal unit output model is established based on the network data, specifically:

[0014] The nodes and branches in each provincial power grid are equivalent to a single node, and an equivalent admittance matrix of the provincial power grid is obtained;

[0015] By combining the equivalent admittance matrix and the inter-provincial contact line node admittance matrix, an equivalent global virtual network admittance matrix is ​​generated;

[0016] Based on the global virtual network admittance matrix, with the goal of minimizing the unit output cost, combined with preset constraints, the provincial optimal unit output model is established.

[0017] This optimal solution greatly reduces the number of state variables and model dimensions of the system by simplifying the complex provincial power grid into a single node, thereby significantly reducing the computational complexity and time cost. The equivalent global virtual network admittance matrix obtained by splicing not only includes the electrical characteristics within each province, but also reflects the inter-provincial electrical connections through inter-provincial contact lines, which enables the model to flexibly handle inter-provincial power flows and interactions, providing strong support for the formulation of a more comprehensive power grid dispatching strategy.

[0018] As a preferred solution, based on the global virtual network admittance matrix, with the goal of minimizing the unit output cost, combined with preset constraints, the provincial optimal unit output model is established, specifically:

[0019] According to the total number of units in each provincial power grid and the output of the units at different time periods, a first objective function is established with the goal of minimizing the unit output cost;

[0020] A first set of constraint conditions is established based on load balance between nodes, operation restrictions of the units and line transmission capacity;

[0021] Based on the global virtual network admittance matrix, according to the first objective function and the first constraint condition, the provincial optimal unit output model is established.

[0022] This optimization scheme takes minimizing the unit output cost as its goal, which can ensure that while meeting the electricity demand, the power generation cost is reduced as much as possible and the economic benefits of the power company are improved; in addition, the provincial optimal unit output model takes into account a variety of constraints, including load balance between nodes, unit operation restrictions and line transmission capacity restrictions. These constraints ensure the feasibility and safety of the optimization results and avoid problems in the actual operation of the scheduling scheme caused by ignoring certain constraints; in addition, since the provincial optimal unit output model is established based on the global virtual network admittance matrix, this matrix not only contains the electrical characteristics within the provincial power grid, but also reflects the cross-provincial electrical connections through inter-provincial interconnection lines, which enables the model to more accurately reflect the actual operation of the power grid and improve the accuracy of the optimization model.

[0023] As a preferred solution, the tie line power model of each provincial power grid is optimized to obtain the optimal tie line output power, specifically:

[0024] Based on the tie line node output power of each provincial power grid, an optimal tie line power model is constructed with the goal of minimizing the tie line node output price;

[0025] Outputting the optimal output power of each tie line according to the optimal tie line power model;

[0026] Updating the optimal output power of the optimal tie-line power model by training a preset price response function;

[0027] If the adjustment amount of the power injected by all tie-line nodes reaches a preset convergence condition, the updated optimal output power is used as the optimal tie-line output power.

[0028] This preferred solution aims to minimize the output price of the tie line nodes and construct an optimal tie line power model, which helps to achieve the optimal configuration of power resources while ensuring the safety of power grid operation, thereby saving costs. The optimal tie line power model is updated by training the price response function. This method enables the model to be dynamically adjusted according to price changes in the actual electricity market, ensuring that the model can still maintain its optimization performance in the face of market price fluctuations. By introducing convergence conditions to determine whether the adjustment amount of the power injected into the tie line node has reached a stable state, this ensures the stability and reliability of the model optimization process and avoids infinite loops or divergences.

[0029] As a preferred solution, based on the tie line node output power of each provincial power grid, with the goal of minimizing the tie line node output price, an optimal tie line power model is constructed, specifically:

[0030] According to the number of tie line nodes of the provincial power grids, combined with the corresponding relationship between the change in the output power of the tie line nodes and the price, a second objective function is established with the goal of minimizing the output price of the tie line nodes;

[0031] A second set of constraint conditions is established according to the transmission capacity of the tie line and the output power of the node;

[0032] Based on the power generation power in each time period and the equivalent injection power of the interconnection line nodes obtained by optimizing the provincial optimal unit output, the optimal interconnection line power model is established according to the second objective function and the second constraint condition.

[0033] This preferred solution combines the corresponding relationship between the change in the output power of the interconnection line node and the price, so that the second objective function can more accurately reflect the changes in the price and supply and demand relationship in the electricity market. And the second constraint condition includes the transmission capacity of the interconnection line and the restriction of the node output power. These comprehensive constraints ensure the feasibility and safety of the optimization model. In addition, the provincial optimal unit output model fully considers the results of the provincial optimal unit output optimization during the establishment process, that is, the power generation power in each period and the equivalent injection power of the interconnection line node. This further model establishment method based on the existing optimization results will help to achieve a higher level of coordination and optimization, and improve the operating efficiency and economy of the entire power grid.

[0034] As a preferred solution, the price response function is specifically:

[0035] According to the initial power of the tie lines of the provincial power grids, the provincial optimal unit output model is optimized to obtain a model optimization result;

[0036] According to the optimization results of the model, the marginal price of the tie line power output and the equivalent injection power of the tie line node at different time periods and different power generation load levels are calculated to obtain the calculation results;

[0037] The power generation state of the current unit is corrected based on the calculation result, and the price response function of the unit after correction is obtained.

[0038] This preferred solution provides detailed information about power trading and market price changes by calculating the marginal price of tie-line power output and the equivalent injected power of tie-line nodes at different time periods and different power generation load levels. The calculation of marginal price helps to understand the sensitivity of prices in the power market to changes in supply and demand, while the calculation of equivalent injected power helps to evaluate the power flow at each node in the power grid. The power generation status of the current unit can be corrected based on the calculation results to improve the price response reliability of the price response function.

[0039] As a preferred solution, by simplifying and equalizing the power sources outside the province, an equivalent quantity-price curve of the virtual node is established, specifically:

[0040] Extract features from a number of power quantity and price curves of the provincial power grids to obtain a data point set;

[0041] Based on the data point set, quantifying the difference between each quantity-price curve and the reference curve by Euclidean distance, eliminating the power quantity-price curves that are continuously higher than the reference curve in preset key features, and obtaining the remaining first power quantity-price curve set;

[0042] Dividing the plurality of power quantity-price curves by clustering to obtain clustering results, and eliminating quantity-price curves that do not meet preset conditions according to the clustering results to obtain the remaining second power quantity-price curve set;

[0043] An equivalent quantity-price curve of the virtual node is established according to the first power quantity-price curve set and the second power quantity-price curve set.

[0044] This preferred solution uses Euclidean distance to quantify the difference between the power quantity and price curve and the reference curve, which can accurately measure the similarity or difference between the curves, and provide a reliable basis for the subsequent elimination of abnormal curves. By eliminating power quantity and price curves that are continuously higher than the reference curve in preset key features, those abnormal or unreasonable data points can be excluded, thereby avoiding their adverse effects on subsequent analysis. The use of clustering to divide the power quantity and price curves can not only further identify groups of curves with similar characteristics, but also help to discover potential power market structures or patterns, and by eliminating quantity and price curves that do not meet the preset conditions, the quality of the data can be further improved.

[0045] As a preferred solution, the plurality of power quantity-price curves are divided by clustering to obtain clustering results, specifically:

[0046] Discretizing the plurality of power quantity-price curves to obtain a first data point set;

[0047] Based on the preset number of clusters to be divided, the first data point set is subjected to standardization and clustering processing to obtain a second data point set;

[0048] In the second data point set, updating the corresponding cluster centroids according to the means of all points in the clusters;

[0049] If the centroid change is less than a preset threshold or reaches a preset number of iterations, the power quantity price curves corresponding to the plurality of clusters are taken as the clustering result.

[0050] This preferred solution discretizes the power quantity and price curve into a data point set, which can significantly reduce the amount of data, reduce the computational complexity of subsequent processing, and improve processing efficiency. The data point set is clustered based on the preset number of clusters, and the degree of clustering sophistication can be adjusted according to actual needs. By iteratively updating the cluster centroids and continuously optimizing the clustering results until the centroid change is less than the preset threshold or reaches the preset number of iterations, the iterative optimization process helps to reduce clustering errors and improve the accuracy and stability of clustering results.

[0051] As a preferred solution, the optimal tie line output power is introduced into the provincial optimal unit output model for calculation to obtain the average network cost of the virtual node, which is specifically:

[0052] Acquire simplified power data of each node according to the first power quantity-price curve set and the second power quantity-price curve set;

[0053] The optimal tie line output power is introduced into the provincial optimal unit output model, and the simplified power supply data of each node is used as the power supplier to calculate the power supply allocated to different units;

[0054] Based on the power sources allocated to the different units, the average grid-connection fee of the virtual node is calculated according to the grid-connection fee of the transmission cost.

[0055] In this preferred solution, the simplified power supply data of each node retains key information and reduces data redundancy, so the complexity and amount of calculation of data processing can be effectively reduced. By using the simplified power supply data of each node as the power supplier, the power supply allocated to different units is calculated. This process reflects the balance of supply and demand in the power market and the optimal allocation of power resources, which can reduce operating costs and transmission losses as much as possible.

[0056] The present invention also provides an out-of-province power supply equivalent market clearing optimization device, comprising a data module, a model module, a power module, a curve module, a cost module and a scheduling module;

[0057] Wherein, the data module is used to obtain network data of each provincial power grid;

[0058] The model module is used to establish a provincial optimal unit output model according to the network data;

[0059] The power module is used to optimize the tie line power function of each provincial power grid to obtain the optimal tie line output power;

[0060] The curve module is used to establish an equivalent quantity-price curve of a virtual node by simplifying and equivalent processing the power supply outside the province;

[0061] The cost module is used to introduce the optimal tie line output power into the provincial optimal unit output model for calculation to obtain the average network cost of the virtual node;

[0062] The dispatching module is used to calculate the unit output arrangement according to the optimal interconnection line output power, the equivalent quantity-price curve and the average grid access fee, and configure the output of each provincial power grid according to the unit output arrangement.

[0063] As a preferred solution, the model module includes an equivalent unit, a splicing unit and a model unit;

[0064] The equivalent unit is used to equate the nodes and branches in each provincial power grid to a single node, and obtain an equivalent admittance matrix after the provincial power grid is equivalent;

[0065] The merging unit is used to generate an equivalent global virtual network admittance matrix by merging the equivalent admittance matrix and the inter-provincial tie line node admittance matrix;

[0066] The model unit is used to establish the provincial optimal unit output model based on the global virtual network admittance matrix, with the goal of minimizing the unit output cost and combined with preset constraints.

[0067] As a preferred solution, the model unit includes a target subunit, a constraint subunit and a model subunit;

[0068] The target subunit is used to establish a first objective function based on the total number of units in each provincial power grid and the output of the units at different time periods, with the goal of minimizing the unit output cost;

[0069] The constraint subunit is used to establish a first constraint condition set based on load balance between nodes, operation restrictions of the units and line transmission capacity;

[0070] The model subunit is used to establish the provincial optimal unit output model based on the global virtual network admittance matrix, according to the first objective function and the first constraint condition.

[0071] As a preferred solution, the power module includes a construction unit, an output unit, a training unit and a judgment unit;

[0072] The construction unit is used to construct an optimal tie line power model based on the tie line node output power of each provincial power grid with the goal of minimizing the tie line node output price;

[0073] The output unit is used to output the optimal output power of each tie line according to the optimal tie line power model;

[0074] The training unit is used to update the optimal output power of the optimal tie-line power model by training a preset price response function;

[0075] The judgment unit is used to use the updated optimal output power as the optimal tie-line output power if the adjustment amount of the power injected by all tie-line nodes reaches a preset convergence condition.

[0076] As a preferred solution, the construction unit includes an establishment subunit, a condition subunit and an output subunit;

[0077] The establishing subunit is used to establish a second objective function based on the number of tie line nodes of each provincial power grid and the corresponding relationship between the change in the output power of the tie line nodes and the price, with the goal of minimizing the output price of the tie line nodes;

[0078] The condition subunit is used to establish a second constraint condition set according to the transmission capacity of the tie line and the node output power;

[0079] The output subunit is used to establish the optimal interconnection line power model based on the power generation power in each time period and the equivalent injection power of the interconnection line node obtained by optimizing the provincial optimal unit output, according to the second objective function and the second constraint condition.

[0080] As a preferred solution, the price response function is specifically:

[0081] According to the initial power of the tie lines of the provincial power grids, the provincial optimal unit output model is optimized to obtain a model optimization result;

[0082] According to the optimization results of the model, the marginal price of the tie line power output and the equivalent injection power of the tie line node at different time periods and different power generation load levels are calculated to obtain the calculation results;

[0083] The power generation state of the current unit is corrected based on the calculation result, and the price response function of the unit after correction is obtained.

[0084] As a preferred solution, the curve module includes an extraction unit, a removal unit, a clustering unit and a curve unit;

[0085] The extraction unit is used to extract features from a number of power quantity and price curves of the provincial power grids to obtain a data point set;

[0086] The elimination unit is configured to quantify the difference between each quantity-price curve and the reference curve by Euclidean distance based on the data point set, eliminate the power quantity-price curves that are continuously higher than the reference curve in preset key features, and obtain the remaining first power quantity-price curve set;

[0087] The clustering unit is used to divide the plurality of power quantity-price curves by clustering to obtain a clustering result, and eliminate the quantity-price curves that do not meet the preset conditions according to the clustering result to obtain the remaining second power quantity-price curve set;

[0088] The curve unit is used to establish an equivalent quantity-price curve of the virtual node according to the first power quantity-price curve set and the second power quantity-price curve set.

[0089] As a preferred solution, the clustering unit includes a discrete subunit, a processing subunit, an updating subunit and an iterative subunit;

[0090] The discrete subunit is used to discretize the plurality of power quantity-price curves to obtain a first data point set;

[0091] The processing subunit is used to perform standardization and clustering processing on the first data point set based on a preset number of clusters to be divided, so as to obtain a second data point set;

[0092] The updating subunit is used to update the corresponding cluster centroids in the second data point set according to the means of all points in the plurality of clusters;

[0093] The iterative subunit is used to take the power quantity price curves corresponding to the plurality of clusters as the clustering result if the centroid change is less than a preset threshold or reaches a preset number of iterations.

[0094] As a preferred solution, the cost module includes a simplification unit, a power supply unit and a cost unit;

[0095] The simplification unit is used to obtain simplified power data of each node according to the first power quantity-price curve set and the second power quantity-price curve set;

[0096] The power supply unit is used to introduce the optimal tie line output power into the provincial optimal unit output model, and use the simplified power supply data of each node as the power supplier to calculate the power supply allocated to different units;

[0097] The cost unit is used to calculate the average grid-connection cost of the virtual node based on the power sources allocated to the different units and the grid-connection cost of the transmission cost.

[0098] The present application also provides a storage medium having a computer program stored thereon. The computer program is called and executed by a computer to implement the above-mentioned method for optimizing the clearing of an out-of-province power equivalent market. BRIEF DESCRIPTION OF THE DRAWINGS

[0099] FIG1 is a flow chart of an out-of-province power supply equivalent market clearing optimization method provided in an embodiment of the present application;

[0100] 2 is a flow chart of an optimization method for clearing an out-of-province power supply equivalent market based on a price function response provided in an embodiment of the present application;

[0101] FIG3 is a schematic diagram of the structure of an out-of-province power source equivalent market clearing optimization device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0102] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0103] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "several" means two or more.

[0104] An out-of-province power source equivalent market clearing optimization method provided in an embodiment of the present application is mainly used in situations where it is necessary to consider the influence of multiple factors in optimizing power resource allocation and improve the optimized scheduling of the power system and the efficient operation of market coupling.

[0105] Embodiment 1:

[0106] Please refer to FIG1 . The embodiment of the present application provides an optimization method for clearing the out-of-province power supply equivalent market, including S1 to S6. The specific implementation steps are as follows:

[0107] S1. Obtain network data of each provincial power grid.

[0108] Step S1 of the embodiment of the present application is specifically as follows:

[0109] Obtain network data of provincial power grids.

[0110] S2. Establish the provincial optimal unit output model based on network data.

[0111] Step S2 of the embodiment of the present application includes S2.1 to S2.4; wherein S2.1 is the process of calculating the equivalent admittance matrix of the provincial power grid after being equivalent, S2.2 is the process of calculating the global virtual network admittance matrix, S2.3 is the process of establishing the provincial optimal unit output model, and S2.4 is the process of establishing the price response function, specifically:

[0112] S2.1. Use the WARD method to equalize the provincial power grid networks, take the tie line nodes as the reserved nodes, eliminate the internal nodes other than the tie line nodes, so that the nodes and branches in each provincial power grid are equalized to a single node, and obtain the equivalent admittance matrix of the provincial power grid after equalization and the equivalent injection power on the tie line nodes;

[0113] Specifically:

[0114] 1) N represents the set of external nodes (tie line nodes) in the provincial network, and I represents the set of internal nodes.

[0115] 2) The local initial admittance matrix corresponding to the retained node is , the local admittance matrix corresponding to the node to be eliminated is , the initial admittance matrix between internal nodes and external nodes is ; Among them, the retained nodes represent the internal nodes retained for the computational analysis, and the nodes to be eliminated represent the external nodes;

[0116] Among them, the equations of the power system to be analyzed before WARD equivalence are as follows:

[0117]

[0118] 3) Use internal node parameters and variables to represent external node voltage variables to eliminate internal nodes, thereby obtaining a linear equation system after WARD equivalent;

[0119] Among them, the external node voltage variables are represented by internal node parameters and variables, which can be expressed as:

[0120]

[0121] The linear equations after WARD equivalence are:

[0122]

[0123] For parameters and have:

[0124] ;

[0125]

[0126] in, and They represent the equalized node admittance matrix and injected current vector respectively.

[0127] 4) The power constraint is processed in the same way as above to obtain the linear equations after WARD is equalized;

[0128] Among them, based on the DC power flow model, the equation group of the power system to be analyzed before WARD equivalent can be expressed as follows:

[0129] The equations obtained after WARD is equal to:

[0130]

[0131] For parameters and have:

[0132] ;

[0133]

[0134] in, represents the corresponding susceptance matrix, represents the node phase angle vector, represents the node injection power vector, and They represent the node susceptance matrix and injected active power vector after equalization, , and They represent the local block admittance matrices corresponding to the equalization, Represents the injected active power of the internal nodes before equalization.

[0135] This embodiment S2.1 greatly reduces the number of state variables and model dimensions of the system by simplifying the complex provincial power grid into a single node, thereby significantly reducing the computational complexity and time cost. The equivalent global virtual network admittance matrix obtained by splicing not only includes the electrical characteristics within each province, but also reflects the inter-provincial electrical connections through inter-provincial contact lines, which enables the model to flexibly handle inter-provincial power flows and interactions, providing strong support for the formulation of a more comprehensive power grid dispatching strategy;

[0136] In addition, WARD equivalence is a method used to simplify network analysis in power systems, especially when performing power flow calculations and optimization of large-scale power grids. By equating some nodes and branches in the power grid to a single node, the complexity of calculations can be reduced. Therefore, this embodiment uses WARD equivalence to simplify the power grid model while keeping the key characteristics of the power grid unchanged.

[0137] S2.2. Combine the equivalent admittance matrix and the inter-provincial contact line node admittance matrix according to the unified contact line node number to generate an equivalent global virtual network admittance matrix ;

[0138] Among them, assuming that there are a total of The tie line node admittance matrices after equalization are , then the global virtual network admittance matrix is ​​expressed as:

[0139]

[0140] Where diag(●) means constructing a block diagonal matrix; It means that only the tie line nodes of each provincial power grid are retained, and only the partial admittance matrix contributed by the inter-provincial tie lines is considered.

[0141] The global virtual network admittance matrix generated by S2.2 of this embodiment includes impedance parameters of each provincial network and tie line, which is helpful for accurately calculating the tie line power flow.

[0142] S2.3. Based on the total number of units in each provincial power grid and the output of units at different times, the first objective function is established with the goal of minimizing the unit output cost. ;

[0143] Based on the load balance between nodes, the operation restrictions of the units and the line transmission capacity, a first set of constraints is established; wherein the first set of constraints includes node load balance, upper and lower limits of unit output, unit climbing capability and line transmission capacity constraints;

[0144] Based on the global virtual network admittance matrix, according to the first objective function and the first constraint condition, establish the provincial optimal unit output model;

[0145] Among them, the first objective function is:

[0146]

[0147] The node load balance is:

[0148]

[0149] The upper and lower limits of the unit output are:

[0150]

[0151] The unit climbing ability is:

[0152]

[0153] The line transmission capacity constraint is:

[0154]

[0155] Among them, N represents the total number of units, including non-market units and market units, excluding independent energy storage trading units; T represents the total number of time periods considered, where D day has a time period of 15 minutes, and 96 time periods are considered. D+1 day considers two time periods of load peak and trough, so T is 98; represents the output of unit i in period t; represents the operating cost of unit i in time period t, which is a multi-segment linear function related to the output intervals reported by the unit and the corresponding energy price; M represents the network power flow constraint relaxation penalty factor used for market clearing optimization, represents the penalty factor for the relaxation of the curtailment constraint of renewable energy; They represent the forward and reverse power flow relaxation variables of line l respectively; NL represents the total number of lines; They represent the forward and reverse flow relaxation variables of section s respectively; NS represents the total number of sections; represents the amount of abandoned electricity in the new energy trading unit c; ES represents the number of independent energy storage trading units; Indicates the charging price and discharging price declared by independent energy storage; They represent the charging and discharging power of independent energy storage respectively;

[0156] are the node voltage phase angle, unit output and the components of node load in time period t respectively; It represents the DC power flow admittance matrix established by the inverse of the line reactance value;

[0157] are the components of the maximum and minimum output of the unit in time period t respectively;

[0158] and are the maximum positive and negative climbing capabilities of unit i respectively; hour, is the output of unit i in the last period of the previous decision cycle;

[0159] represents the maximum transmission capacity of line l; for Elements of represents the line-node association matrix; Represents the inverse of the diagonal matrix of line reactance values; Represents the phase angle vector of the internal node in time period t.

[0160] In this embodiment S2.3, minimizing the unit output cost is taken as the goal, which can ensure that while meeting the electricity demand, the power generation cost is reduced as much as possible and the economic benefits of the power company are improved; and the provincial optimal unit output model takes into account a variety of constraints, including load balance between nodes, unit operation restrictions and line transmission capacity restrictions. These constraints ensure the feasibility and safety of the optimization results and avoid problems in the actual operation of the scheduling plan caused by ignoring certain constraints; in addition, since the provincial optimal unit output model is established based on the global virtual network admittance matrix, this matrix not only includes the electrical characteristics within the provincial power grid, but also reflects the cross-provincial electrical connections through inter-provincial interconnection lines, which enables the model to more accurately reflect the actual operation of the power grid and improve the accuracy of the optimization model.

[0161] S2.4, based on the initial power of the tie lines of each provincial power grid, the initial tie line power is set to 0MW, the provincial optimal unit output model is optimized, and the model optimization result is obtained;

[0162] According to the model optimization results, the marginal price of tie-line power output and the equivalent injected power of tie-line nodes at different time periods and different generation load levels are calculated to obtain the calculation results;

[0163] Based on the calculation results, the power generation status of the current unit is corrected to obtain the price response function of the unit after correction. (Price response function curve).

[0164] This embodiment S2.4 provides detailed information about power trading and market price changes by calculating the marginal price of tie-line power output and the equivalent injected power of tie-line nodes at different time periods and different power generation load levels. The calculation of marginal price helps to understand the sensitivity of prices in the power market to changes in supply and demand, while the calculation of equivalent injected power helps to evaluate the power flow of each node in the power grid. Based on the calculation results, the power generation status of the current unit is corrected to improve the price response reliability of the price response function.

[0165] S3. Optimize the interconnection line power function of each provincial power grid to obtain the optimal interconnection line output power.

[0166] Step S3 of the embodiment of the present application includes S3.1 to S3.2; specifically, S3.1 is a process of establishing an optimal tie line power model, and S3.2 is a process of calculating the optimal tie line output power, specifically:

[0167] S3.1. Based on the number of tie-line nodes in each provincial power grid, combined with the corresponding relationship between the change in tie-line node output power and price, the second objective function is established with the goal of minimizing the tie-line node output price. ;

[0168] According to the transmission capacity of the tie line and the node output power, a second set of constraint conditions is established; wherein the second set of constraint conditions includes the tie line node power balance, the tie line transmission capacity constraint, the tie line node output power expression and the piecewise linear constraint of the price response function;

[0169] Based on the provincial optimal unit output optimization, the power generation power in each period and the equivalent injection power of the tie line node are obtained. According to the second objective function and the second constraint condition, establish the optimal tie line power model;

[0170] The construction method of the piecewise linear constraint of the price response function is as follows: Assume that the output power-cost increment curve of the tie line node b is The slope of each segment in each time period is calculated using the power generation load-cost curve. and The intercept, from which the piecewise linear power output cost of the tie line node b in time period t can be obtained The constraint expression form;

[0171] Among them, the second objective function is:

[0172]

[0173] The power balance of the tie line node is:

[0174]

[0175] The transmission capacity constraint of the tie line is:

[0176]

[0177] The output power expression of the tie line node is:

[0178]

[0179] The piecewise linear constraints on the price response function are:

[0180]

[0181] For parameters have:

[0182]

[0183] in, It is expressed as the number of tie line nodes; It represents the power output cost of tie line node b in time period t, that is, the price response function corresponding to the change of tie line node output power;

[0184] and They are respectively represented as the tie line node voltage phase angle, tie line node equivalent injection and tie line node injection power regulation in time period t;

[0185] represents the maximum transmission capacity of the tie line l; Expressed as Elements of represents the tie line-node association matrix; Represents the inverse of the diagonal matrix formed by the reactance values ​​of the tie lines; Represents the phase angle vector of the tie line node in time period t;

[0186] It is expressed as the component of the tie line node output power in time period t; represents the node-node association matrix;

[0187] is the output power of tie line node b in time period t; It represents the basic power of the tie line node b in time period t under the equivalent injection power under the tie.

[0188] In this embodiment S3.1, by combining the corresponding relationship between the change in the output power of the interconnection line node and the price, the second objective function can more accurately reflect the changes in the price and supply and demand relationship in the electricity market. And the second constraint condition includes the transmission capacity of the interconnection line and the limitation of the node output power. These comprehensive constraints ensure the feasibility and safety of the optimization model. In addition, the provincial optimal unit output model fully considers the results of the provincial optimal unit output optimization during the establishment process, that is, the power generation power in each time period and the equivalent injection power of the interconnection line node. This further model establishment method based on the existing optimization results helps to achieve a higher level of coordination and optimization, and improve the operating efficiency and economy of the entire power grid.

[0189] S3.2. Output the optimal output power of each tie line according to the optimal tie line power model;

[0190] Through the price response function Conduct training to update the optimal output power of the optimal tie line power model;

[0191] If the adjustment amount of the power injected into all tie line nodes is The absolute value of is less than a given threshold , then convergence is determined, and the updated optimal output power is used as the optimal tie line output power; if the convergence condition is not met, the optimal output power is repeatedly updated until the convergence condition is met; where, is a small positive number;

[0192] Among them, taking the tie line node b in the provincial optimal unit output model as an example, the tie line output power of this node is The training process of the price response function curve is:

[0193] 1) In the Kth iteration Optimize and obtain the optimal unit output plan for each time period and the node marginal electricity price of node b.

[0194] 2) Calculate the sum of the unit output in each period, and pair the total power generation load in each period with the marginal electricity price of node b in each period to obtain Bigram ;in, The number of optimization periods.

[0195] 3) The obtained The binary groups are added to the binary group sequence in the historical iteration process, and the system generation load-cost segment characteristic curve is drawn according to this sequence, and the slope of each segment of the curve is the corresponding value.

[0196] 4) Given the current state of power generation load By translating the origin of the system power generation load-cost segmented characteristic curve to the position of the given current state power point load, we can get the relationship between the output power increment and the cost increment of node b in the current state, that is, we can get the output power-cost increment curve of node b, which represents the change of power generation cost of the provincial power system with the output power of the interconnection line node.

[0197] This embodiment S3.2 aims to minimize the output price of the tie line node and constructs an optimal tie line power model, which helps to achieve the optimal configuration of power resources while ensuring the safety of power grid operation, thereby saving costs. The optimal tie line power model is updated by training the price response function. This method enables the model to be dynamically adjusted according to price changes in the actual electricity market, ensuring that the model can still maintain its optimization performance in the face of market price fluctuations. By introducing a convergence condition to determine whether the adjustment amount of the power injected into the tie line node has reached a stable state, this ensures the stability and reliability of the model optimization process and avoids infinite loops or divergences.

[0198] S4. By simplifying and equivalence processing the power supply outside the province, an equivalent quantity-price curve of the virtual node is established.

[0199] Step S4 of the embodiment of the present application includes S4.1 to S4.4; wherein S4.1 is a process of establishing a reference curve, S4.2 is a process of eliminating data to establish a first power quantity-price curve set, S4.3 is a process of aggregating data using a K-Means algorithm, and S4.4 is a process of eliminating data again and establishing an equivalent quantity-price curve for a virtual node, specifically:

[0200] S4.1. First, collect historical clearing data including clearing prices of different power sources at various transaction volume levels;

[0201] Based on historical clearing data, the polynomial coefficients are estimated using the Least Squares Method (LSM) , that is, for each data point , the computational model predicts The error between the actual observation ,have:

[0202]

[0203] Secondly, for polynomial fitting, a design matrix X is constructed, and the columns of the design matrix X are composed of the powers of the independent variables x;

[0204] Use the design matrix and the observation vector y to construct the normal equations to solve for the vector of model parameters ;in, The matrix representing the polynomial coefficients, the specific mathematical formula is:

[0205]

[0206] like is reversible, then the parameter vector is calculated by matrix inversion ,Right now ; According to the obtained parameters and the polynomial coefficients Calculate the reference curve ,as follows:

[0207]

[0208] Where N represents the number of samples in the data set. represents the i-th observation value, Representing polynomials The predicted value at; and, the reference curve Used to reflect the general situation of power supply being cleared historically.

[0209] It is represented as the transpose of X;

[0210] x is the independent variable, usually the trading volume is used as the independent variable; represents the coefficients of the polynomial, and n is the order of the polynomial.

[0211] S4.2. Discretize the quantity-price curves of several power sources at the same trading volume level to obtain a series of data point sets; where the data point sets represent prices at different trading volumes, and there is The corresponding prices are ;

[0212] For each volume-price curve, calculate its difference with the reference curve Euclidean distance D at each volume level;

[0213] Quantify the difference between each quantity-price curve and the reference curve in several power quantity-price curves by using the Euclidean distance D The difference between them; the smaller the Euclidean distance D is, the closer the volume-price curve is to the reference curve. The closer they are, the smaller the difference is.

[0214] Eliminate the power quantity-price curves that are continuously higher than the reference curve in terms of preset key features according to a preset threshold value of the Euclidean distance D, and obtain the remaining first power quantity-price curve set;

[0215] Among them, the Euclidean distance D is expressed as:

[0216]

[0217] in, Indicates that the volume price curve is in the trading volume The corresponding price below, represents the price of the reference curve under the same trading volume; n is the total number of trading volume levels, that is, the number of discrete points of the volume-price curve.

[0218] S4.3, discretizing a number of power quantity-price curves to obtain cost values ​​at different processing levels to form a first data point set;

[0219] Based on the K-Means algorithm, the first data point set is standardized so that the mean of each feature is 0 and the standard deviation is 1;

[0220] Based on the preset number of clusters k to be divided, select k initialized centroids in the first data point set , expressed as ,in is the centroid of the kth cluster, or the K-means++ algorithm can be used to initialize the centroid to obtain the second set of data points; where k is obtained based on prior experience or through cross-validation;

[0221] Set the maximum number of iterations N, the control algorithm will perform up to N iterations to update the centroid and cluster division;

[0222] For each iteration n (from 1 to N): the clustering needs to be initialized is zero or an empty set, where t represents the number of iterations, ;

[0223] And, for each sample curve in the second data point set , the following calculations are required:

[0224] 1) Calculate the sample curve With all centroid vectors The distance between . Usually calculated using Euclidean distance;

[0225] 2) Find the centroid with the smallest distance ,Right now Yes all the smallest of the

[0226] 3) The sample curve Assign to The corresponding cluster Update

[0227] In the second data point set, update the corresponding cluster centroid according to the mean or median of all points in several clusters ;

[0228] Among them, the K-Means++ algorithm is an optimization of the K-Means random initialization centroid method, and its optimization strategy can be expressed as follows:

[0229] 1) Randomly select a point from the input data point set as the first cluster center ;

[0230] 2) For each point in the data set , calculate its distance to the nearest cluster center among the selected cluster centers ;

[0231] 3) Select a new data point as the new cluster center. The selection principle is: Larger points have a greater probability of being cluster centers;

[0232] 4) Repeat steps 2 and 3 above until k cluster centroids are selected;

[0233] 5) Use these k centroids as the initial centroids to run the standard K-Means algorithm.

[0234] S4.4. If the centroid change is less than the preset threshold or reaches the preset number of iterations N, the final cluster division is output , the power quantity-price curves corresponding to several clusters are taken as clustering results, and the quantity-price curves that are on the edge or have large differences with the main cluster are eliminated according to the clustering results, and the remaining second power quantity-price curve set after inverse standardization screening is obtained; wherein, the cluster division Each of Contains the allocation All sample points of the centroid;

[0235] Extracting key features, such as marginal cost, from power data of each node according to the first power quantity price curve set and the second power quantity price curve set;

[0236] Based on the three specific characteristics of power generation, cost and unit reliability among the key characteristics, a multi-objective optimization method is used to calculate the weight of each node power supply under different characteristics to represent the equivalent characteristics of the virtual node, and the quantity-price curve of each node is fitted to obtain the fitting curve;

[0237] The equivalent quantity-price curve of the virtual node is calculated based on the fitting curve;

[0238] where, for each treatment level q, the cost of the isoquantity-price curve is It can be expressed as:

[0239]

[0240] in, represents the weight of the i-th node, It represents the operating cost of the ith node at the output level q.

[0241] Among them, the multi-objective optimization method is:

[0242] 1) Establish the power generation cost objective function and environmental cost / carbon emission cost objective function ,Right now:

[0243] ;

[0244]

[0245] For parameters and have:

[0246] ;

[0247]

[0248] in, represents the carbon emission cost of the ith node at time t, represents the carbon emissions of the power system, represents the weight of the i-th node, and n represents the total number of node power supplies;

[0249] , as well as represents the fuel cost coefficient of thermal power unit i; It represents the carbon emission intensity of thermal power unit i, which varies with the load rate of the unit. When the thermal power is in the deep peak regulation DPR state, its carbon emission intensity is significantly increased compared with the conventional output; Represents the time interval from t-1 to t.

[0250] 2) Establish decision variables is used to represent the weight of the i-th node, and Satisfy the normalization condition .

[0251] 3) Use multi-objective optimization algorithm - multi-objective particle swarm optimization PSO algorithm for calculation:

[0252] A. Iterative formula for particle velocity and position:

[0253] Suppose there are m particles in D-dimensional space, and the position of the i-th particle is a vector ; The velocity of the i-th particle is a vector ; The optimal position searched by the i-th particle is ; The optimal position searched by the entire particle swarm is ; Particle speed and location The iteration formula is as follows:

[0254] ;

[0255]

[0256] in, ; is the inertia parameter; , is the learning factor; , is a random number; the three items on the right side of the particle speed iteration formula are the historical speed memory part, the cognitive part and the social part.

[0257] In order to improve the global optimization ability of the initial particle swarm and enable the particle swarm to jump out of the local optimal solution, the specific update rules of the parameters are as follows:

[0258] ;

[0259]

[0260] in, The values ​​of are 0.9 and 0.7 respectively; The values ​​of are 1.5 and 1 respectively.

[0261] B. Adaptive particle optimization strategy:

[0262] During the PSO algorithm solution process, the history of the search particles and the global optimal particle will be continuously updated and guide the remaining search particles to fly to the optimal position, thereby achieving the effect of algorithm convergence. However, the rapid aggregation of search particles will lead to more invalid solutions in close distances and easily cause the algorithm to fall into the local optimum. Therefore, the adaptive operator is introduced to set the particle optimization strategy as follows:

[0263]

[0264] When the particle meets the following position update conditions:

[0265]

[0266] Perform position transformation optimization according to the following method:

[0267]

[0268] Among them, e represents the adaptive operator, Represented as a random number function, Denoted as adaptive threshold, is the Euclidean distance between particle i and the nearest particle j in the D-dimensional space formed by the objective function, It is expressed as the Euclidean distance between particle i and the optimal particle in the population; Q is the optimization judgment threshold.

[0269] Moreover, the initial particles are far away from the optimal particles at the beginning of the algorithm. The particle position update strategy is mainly affected by the adaptive operator term, and the particle diversity is enhanced. In the later stage of the algorithm, the particles gradually approach the global optimal solution. At this time, the adaptive operator term has little effect on the particle position update, and the particles mainly rely on their own positions for optimization, ensuring the algorithm's ability to solve problems in a small range.

[0270] C. Fuzzy strategy:

[0271] The fuzzy decision-making method is used to fuzzify the two objective functions of operating cost and environmental cost. The processing functions are as follows:

[0272]

[0273] in, is the single objective function value of the i-th Pareto optimal solution, is the objective function value after fuzzy processing, and are the maximum and minimum values ​​of the corresponding objective function.

[0274] Then, [operating cost, environmental cost] is taken as a factor set, and a single factor evaluation is performed on the non-inferior solution set obtained by the PSO algorithm to form a fuzzy evaluation matrix. For ease of analysis, the weight vector of the evaluation factor in this embodiment is [0.5, 0.5];

[0275] Finally, a fuzzy comprehensive evaluation is carried out and the best one is selected as one of the components of the final scheduling plan.

[0276] In this embodiment, S4.3~S4.4 discretizes the power quantity price curve into a data point set, which can significantly reduce the amount of data, reduce the computational complexity of subsequent processing, and improve processing efficiency. The data point set is clustered based on the preset number of clusters, and the degree of clustering sophistication can be adjusted according to actual needs. The clustering results are continuously optimized by iteratively updating the cluster centroid until the centroid change is less than the preset threshold or reaches the preset number of iterations. This iterative optimization process helps to reduce clustering errors and improve the accuracy and stability of clustering results;

[0277] In addition, the use of Euclidean distance to quantify the difference between the power quantity and price curve and the reference curve can accurately measure the similarity or difference between the curves, providing a reliable basis for the subsequent elimination of abnormal curves. By eliminating power quantity and price curves that are continuously higher than the reference curve in preset key features, those abnormal or unreasonable data points can be excluded, thereby avoiding their adverse effects on subsequent analysis. The use of clustering to divide the power quantity and price curves can not only further identify groups of curves with similar characteristics, but also help to discover potential power market structures or patterns, and by eliminating quantity and price curves that do not meet the preset conditions, the quality of the data can be further improved.

[0278] S5. Introduce the optimal interconnection line output power into the provincial optimal unit output model for calculation to obtain the average network access cost of the virtual node.

[0279] Step S5 of the embodiment of the present application is specifically as follows:

[0280] Acquire simplified power data of each node according to the first power quantity-price curve set and the second power quantity-price curve set;

[0281] The optimal tie line output power is introduced into the provincial optimal unit output model, and the simplified power supply data of each node is used as the power supplier to calculate the power supply allocated to different units;

[0282] Based on the power sources allocated to different units, the average grid-connection cost of the virtual node is calculated according to the grid-connection cost of the transmission cost.

[0283] In this embodiment S5, the simplified power supply data of each node retains key information and reduces data redundancy, so the complexity and calculation amount of data processing can be effectively reduced. By using the simplified power supply data of each node as the power supplier, the power source allocated to different units is calculated. This process reflects the balance of supply and demand in the power market and the optimal allocation of power resources, which can reduce operating costs and transmission losses as much as possible.

[0284] S6. The unit output arrangement is calculated based on the optimal interconnection line output power, equivalent quantity-price curve and average grid access fee, and the output of each provincial power grid is configured according to the unit output arrangement.

[0285] Step S6 of the embodiment of the present application includes S6.1 to S6.2; wherein S6.1 is the process of establishing a regional market clearing function, and S6.2 is the process of configuring the output of each provincial power grid, specifically:

[0286] S6.1. By treating the power supply nodes of other provinces in each provincial power grid as a virtual node, a regional market clearing function is established as follows:

[0287]

[0288] For parameters have:

[0289]

[0290] Among them, N represents the total number of generator sets; T represents the total number of time periods considered, where one time period is 15 minutes on D day, and 96 time periods are considered. On D+1 day, two time periods, peak and valley load, are considered, so T is 98; represents the output of unit i in period t; unit operating cost It is a multi-segment linear function related to each output range declared by the unit and the corresponding energy price; Indicates the average price of Internet access fee; Indicates the power transmission power of the inter-provincial transmission channel; Represents the penalty factor for power abandonment of new energy trading units; represents the abandoned power of the new energy trading unit r in time period t;

[0291] It represents the sum of inter-provincial transmission fees (including network loss fees), transmission and distribution fees within the province on the sending side, and transmission fee sharing space. represents the transmission power of inter-provincial power transmission component i in time period t.

[0292] S6.2. Based on the output power of the optimal interconnection line, the equivalent quantity-price curve of the virtual node, the average transmission fee of the virtual node, the operating cost of the power supply within the province, and the regional market clearing function, the unit output arrangement is calculated, and the output of each provincial power grid is configured according to the unit output arrangement.

[0293] To apply the first embodiment of the present application, please refer to Figure 2, which is a flow chart of the out-of-province power equivalent market clearing optimization method based on price function response provided in the embodiment of the present application, representing the general process of calculating various parameters and performing corresponding processing from the beginning to the end of this embodiment.

[0294] Overall, this embodiment has the following beneficial effects:

[0295] The present invention establishes a provincial optimal unit output model by collecting and analyzing network data, which fully considers the actual situation and needs within each provincial power grid. By simplifying and equivalence processing of out-of-province power sources, an equivalent quantity-price curve of a virtual node is established, which simplifies the complex inter-provincial power trading problem and improves the accuracy and efficiency of the calculation. Introducing the optimal interconnection line output power into the provincial optimal unit output model for calculation can accurately predict and evaluate the impact of different scheduling schemes on the power system, providing strong data support for making scientific scheduling decisions. For unit output arrangement, the introduction of the optimal interconnection line output power ensures the coordination and balance of power transmission and exchange between different provincial power grids; considering the economic factor of network access fees, the unit output arrangement is more in line with market laws, which helps to reduce the cost of power transmission and trading; by comparing the equivalent quantity-price curves of different virtual nodes, it is possible to identify which power sources are more competitive under current market conditions, thereby giving priority to the output of these power sources and achieving optimal allocation of power resources;

[0296] In addition, the provincial network equivalence adopts the Ward equivalence method, with the tie line node as the reserved node, eliminating the internal nodes other than the tie line node, so as to obtain the equivalent admittance matrix of the tie line node and the equivalent injected power on the node, so as to accurately consider the influence of the provincial network parameters on the global power flow distribution; the provincial optimal unit output model adopts the minimization of the unit output cost as the optimization function, considering the relevant constraints to obtain the tie line equivalent node power and marginal price under the given tie line power; the regional optimal tie line model adopts the minimization of the tie line node output price, considering the tie line node power balance and other related constraints to obtain the tie line output power, and cyclically generates / updates the given tie line equivalent node power output price function until convergence; the virtual node equivalence method of the out-of-province market is adopted, and the out-of-province power source is equivalent to the same node to participate in the regional market clearing, considering the weight calculation of the multi-objective optimization function and the clearing function of the regional market, so as to obtain the optimization function of efficient clearing. Therefore, this embodiment has a more flexible, comprehensive and accurate modeling capability when processing the power data of multiple provinces with complex coupling such as the power market, which can provide more powerful support for the efficient operation and decision-making of the power market.

[0297] Embodiment 2:

[0298] Please refer to FIG3 , an embodiment of the present application provides an out-of-province power supply equivalent market clearing optimization device, including a data module 10 , a model module 20 , a power module 30 , a curve module 40 , a cost module 50 and a scheduling module 60 ;

[0299] Wherein, the data module 10 is used to obtain network data of each provincial power grid;

[0300] Model module 20, used to establish a provincial optimal unit output model based on network data;

[0301] The power module 30 is used to optimize the tie line power function of each provincial power grid to obtain the optimal tie line output power;

[0302] The curve module 40 is used to establish an equivalent quantity-price curve of the virtual node by simplifying and equivalent processing the power supply outside the province;

[0303] Cost module 50, used to introduce the optimal tie line output power into the provincial optimal unit output model for calculation, and obtain the average network cost of the virtual node;

[0304] The dispatching module 60 is used to calculate the unit output arrangement according to the optimal interconnection line output power, the equivalent quantity-price curve and the average grid access fee, and configure the output of each provincial power grid according to the unit output arrangement.

[0305] In one embodiment, the data module 10 is specifically:

[0306] Obtain network data of provincial power grids.

[0307] In one embodiment, the model module 20 includes an equivalent unit, a splicing unit, a target subunit, a constraint subunit, a model subunit and a function unit; wherein the equivalent unit is a process of calculating the equivalent admittance matrix of the provincial power grid after the equivalent value, the splicing unit is a process of calculating the global virtual network admittance matrix, the target subunit, the constraint subunit and the model subunit are the processes of establishing the provincial optimal unit output model, and the function unit is the process of establishing the price response function, specifically:

[0308] The equivalent unit is used to use the WARD method to equalize the provincial power grid networks, taking the tie line nodes as the reserved nodes and eliminating the internal nodes other than the tie line nodes, so as to equalize the nodes and branches in each provincial power grid to a single node, and obtain the equivalent admittance matrix of the provincial power grid after equalization and the equivalent injection power on the tie line nodes;

[0309] Specifically:

[0310] 1) N represents the set of external nodes (tie line nodes) in the provincial network, and I represents the set of internal nodes.

[0311] 2) The local initial admittance matrix corresponding to the retained node is , the local admittance matrix corresponding to the node to be eliminated is , the initial admittance matrix between internal nodes and external nodes is ; Among them, the retained nodes represent the internal nodes retained for the computational analysis, and the nodes to be eliminated represent the external nodes;

[0312] Among them, the equations of the power system to be analyzed before WARD equivalence are as follows:

[0313]

[0314] 3) Use internal node parameters and variables to represent external node voltage variables to eliminate internal nodes, thereby obtaining a linear equation system after WARD equivalent;

[0315] Among them, the external node voltage variables are represented by internal node parameters and variables, which can be expressed as:

[0316]

[0317] The linear equations after WARD equivalence are:

[0318]

[0319] For parameters and have:

[0320] ;

[0321]

[0322] in, and They represent the equalized node admittance matrix and injected current vector respectively.

[0323] 4) The power constraint is processed in the same way as above to obtain the linear equations after WARD is equalized;

[0324] Among them, based on the DC power flow model, the equation group of the power system to be analyzed before WARD equivalent can be expressed as follows:

[0325] The equations obtained after WARD is equal to:

[0326]

[0327] For parameters and have:

[0328] ;

[0329]

[0330] in, represents the corresponding susceptance matrix, represents the node phase angle vector, represents the node injection power vector, and They represent the node susceptance matrix and injected active power vector after equalization, , and They represent the local block admittance matrices corresponding to the equalization, Represents the injected active power of the internal nodes before equalization.

[0331] The equivalent unit of this embodiment greatly reduces the number of state variables and model dimensions of the system by simplifying the complex provincial power grid into a single node, thereby significantly reducing the computational complexity and time cost. The equivalent global virtual network admittance matrix obtained by splicing not only includes the electrical characteristics within each province, but also reflects the inter-provincial electrical connections through inter-provincial contact lines, which enables the model to flexibly handle inter-provincial power flows and interactions, providing strong support for the formulation of a more comprehensive power grid dispatching strategy;

[0332] In addition, WARD equivalence is a method used to simplify network analysis in power systems, especially when performing power flow calculations and optimization of large-scale power grids. By equating some nodes and branches in the power grid to a single node, the complexity of calculations can be reduced. Therefore, this embodiment uses WARD equivalence to simplify the power grid model while keeping the key characteristics of the power grid unchanged.

[0333] The splicing unit is used to splice the equivalent admittance matrix and the inter-provincial tie line node admittance matrix according to the unified tie line node number to generate an equivalent global virtual network admittance matrix ;

[0334] Among them, assuming that there are a total of The tie line node admittance matrices after equalization are , then the global virtual network admittance matrix is ​​expressed as:

[0335]

[0336] Where diag(●) means constructing a block diagonal matrix; It means that only the tie line nodes of each provincial power grid are retained, and only the partial admittance matrix contributed by the inter-provincial tie lines is considered.

[0337] The global virtual network admittance matrix generated by the splicing unit in this embodiment includes impedance parameters of each provincial network and tie line, which is helpful for accurately calculating the tie line power flow.

[0338] The target subunit is used to establish the first objective function based on the total number of units in each provincial power grid and the output of the units at different times, with the goal of minimizing the unit output cost. ;

[0339] The constraint subunit is used to establish a first set of constraint conditions based on the load balance between nodes, the operation restrictions of the units and the line transmission capacity; wherein the first set of constraint conditions includes the node load balance, the upper and lower limits of the unit output, the unit climbing capability and the line transmission capacity constraints;

[0340] A model subunit is used for, based on the global virtual network admittance matrix, according to the first objective function and the first constraint condition, establish the provincial optimal unit output model;

[0341] Among them, the first objective function is:

[0342]

[0343] The node load balance is:

[0344]

[0345] The upper and lower limits of the unit output are:

[0346]

[0347] The unit climbing ability is:

[0348]

[0349] The line transmission capacity constraint is:

[0350]

[0351] Among them, N represents the total number of units, including non-market units and market units, excluding independent energy storage trading units; T represents the total number of time periods considered, where D day has a time period of 15 minutes, and 96 time periods are considered. D+1 day considers two time periods of load peak and trough, so T is 98; represents the output of unit i in period t; represents the operating cost of unit i in time period t, which is a multi-segment linear function related to the output intervals reported by the unit and the corresponding energy price; M represents the network power flow constraint relaxation penalty factor used for market clearing optimization, represents the penalty factor for the relaxation of the curtailment constraint of renewable energy; They represent the forward and reverse power flow relaxation variables of line l respectively; NL represents the total number of lines; They represent the forward and reverse flow relaxation variables of section s respectively; NS represents the total number of sections; represents the amount of abandoned electricity in the new energy trading unit c; ES represents the number of independent energy storage trading units; Indicates the charging price and discharging price declared by independent energy storage; They represent the charging and discharging power of independent energy storage respectively;

[0352] are the node voltage phase angle, unit output and the components of node load in time period t respectively; It represents the DC power flow admittance matrix established by the inverse of the line reactance value;

[0353] are the components of the maximum and minimum output of the unit in time period t respectively;

[0354] and are the maximum positive and negative climbing capabilities of unit i respectively; hour, is the output of unit i in the last period of the previous decision cycle;

[0355] represents the maximum transmission capacity of line l; for Elements of represents the line-node association matrix; Represents the inverse of the diagonal matrix of line reactance values; Represents the phase angle vector of the internal node in time period t.

[0356] The target subunit, constraint subunit and model subunit of this embodiment take minimizing the unit output cost as the goal, which can ensure that while meeting the power demand, the power generation cost is reduced as much as possible and the economic benefits of the power company are improved; and the provincial optimal unit output model takes into account a variety of constraints, including load balance between nodes, unit operation restrictions and line transmission capacity restrictions. These constraints ensure the feasibility and safety of the optimization results and avoid problems in the actual operation of the scheduling plan caused by ignoring certain constraints; in addition, since the provincial optimal unit output model is established based on the global virtual network admittance matrix, this matrix not only contains the electrical characteristics within the provincial power grid, but also reflects the cross-provincial electrical connections through inter-provincial interconnection lines, which enables the model to more accurately reflect the actual operation of the power grid and improve the accuracy of the optimization model.

[0357] The function unit is used to set the initial tie line power to 0MW based on the initial tie line power of each provincial power grid, optimize the provincial optimal unit output model, and obtain the model optimization result;

[0358] The function unit is also used to calculate the marginal price of tie-line power output and the equivalent injection power of tie-line nodes at different time periods and different power generation load levels according to the model optimization results, and obtain the calculation results;

[0359] The function unit is also used to correct the power generation status of the current unit based on the calculation results to obtain the price response function of the unit after correction. (Price response function curve).

[0360] The function unit of this embodiment provides detailed information about power trading and market price changes by calculating the marginal price of tie-line power output and the equivalent injected power of tie-line nodes at different time periods and different power generation load levels. The calculation of marginal price helps to understand the sensitivity of prices in the power market to changes in supply and demand, while the calculation of equivalent injected power helps to evaluate the power flow of each node in the power grid. The power generation status of the current unit is corrected based on the calculation results, which can improve the price response reliability of the price response function.

[0361] In one embodiment, the power module 30 includes an establishment subunit, a condition subunit, an output subunit, an output unit, a training unit and a judgment unit; wherein the establishment subunit, the condition subunit and the output subunit are the process of establishing the optimal tie line power model, and the output unit, the training unit and the judgment unit are the process of calculating the optimal tie line output power, specifically:

[0362] A subunit is established to establish the second objective function based on the number of tie line nodes of each provincial power grid, combined with the corresponding relationship between the change in the output power of the tie line nodes and the price, with the goal of minimizing the output price of the tie line nodes. ;

[0363] A condition subunit is used to establish a second set of constraint conditions according to the transmission capacity of the tie line and the node output power; wherein the second set of constraint conditions includes a tie line node power balance, a tie line transmission capacity constraint, a tie line node output power expression, and a piecewise linear constraint of a price response function;

[0364] The output subunit is used to optimize the power generation power of each period and the equivalent injection power of the tie line node based on the provincial optimal unit output optimization, according to the second objective function and the second constraint condition, establish the optimal tie line power model;

[0365] The construction method of the piecewise linear constraint of the price response function is as follows: Assume that the output power-cost increment curve of the tie line node b is The slope of each segment in each time period is calculated using the power generation load-cost curve. and The intercept, from which the piecewise linear power output cost of the tie line node b in time period t can be obtained The constraint expression form;

[0366] Among them, the second objective function is:

[0367]

[0368] The power balance of the tie line node is:

[0369]

[0370] The transmission capacity constraint of the tie line is:

[0371]

[0372] The output power expression of the tie line node is:

[0373]

[0374] The piecewise linear constraints on the price response function are:

[0375]

[0376] For parameters have:

[0377]

[0378] in, It is expressed as the number of tie line nodes; It represents the power output cost of tie line node b in time period t, that is, the price response function corresponding to the change of tie line node output power;

[0379] and They are respectively represented as the tie line node voltage phase angle, tie line node equivalent injection and tie line node injection power regulation in time period t;

[0380] represents the maximum transmission capacity of the tie line l, Expressed as Elements of represents the tie line-node association matrix, represents the inverse of the diagonal matrix of the tie line reactance values, Represents the phase angle vector of the tie line node in time period t;

[0381] It is expressed as the component of the tie line node output power in time period t, represents the node-node association matrix;

[0382] is the output power of tie line node b in time period t; It represents the basic power of the tie line node b in time period t under the equivalent injection power under the tie.

[0383] This embodiment establishes sub-units, conditional sub-units, and output sub-units by combining the corresponding relationship between the change in the output power of the interconnection line node and the price, so that the second objective function can more accurately reflect the changes in price and supply and demand in the electricity market. And the second constraint condition includes the transmission capacity of the interconnection line and the restrictions on the output power of the node. These comprehensive constraints ensure the feasibility and safety of the optimization model. In addition, the provincial optimal unit output model fully considers the results of the provincial optimal unit output optimization during the establishment process, that is, the power generation power in each time period and the equivalent injection power of the interconnection line node. This further model establishment method based on the existing optimization results helps to achieve a higher level of coordination and optimization, and improve the operating efficiency and economy of the entire power grid.

[0384] An output unit, used for outputting the optimal output power of each tie line according to the optimal tie line power model;

[0385] Training unit, used to respond to the price function Conduct training to update the optimal output power of the optimal tie line power model;

[0386] The judgment unit is used to adjust the amount of power injected into all tie line nodes. The absolute value of is less than a given threshold , then convergence is determined, and the updated optimal output power is used as the optimal tie line output power; if the convergence condition is not met, the optimal output power is repeatedly updated until the convergence condition is met; where, is a small positive number;

[0387] Among them, taking the tie line node b in the provincial optimal unit output model as an example, the tie line output power of this node is The training process of the price response function curve is:

[0388] 1) In the Kth iteration Optimize and obtain the optimal unit output plan for each time period and the node marginal electricity price of node b.

[0389] 2) Calculate the sum of the unit output in each period, and pair the total power generation load in each period with the marginal electricity price of node b in each period to obtain Bigram ;in, The number of optimization periods.

[0390] 3) The obtained The binary groups are added to the binary group sequence in the historical iteration process, and the system generation load-cost segment characteristic curve is drawn according to this sequence, and the slope of each segment of the curve is the corresponding value.

[0391] 4) Given the current state of power generation load By translating the origin of the system power generation load-cost segmented characteristic curve to the position of the given current state power point load, we can get the relationship between the output power increment and the cost increment of node b in the current state, that is, we can get the output power-cost increment curve of node b, which represents the change of power generation cost of the provincial power system with the output power of the interconnection line node.

[0392] The output unit, training unit and judgment unit of this embodiment aim to minimize the output price of the tie line node and construct an optimal tie line power model, which helps to achieve the optimal configuration of power resources while ensuring the safety of power grid operation, thereby saving costs. The optimal tie line power model is updated by training the price response function. This method enables the model to be dynamically adjusted according to price changes in the actual power market, ensuring that the model can still maintain its optimization performance in the face of market price fluctuations. By introducing a convergence condition to determine whether the adjustment amount of the power injected into the tie line node has reached a stable state, this ensures the stability and reliability of the model optimization process and avoids infinite loops or divergences.

[0393] In one embodiment, the curve module 40 includes a curve unit, an extraction unit, a removal unit, a discrete subunit, a processing subunit and an iteration subunit; wherein the curve unit is a process of establishing a reference curve, the extraction unit and the removal unit are processes of performing data removal to establish a first power quantity-price curve set, the discrete subunit and the processing subunit are processes of performing data aggregation through the K-Means algorithm, and the iteration subunit is a process of performing data removal again and establishing an equivalent quantity-price curve of a virtual node, specifically:

[0394] The curve unit is used to first collect historical clearing data including clearing prices of different power sources at various trading volume levels;

[0395] Curve units are also used to estimate polynomial coefficients using the least squares method (LSM) based on historical clearing data. , that is, for each data point , the computational model predicts The error between the actual observation ,have:

[0396]

[0397] Secondly, for polynomial fitting, a design matrix X is constructed, and the columns of the design matrix X are composed of the powers of the independent variables x;

[0398] Curve unit, also used to construct normal equations using the design matrix and observation vector y to solve the vector of model parameters ;in, The matrix representing the polynomial coefficients, the specific mathematical formula is:

[0399]

[0400] Curve unit, also used if is reversible, then the parameter vector is calculated by matrix inversion ,Right now ; According to the obtained parameters and the polynomial coefficients Calculate the reference curve ,as follows:

[0401]

[0402] Where N represents the number of samples in the data set. represents the i-th observation value, Representing polynomials The predicted value at; and, the reference curve Used to reflect the general situation of power supply being cleared historically.

[0403] It is represented as the transpose of X;

[0404] x is the independent variable, usually the trading volume is used as the independent variable; represents the coefficients of the polynomial, and n is the order of the polynomial.

[0405] The extraction unit is used to discretize the quantity-price curves of several power supplies at the same trading volume level to obtain a series of data point sets; wherein the data point sets represent prices under different trading volumes, and there is The corresponding prices are ;

[0406] Eliminate the unit, which is used to calculate the difference between each volume price curve and the reference curve Euclidean distance D at each volume level;

[0407] The elimination unit is also used to quantify the difference between each quantity-price curve and the reference curve in several power quantity-price curves through the Euclidean distance D. The difference between them; the smaller the Euclidean distance D is, the closer the volume-price curve is to the reference curve. The closer they are, the smaller the difference is.

[0408] The elimination unit is further used to eliminate the power quantity price curves that are continuously higher than the reference curve in the preset key features according to the preset threshold value of the Euclidean distance D, so as to obtain the remaining first power quantity price curve set;

[0409] Among them, the Euclidean distance D is expressed as:

[0410]

[0411] in, Indicates that the volume price curve is in the trading volume The corresponding price below, represents the price of the reference curve under the same trading volume; n is the total number of trading volume levels, that is, the number of discrete points of the volume-price curve.

[0412] A discretization subunit, used for discretizing a plurality of power quantity-price curves to obtain cost values ​​under different processing levels to form a first data point set;

[0413] A processing subunit, for performing standardization processing on the first data point set based on a K-Means algorithm so that the mean value of each feature is 0 and the standard deviation is 1;

[0414] The processing subunit is further used to select k initialized centroids in the first data point set based on the preset number of clusters k to be divided , expressed as ,in is the centroid of the kth cluster, or the K-means++ algorithm can be used to initialize the centroid to obtain the second set of data points; where k is obtained based on prior experience or through cross-validation;

[0415] The processing subunit is also used to set the maximum number of iterations N. The control algorithm will perform a maximum of N iterations to update the centroid and cluster division;

[0416] Processing subunit, also used for each iteration n (from 1 to N): the cluster partition needs to be initialized is zero or an empty set, where t represents the number of iterations, ;

[0417] And, for each sample curve in the second data point set , the following calculations are required:

[0418] 1) Calculate the sample curve With all centroid vectors The distance between . Usually calculated using Euclidean distance;

[0419] 2) Find the centroid with the smallest distance ,Right now Yes all the smallest of the

[0420] 3) The sample curve Assign to The corresponding cluster Update

[0421] In the second data point set, update the corresponding cluster centroid according to the mean or median of all points in several clusters ;

[0422] Among them, the K-Means++ algorithm is an optimization of the K-Means random initialization centroid method, and its optimization strategy can be expressed as follows:

[0423] 1) Randomly select a point from the input data point set as the first cluster center ;

[0424] 2) For each point in the data set , calculate its distance to the nearest cluster center among the selected cluster centers ;

[0425] 3) Select a new data point as the new cluster center. The selection principle is: Larger points have a greater probability of being cluster centers;

[0426] 4) Repeat steps 2 and 3 above until k cluster centroids are selected;

[0427] 5) Use these k centroids as the initial centroids to run the standard K-Means algorithm.

[0428] Iteration subunit, used to output the final cluster division if the centroid change is less than the preset threshold or reaches the preset number of iterations N , the power quantity-price curves corresponding to several clusters are taken as clustering results, and the quantity-price curves that are on the edge or have large differences with the main cluster are eliminated according to the clustering results, and the remaining second power quantity-price curve set after inverse standardization screening is obtained; wherein, the cluster division Each of Contains the allocation All sample points of the centroid;

[0429] The iterative subunit is further used to extract key features, such as marginal cost, from the power data of each node according to the first power quantity price curve set and the second power quantity price curve set;

[0430] The iterative subunit is also used to calculate the weight of each node power source under different characteristics based on the three specific characteristics of power generation, cost and unit reliability among the key characteristics by using a multi-objective optimization method to represent the equivalent characteristics of the virtual node, and to fit the quantity-price curve of each node to obtain a fitting curve;

[0431] The iterative subunit is also used to calculate the equivalent quantity-price curve of the virtual node according to the fitting curve;

[0432] where, for each treatment level q, the cost of the isoquantity-price curve is It can be expressed as:

[0433]

[0434] in, represents the weight of the i-th node, It represents the operating cost of the ith node at the output level q.

[0435] Among them, the multi-objective optimization method is:

[0436] 1) Establish the power generation cost objective function and environmental cost / carbon emission cost objective function ,Right now:

[0437] ;

[0438]

[0439] For parameters and have:

[0440] ;

[0441]

[0442] in, represents the carbon emission cost of the i-th node at time t; represents the carbon emissions of the power system; represents the weight of the i-th node; n represents the total number of node power supplies;

[0443] , as well as represents the fuel cost coefficient of thermal power unit i; It represents the carbon emission intensity of thermal power unit i, which varies with the load rate of the unit. When the thermal power is in the deep peak regulation DPR state, its carbon emission intensity is significantly increased compared with the conventional output; Represents the time interval from t-1 to t.

[0444] 2) Establish decision variables is used to represent the weight of the i-th node, and Satisfy the normalization condition .

[0445] 3) Use multi-objective optimization algorithm - multi-objective particle swarm optimization PSO algorithm for calculation:

[0446] A. Iterative formula for particle velocity and position:

[0447] Suppose there are m particles in D-dimensional space, and the position of the i-th particle is a vector ; The velocity of the i-th particle is a vector ; The optimal position searched by the i-th particle is ; The optimal position searched by the entire particle swarm is ; Particle speed and location The iteration formula is as follows:

[0448] ;

[0449]

[0450] in, ; is the inertia parameter; , is the learning factor; , is a random number; the three items on the right side of the particle speed iteration formula are the historical speed memory part, the cognitive part and the social part.

[0451] In order to improve the global optimization ability of the initial particle swarm and enable the particle swarm to jump out of the local optimal solution, the specific update rules of the parameters are as follows:

[0452] ;

[0453]

[0454] in, The values ​​of are 0.9 and 0.7 respectively; The values ​​of are 1.5 and 1 respectively.

[0455] B. Adaptive particle optimization strategy:

[0456] During the PSO algorithm solution process, the history of the search particles and the global optimal particle will be continuously updated and guide the remaining search particles to fly to the optimal position, thereby achieving the effect of algorithm convergence. However, the rapid aggregation of search particles will lead to more invalid solutions in close distances and easily cause the algorithm to fall into the local optimum. Therefore, the adaptive operator is introduced to set the particle optimization strategy as follows:

[0457]

[0458] When the particle meets the following position update conditions:

[0459]

[0460] Perform position transformation optimization according to the following method:

[0461]

[0462] Among them, e represents the adaptive operator, Represented as a random number function, Denoted as adaptive threshold, is the Euclidean distance between particle i and the nearest particle j in the D-dimensional space formed by the objective function, It is expressed as the Euclidean distance between particle i and the optimal particle in the population; Q is the optimization judgment threshold.

[0463] Moreover, the initial particles are far away from the optimal particles at the beginning of the algorithm. The particle position update strategy is mainly affected by the adaptive operator term, and the particle diversity is enhanced. In the later stage of the algorithm, the particles gradually approach the global optimal solution. At this time, the adaptive operator term has little effect on the particle position update, and the particles mainly rely on their own positions for optimization, ensuring the algorithm's ability to solve problems in a small range.

[0464] C. Fuzzy strategy:

[0465] The fuzzy decision-making method is used to fuzzify the two objective functions of operating cost and environmental cost. The processing functions are as follows:

[0466]

[0467] in, is the single objective function value of the i-th Pareto optimal solution, is the objective function value after fuzzy processing, and are the maximum and minimum values ​​of the corresponding objective function.

[0468] Then, [operating cost, environmental cost] is taken as a factor set, and a single factor evaluation is performed on the non-inferior solution set obtained by the PSO algorithm to form a fuzzy evaluation matrix. For ease of analysis, the weight vector of the evaluation factor in this embodiment is [0.5, 0.5];

[0469] Finally, a fuzzy comprehensive evaluation is carried out and the best one is selected as one of the components of the final scheduling plan.

[0470] The discrete subunit, processing subunit and iteration subunit of this embodiment discretize the power quantity price curve into a data point set, which can significantly reduce the amount of data, reduce the computational complexity of subsequent processing, and improve processing efficiency. The data point set is clustered based on the preset number of clusters, and the degree of clustering sophistication can be adjusted according to actual needs. The clustering results are continuously optimized by iteratively updating the cluster centroid until the centroid change is less than the preset threshold or reaches the preset number of iterations. This iterative optimization process helps to reduce clustering errors and improve the accuracy and stability of clustering results.

[0471] In addition, the use of Euclidean distance to quantify the difference between the power quantity and price curve and the reference curve can accurately measure the similarity or difference between the curves, providing a reliable basis for the subsequent elimination of abnormal curves. By eliminating power quantity and price curves that are continuously higher than the reference curve in preset key features, those abnormal or unreasonable data points can be excluded, thereby avoiding their adverse effects on subsequent analysis. The use of clustering to divide the power quantity and price curves can not only further identify groups of curves with similar characteristics, but also help to discover potential power market structures or patterns, and by eliminating quantity and price curves that do not meet the preset conditions, the quality of the data can be further improved.

[0472] In one embodiment, the cost module 50 includes a simplification unit, a power supply unit, and a cost unit;

[0473] The simplification unit is used to obtain simplified power data of each node according to the first power quantity-price curve set and the second power quantity-price curve set;

[0474] The power supply unit is used to introduce the optimal tie line output power into the provincial optimal unit output model, and use the simplified power supply data of each node as the power supplier to calculate the power supply allocated to different units;

[0475] The cost unit is used to calculate the average grid-connection cost of the virtual node based on the power sources allocated to different units and the grid-connection cost of the transmission cost.

[0476] In the cost module 50 of this embodiment, the simplified power supply data of each node retains key information and reduces data redundancy, so the complexity and calculation amount of data processing can be effectively reduced. By using the simplified power supply data of each node as the power supplier, the power source allocated to different units is calculated. This process reflects the balance of supply and demand in the power market and the optimal allocation of power resources, which can reduce operating costs and transmission losses as much as possible.

[0477] In one embodiment, the dispatch module 60 includes a clearing unit and a configuration unit; wherein the clearing unit is a process of establishing a regional market clearing function, and the configuration unit is a process of configuring the output of each provincial power grid, specifically:

[0478] The clearing unit is used to establish the regional market clearing function by treating the power supply nodes of other provinces in each provincial power grid as a virtual node, as follows:

[0479]

[0480] For parameters have:

[0481]

[0482] Among them, N represents the total number of generator sets; T represents the total number of time periods considered, where one time period is 15 minutes on D day, and 96 time periods are considered. On D+1 day, two time periods, peak and valley load, are considered, so T is 98; represents the output of unit i in period t; unit operating cost It is a multi-segment linear function related to each output range declared by the unit and the corresponding energy price; Indicates the average price of Internet access fee; Indicates the power transmission power of the inter-provincial transmission channel; Represents the penalty factor for power abandonment of new energy trading units; represents the abandoned power of the new energy trading unit r in time period t;

[0483] It represents the sum of inter-provincial transmission fees (including network loss fees), transmission and distribution fees within the province on the sending side, and transmission fee sharing space. represents the transmission power of inter-provincial power transmission component i in time period t.

[0484] The configuration unit is used to calculate the unit output arrangement according to the optimal interconnection line output power, the equivalent quantity-price curve of the virtual node, the average transmission fee of the virtual node, the operating cost of the power supply in the province, and the regional market clearing function, and configure the output of each provincial power grid according to the unit output arrangement.

[0485] To apply the second embodiment of the present application, please refer to Figure 2, which is a flow chart of the out-of-province power equivalent market clearing optimization method based on price function response provided in the embodiment of the present application, representing the general process of calculating various parameters and performing corresponding processing from the beginning to the end of this embodiment.

[0486] Overall, this embodiment has the following beneficial effects:

[0487] The present invention establishes a provincial optimal unit output model by collecting and analyzing network data, which fully considers the actual situation and needs within each provincial power grid. By simplifying and equivalence processing of out-of-province power sources, an equivalent quantity-price curve of a virtual node is established, which simplifies the complex inter-provincial power trading problem and improves the accuracy and efficiency of the calculation. Introducing the optimal interconnection line output power into the provincial optimal unit output model for calculation can accurately predict and evaluate the impact of different scheduling schemes on the power system, providing strong data support for making scientific scheduling decisions. For unit output arrangement, the introduction of the optimal interconnection line output power ensures the coordination and balance of power transmission and exchange between different provincial power grids; considering the economic factor of network access fees, the unit output arrangement is more in line with market laws, which helps to reduce the cost of power transmission and trading; by comparing the equivalent quantity-price curves of different virtual nodes, it is possible to identify which power sources are more competitive under current market conditions, thereby giving priority to the output of these power sources and achieving optimal allocation of power resources;

[0488] In addition, the provincial network equivalence adopts the Ward equivalence method, with the tie line node as the reserved node, eliminating the internal nodes other than the tie line node, so as to obtain the equivalent admittance matrix of the tie line node and the equivalent injected power on the node, so as to accurately consider the influence of the provincial network parameters on the global power flow distribution; the provincial optimal unit output model adopts the minimization of the unit output cost as the optimization function, considering the relevant constraints to obtain the tie line equivalent node power and marginal price under the given tie line power; the regional optimal tie line model adopts the minimization of the tie line node output price, considering the tie line node power balance and other related constraints to obtain the tie line output power, and cyclically generates / updates the given tie line equivalent node power output price function until convergence; the virtual node equivalence method of the out-of-province market is adopted, and the out-of-province power source is equivalent to the same node to participate in the regional market clearing, considering the weight calculation of the multi-objective optimization function and the clearing function of the regional market, so as to obtain the optimization function of efficient clearing. Therefore, this embodiment has a more flexible, comprehensive and accurate modeling capability when processing the power data of multiple provinces with complex coupling such as the power market, which can provide more powerful support for the efficient operation and decision-making of the power market.

[0489] Embodiment three:

[0490] The embodiment of the present application provides a computer-readable storage medium, the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the method for optimizing the clearing of an out-of-province power supply equivalent market;

[0491] Wherein, the method for optimizing the clearing of the equivalent market of power sources outside the province, if implemented in the form of a software functional unit and used as an independent product, can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0492] The above are preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for optimizing the clearing of the out-of-province power supply equivalent market, characterized in that: include: Obtain network data of provincial power grids; Establishing a provincial optimal unit output model according to the network data; specifically: equating the nodes and branches in the provincial power grids to a single node, obtaining an equivalent admittance matrix of the provincial power grid after being equivalent; generating an equivalent global virtual network admittance matrix by combining the equivalent admittance matrix and the inter-provincial tie line node admittance matrix; Based on the global virtual network admittance matrix, with the goal of minimizing the unit output cost, combined with preset constraints, the provincial optimal unit output model is established; Optimizing the tie line power function of each provincial power grid to obtain the optimal tie line output power; By simplifying and equalizing the power sources outside the province, an equivalent quantity-price curve of the virtual node is established; The optimal tie line output power is introduced into the provincial optimal unit output model for calculation to obtain the average network cost of the virtual node; Calculating the unit output arrangement according to the optimal tie line output power, the equivalent quantity-price curve and the average grid access fee, and configuring the output of each provincial power grid according to the unit output arrangement; Wherein, based on the global virtual network admittance matrix, with the goal of minimizing the unit output cost, combined with preset constraints, the provincial optimal unit output model is established, specifically: According to the total number of units in each provincial power grid and the output of the units at different time periods, a first objective function is established with the goal of minimizing the unit output cost; A first set of constraint conditions is established based on load balance between nodes, operation restrictions of the units and line transmission capacity; Based on the global virtual network admittance matrix, according to the first objective function and the first constraint condition, the provincial optimal unit output model is established.

2. The method for optimizing the clearing of the out-of-province power supply equivalent market as claimed in claim 1, characterized in that: The tie line power model of each provincial power grid is optimized to obtain the optimal tie line output power, which is specifically: Based on the tie line node output power of each provincial power grid, an optimal tie line power model is constructed with the goal of minimizing the tie line node output price; Outputting the optimal output power of each tie line according to the optimal tie line power model; Updating the optimal output power of the optimal tie-line power model by training a preset price response function; If the adjustment amount of the power injected by all tie-line nodes reaches a preset convergence condition, the updated optimal output power is used as the optimal tie-line output power.

3. The method for optimizing the clearing of the out-of-province power supply equivalent market as claimed in claim 2, characterized in that: Based on the tie line node output power of each provincial power grid, with the goal of minimizing the tie line node output price, an optimal tie line power model is constructed, specifically: According to the number of tie line nodes of the provincial power grids, combined with the corresponding relationship between the change in the output power of the tie line nodes and the price, a second objective function is established with the goal of minimizing the output price of the tie line nodes; A second set of constraint conditions is established according to the transmission capacity of the tie line and the output power of the node; Based on the power generation power in each time period and the equivalent injection power of the interconnection line nodes obtained by optimizing the provincial optimal unit output, the optimal interconnection line power model is established according to the second objective function and the second constraint condition.

4. The method for optimizing the clearing of the out-of-province power supply equivalent market as claimed in claim 2, characterized in that: The price response function is specifically: According to the initial power of the tie lines of the provincial power grids, the provincial optimal unit output model is optimized to obtain a model optimization result; According to the optimization results of the model, the marginal price of the tie line power output and the equivalent injection power of the tie line node at different time periods and different power generation load levels are calculated to obtain the calculation results; The power generation state of the current unit is corrected based on the calculation result, and the price response function of the unit after correction is obtained.

5. The method for optimizing the clearing of the out-of-province power supply equivalent market as claimed in claim 1, characterized in that: By simplifying and equalizing the power sources outside the province, an equivalent quantity-price curve of the virtual node is established, which is as follows: Extract features from a number of power quantity and price curves of the provincial power grids to obtain a data point set; Based on the data point set, quantifying the difference between each quantity-price curve and the reference curve by Euclidean distance, eliminating the power quantity-price curves that are continuously higher than the reference curve in preset key features, and obtaining the remaining first power quantity-price curve set; Dividing the plurality of power quantity-price curves by clustering to obtain clustering results, and eliminating quantity-price curves that do not meet preset conditions according to the clustering results to obtain the remaining second power quantity-price curve set; An equivalent quantity-price curve of the virtual node is established according to the first power quantity-price curve set and the second power quantity-price curve set.

6. The method for optimizing the clearing of the out-of-province power supply equivalent market as claimed in claim 5, characterized in that: The power quantity-price curves are divided by clustering to obtain clustering results, which are specifically: Discretizing the plurality of power quantity-price curves to obtain a first data point set; Based on the preset number of clusters to be divided, the first data point set is subjected to standardization and clustering processing to obtain a second data point set; In the second data point set, updating the corresponding cluster centroids according to the means of all points in the plurality of clusters; If the centroid change is less than a preset threshold or reaches a preset number of iterations, the power quantity price curves corresponding to the plurality of clusters are taken as the clustering result.

7. The method for optimizing the clearing of the out-of-province power supply equivalent market as claimed in claim 5, characterized in that: The optimal tie line output power is introduced into the provincial optimal unit output model for calculation to obtain the average network cost of the virtual node, which is specifically: Acquire simplified power data of each node according to the first power quantity-price curve set and the second power quantity-price curve set; The optimal tie line output power is introduced into the provincial optimal unit output model, and the simplified power supply data of each node is used as the power supplier to calculate the power supply allocated to different units; Based on the power sources allocated to the different units, the average grid-connection fee of the virtual node is calculated according to the grid-connection fee of the transmission cost.

8. An out-of-province power supply equivalent market clearing optimization device, characterized in that: It includes data module, model module, power module, curve module, cost module and scheduling module; Wherein, the data module is used to obtain network data of each provincial power grid; The model module is used to establish a provincial optimal unit output model according to the network data; specifically: the nodes and branches in the provincial power grids are equivalent to a single node to obtain an equivalent admittance matrix of the provincial power grid; by splicing the equivalent admittance matrix and the inter-provincial tie line node admittance matrix, an equivalent global virtual network admittance matrix is ​​generated; based on the global virtual network admittance matrix, with the goal of minimizing the unit output cost, combined with preset constraints, the provincial optimal unit output model is established; The power module is used to optimize the tie line power function of each provincial power grid to obtain the optimal tie line output power; The curve module is used to establish an equivalent quantity-price curve of a virtual node by simplifying and equivalent processing the power supply outside the province; The cost module is used to introduce the optimal tie line output power into the provincial optimal unit output model for calculation to obtain the average network cost of the virtual node; The dispatching module is used to calculate the unit output arrangement according to the optimal tie line output power, the equivalent quantity-price curve and the average grid access fee, and configure the output of each provincial power grid according to the unit output arrangement; Wherein, based on the global virtual network admittance matrix, with the goal of minimizing the unit output cost, combined with preset constraints, the provincial optimal unit output model is established, specifically: According to the total number of units in each provincial power grid and the output of the units at different time periods, a first objective function is established with the goal of minimizing the unit output cost; A first set of constraint conditions is established based on load balance between nodes, operation restrictions of the units and line transmission capacity; Based on the global virtual network admittance matrix, according to the first objective function and the first constraint condition, the provincial optimal unit output model is established.

9. The out-of-province power supply equivalent market clearing optimization device as claimed in claim 8, characterized in that: The power module includes a construction unit, an output unit, a training unit and a judgment unit; The construction unit is used to construct an optimal tie line power model based on the tie line node output power of each provincial power grid with the goal of minimizing the tie line node output price; The output unit is used to output the optimal output power of each tie line according to the optimal tie line power model; The training unit is used to update the optimal output power of the optimal tie-line power model by training a preset price response function; The judgment unit is used to use the updated optimal output power as the optimal tie-line output power if the adjustment amount of the power injected by all tie-line nodes reaches a preset convergence condition.

10. The out-of-province power supply equivalent market clearing optimization device as claimed in claim 9, characterized in that: The construction unit includes a building subunit, a condition subunit and an output subunit; The establishing subunit is used to establish a second objective function based on the number of tie line nodes of each provincial power grid and the corresponding relationship between the change in the output power of the tie line nodes and the price, with the goal of minimizing the output price of the tie line nodes; The condition subunit is used to establish a second constraint condition set according to the transmission capacity of the tie line and the node output power; The output subunit is used to establish the optimal interconnection line power model based on the power generation power in each time period and the equivalent injection power of the interconnection line node obtained by optimizing the provincial optimal unit output, according to the second objective function and the second constraint condition.

11. The out-of-province power supply equivalent market clearing optimization device as claimed in claim 9, characterized in that: The price response function is specifically: According to the initial power of the tie lines of the provincial power grids, the provincial optimal unit output model is optimized to obtain a model optimization result; According to the optimization results of the model, the marginal price of the tie line power output and the equivalent injection power of the tie line node at different time periods and different power generation load levels are calculated to obtain the calculation results; The power generation state of the current unit is corrected based on the calculation result, and the price response function of the unit after correction is obtained.

12. The out-of-province power supply equivalent market clearing optimization device as claimed in claim 8, characterized in that: The curve module includes an extraction unit, a elimination unit, a clustering unit and a curve unit; The extraction unit is used to extract features from a number of power quantity and price curves of the provincial power grids to obtain a data point set; The elimination unit is configured to quantify the difference between each quantity-price curve and the reference curve by Euclidean distance based on the data point set, eliminate the power quantity-price curves that are continuously higher than the reference curve in preset key features, and obtain the remaining first power quantity-price curve set; The clustering unit is used to divide the plurality of power quantity-price curves by clustering to obtain a clustering result, and eliminate the quantity-price curves that do not meet the preset conditions according to the clustering result to obtain the remaining second power quantity-price curve set; The curve unit is used to establish an equivalent quantity-price curve of the virtual node according to the first power quantity-price curve set and the second power quantity-price curve set.

13. The out-of-province power supply equivalent market clearing optimization device as claimed in claim 12, characterized in that: The clustering unit includes a discrete subunit, a processing subunit, an updating subunit and an iterative subunit; The discrete subunit is used to discretize the plurality of power quantity-price curves to obtain a first data point set; The processing subunit is used to perform standardization and clustering processing on the first data point set based on a preset number of clusters to be divided, so as to obtain a second data point set; The updating subunit is used to update the corresponding cluster centroids in the second data point set according to the means of all points in the plurality of clusters; The iterative subunit is used to take the power quantity price curves corresponding to the plurality of clusters as the clustering result if the centroid change is less than a preset threshold or reaches a preset number of iterations.

14. The out-of-province power supply equivalent market clearing optimization device as claimed in claim 12, characterized in that: The cost module includes a simplification unit, a power supply unit and a cost unit; The simplification unit is used to obtain simplified power data of each node according to the first power quantity-price curve set and the second power quantity-price curve set; The power supply unit is used to introduce the optimal tie line output power into the provincial optimal unit output model, and use the simplified power supply data of each node as the power supplier to calculate the power supply allocated to different units; The cost unit is used to calculate the average grid-connection cost of the virtual node based on the power sources allocated to the different units and the grid-connection cost of the transmission cost.

15. A storage medium, characterized in that: The storage medium stores a computer program, which is called and executed by a computer to implement an out-of-province power equivalent market clearing optimization method as described in any one of claims 1 to 7 above.

Images