Method for accelerating scheduling of an interconnected electrical system based on branch-and-bound search information

Abstract

This invention discloses a method for accelerating the scheduling of electrical interconnected systems based on branch-and-bound search information, comprising the following steps: 1) constructing an electrical-gas interconnected system scheduling model; 2) solving the electrical-gas interconnected system scheduling model to obtain N relaxed solutions, thereby constructing a search information dataset; 3) evaluating the search information dataset to obtain an upper bound of the optimal solution of the electrical-gas interconnected system scheduling model; 4) based on the upper bound of the optimal solution of the electrical-gas interconnected system scheduling model, solving the electrical-gas interconnected system scheduling model to obtain the optimal solution of the electrical-gas interconnected system scheduling model. This invention can accelerate the convergence of the electrical-gas interconnected system scheduling solution while ensuring optimality.

Description

Technical Field

[0001] This invention relates to the field of power systems and their automation, specifically to a method for accelerating the scheduling of interconnected electrical systems based on branch-bound search information. Background Technology

[0002] Electricity-gas interconnection systems enable flexible complementary support between electrical and natural gas energy sources, serving as an important means to improve energy utilization efficiency and facilitate energy structure transformation. The mathematical essence of the electricity-gas interconnection system scheduling problem lies in operational optimization considering the physical constraints of the natural gas and electrical systems. Under the premise of ensuring the electricity-gas interconnection system can effectively cope with changes in operating conditions such as load demand and clean energy output, it achieves the optimal allocation of adjustable resources. Its accuracy and efficiency directly affect the safety and economy of the electricity-gas interconnection system.

[0003] Due to the discrete variables in the model and the convergence and robustness requirements in the computation, the scheduling decision problem of the electric-gas interconnected system is usually constructed as a mixed-integer linear programming (MILP) problem and is commonly solved using commercial MILP solvers such as CPLEX and GUROBI. However, the piecewise linear formulation of the nonlinear physical operation of the natural gas system introduces a large number of auxiliary discrete variables and their corresponding constraints, making the scheduling decision model larger in scale and more complex, facing a "combinatorial explosion" bottleneck in the solution, and posing a more severe challenge to current operations research optimization techniques.

[0004] Existing research on accelerating the solution of scheduling decision problems in electric-gas interconnected systems often focuses on peripheral model processing. This involves reducing and reconstructing the variables, constraints, and modeling methods of the MILP model using the experience of power system experts or domain knowledge before calling the MILP solver, aiming to minimize the complexity of subsequent solver processing. While some of these processing methods do not affect the optimality of the MILP model, many peripheral model processing methods employ heuristic approaches to reduce the operations research optimization space in order to significantly improve computational efficiency, failing to guarantee optimality or even feasibility. Furthermore, they cannot achieve accurate assessment of error boundaries in engineering applications. Summary of the Invention

[0005] The purpose of this invention is to provide a method for accelerating the scheduling of electrical interconnection systems based on branch-bound search information, comprising the following steps:

[0006] 1) Construct a scheduling model for the electrical-pneumatic interconnection system;

[0007] 2) Perform preliminary calculations on the scheduling model of the electric-gas interconnection system to obtain N relaxed solutions, thereby constructing a search information dataset;

[0008] 3) Evaluate the search information dataset to obtain the upper bound of the optimal solution of the electrical-electric interconnection system scheduling model;

[0009] 4) Based on the upper bound of the optimal solution of the electric-gas interconnection system scheduling model, the electric-gas interconnection system scheduling model is subsequently solved to obtain the optimal solution of the electric-gas interconnection system scheduling model.

[0010] Furthermore, the power-gas interconnection system scheduling model includes a power system scheduling model and a natural gas system scheduling model.

[0011] Furthermore, the objective function of the power system dispatch model is as follows:

[0012]

[0013] In the formula, Represent the set of scheduling time periods and the set of scheduling units, respectively; constants These represent the output cost coefficient and start-up cost coefficient of each unit, respectively; continuous decision variable p g,t This represents the output of each unit at different times; discrete decision variable x g,t This flag indicates whether unit g is started during time period t; t represents any scheduling time period; g represents any scheduling unit.

[0014] The constraints of the power system dispatch model include load balance constraints, line power flow constraints, logical constraints between discrete decision variables of generating units, generating unit capacity constraints, and generating unit ramping constraints.

[0015] The load balancing constraints are as follows:

[0016]

[0017] In the formula, d represents the set of system loads; d represents any system load; the constant D d,t This indicates the load of each load node at different times.

[0018] The power flow constraints for the lines are as follows:

[0019]

[0020] In the formula, b represents the set of network lines; b represents any network line; the constant F b Represents the maximum capacity of each line; constant PTDF g,b PTDF constant d,b Indicates the power transfer distribution factor;

[0021] The logical constraints between the discrete decision variables of the unit are shown below:

[0022]

[0023]

[0024]

[0025] In the formula, discrete decision variable y g,t ,z g,t These represent the operating status of unit g during time period t and whether it is shut down; constants. These represent the minimum start-up and shutdown times for each unit; y g,t-1 This indicates the operating status of unit g during the time period t-1; x g,j Discrete decision variable x g,t This indicates whether unit g is started during time period j; z g,j This indicates whether unit g is shut down during time period j;

[0026] The unit capacity constraints are as follows:

[0027]

[0028] In the formula, constant These represent the minimum and maximum output of each unit, respectively.

[0029] The climbing constraints are as follows:

[0030]

[0031]

[0032] In the formula, constant These represent the ramp-up / ramp-down rates and start-up / shutdown rates of each unit, respectively.

[0033] Furthermore, the objective function of the natural gas system scheduling model is as follows:

[0034]

[0035] In the formula, Let represent the gas source and the pipeline set, respectively; s represents any gas source; m and n represent the beginning and end of the pipeline, respectively; constants. These represent the gas source cost coefficient and the pipeline storage cost coefficient, respectively; continuous decision variable f s,t ,L m,n,t These represent the outflow rate of each gas source and the pipe inventory of each pipeline at different time periods;

[0036] The constraints of the natural gas system scheduling model include gas source output constraints, compressor constraints, gas node constraints, and pipeline constraints.

[0037] The gas source outlet constraints are as follows:

[0038]

[0039] In the formula, the constant F s ,constant These represent the minimum and maximum air output of the gas source, respectively;

[0040] The compressor constraints include compressor flow rate upper limit constraints and compressor start and end pressure constraints.

[0041] The compressor flow rate upper limit constraint is as follows:

[0042]

[0043] In the formula, the set This refers to a compressor; the subscript 'v' indicates flow into the compressor, and the subscript 'w' indicates flow out of the compressor; constants. Represents the upper limit of compressor capacity; continuous decision variable f v,w,t This indicates the airflow rate through the compressor at different times.

[0044] The compressor's starting and ending pressure constraints are shown below:

[0045]

[0046] In the formula, constant These represent the lower and upper limits of the compression coefficient for each compressor, respectively; the continuous decision variable π v,t ,π w,t These represent the initial and final pressures of the compressor at different time periods;

[0047] The gas node constraints include node pressure constraints and node airflow balance constraints;

[0048] The nodal pressure constraints are shown below:

[0049]

[0050] In the formula, n' represents the set of gas network nodes; n' represents the set of gas network nodes; constant These represent the upper and lower limits of pressure at each node; the continuous decision variable π. n,t This indicates the pressure at each node during different time periods;

[0051] The nodal airflow balance constraints are shown below:

[0052]

[0053] In the formula, Represent the gas turbine unit set and the gas load set, respectively; constant D d,t ,η v,w These represent the nodal gas load and the compressor gas consumption coefficient, respectively; continuous decision variable f m,t ,f n,t These represent the inflow and outflow airflow at the beginning and end of the pipe, respectively; continuous decision variable f s,t This indicates the outflow rate of each gas source at different time periods; f v,t f w,t These represent the air flow rates flowing into and out of the compressor, respectively.

[0054] Gas consumption f of gas turbine unit g,t As shown below:

[0055]

[0056] In the formula, constant Indicates the gas consumption coefficient of the gas turbine unit;

[0057] Pipe constraints include pipe capacity constraints, pipe inventory constraints, and the Weymouth equation;

[0058] Pipeline capacity constraints are shown below:

[0059]

[0060] In the formula, constant Indicates the maximum capacity of each pipe; f m,n,t This represents the pipeline capacity during time period t;

[0061] The storage constraints are as follows:

[0062]

[0063]

[0064] In the formula, constant L represents the pipe storage coefficient for each pipeline; m,n,t-1 These represent the pipe inventory of each pipe during time period t-1;

[0065] The Weymouth equation is shown below:

[0066]

[0067] In the formula, K m,n It is a constant;

[0068] The linearized Weymouth equation is shown below:

[0069] h(x)=∑ j∈{1,...,P} (A j δ j +B j u j ) (twenty one)

[0070] u j X j ≤δ j ≤u j X j+1 (twenty two)

[0071] ∑ j∈{1,...,P} u j ≤1 (23)

[0072] ∑ j∈{1,...,P} δ j =x (24)

[0073] In the formula, A j B j Represents a linear function within each segment; The segment in which the independent variable is located; p represents the number of segments; auxiliary continuous decision variables. represents the values ​​of the independent variable after linearization and the auxiliary discrete decision variable; h(x) represents the piecewise linear function of the Weymouth equation.

[0074] Furthermore, methods for solving the scheduling model of the electrical-electric interconnection system include the branch and bound algorithm;

[0075] The elements of the search information dataset include sample inputs and corresponding labels; the sample input is the objective function value obj of the relaxation solution, and the label corresponding to the sample input is denoted as . The mapping relationship between sample input and labels is as follows:

[0076] Among them, the label As shown below:

[0077]

[0078] In the formula, j represents the piecewise variable u j The index.

[0079] Furthermore, the steps for evaluating the search information dataset include:

[0080] 1) Calculate the distance d between all samples in the search information dataset and the test target; the test target is the current lower bound of the optimal solution;

[0081] The distance d between the sample and the test target is shown below:

[0082]

[0083] In the formula, q belongs to d(obj,obj) q The smallest set of k samples; obj q To determine the value of the objective function for the sample;

[0084] 2) Calculate the effective range of values ​​for the piecewise auxiliary discrete decision variables. and

[0085] The parameter Δ is shown below:

[0086]

[0087] In the formula, the parameters

[0088] 3) Based on the valid value range and Establish a small-scale auxiliary MILP model;

[0089] 4) Solve the small-scale auxiliary MILP model to obtain the upper bound of the optimal solution.

[0090] Furthermore, a small-scale auxiliary MILP model is shown below:

[0091]

[0092]

[0093] A system applying the aforementioned electrical interconnection system scheduling acceleration method includes:

[0094] Model building unit: Constructing a scheduling model for an electrical-pneumatic interconnected system;

[0095] Search information dataset construction unit: Solve the scheduling model of the electric-electric interconnection system to obtain N relaxed solutions, thereby constructing the search information dataset;

[0096] Optimization Unit: Evaluates the search information dataset to obtain the upper bound of the optimal solution for the scheduling model of the electric-electric interconnected system;

[0097] Solution Unit: Based on the upper bound of the optimal solution of the electric-gas interconnected system scheduling model, the electric-gas interconnected system scheduling model is solved to obtain the optimal solution of the electric-gas interconnected system scheduling model.

[0098] An electronic device includes: a memory, a processor, and a data bus, wherein the memory stores all machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the data bus, and the machine-readable instructions are executed by the processor to perform the method described thereon.

[0099] A computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to perform the method.

[0100] The technical effects of this invention are undeniable. This invention utilizes information from the initial search process of branch and bound to identify discrete decision variables that need to be optimized in subsequent stages, thereby constructing and solving the corresponding small-scale integer programming model, and embedding it into existing commercial solvers to guide the branch and bound algorithm to remove more redundant search space, thereby accelerating the convergence of the scheduling solution for the electric-gas interconnection system while ensuring optimality. Attached Figure Description

[0101] Figure 1 This forms the basic framework of the proposed method;

[0102] Figure 2 Validation of the effectiveness for different numbers of segments;

[0103] Figure 3 To verify the effectiveness under different load conditions. Detailed Implementation

[0104] The present invention will be further described below with reference to embodiments, but it should not be construed that the scope of the present invention is limited to the following embodiments. Various substitutions and modifications made based on ordinary technical knowledge and common practices in the art without departing from the above-described technical concept of the present invention should be included within the scope of protection of the present invention.

[0105] Example 1:

[0106] See Figures 1 to 3 A method for accelerating the scheduling of electrical interconnected systems based on branch-bound search information includes the following steps:

[0107] 1) Construct a scheduling model for the electrical-pneumatic interconnection system;

[0108] 2) Perform preliminary calculations on the scheduling model of the electric-gas interconnection system to obtain N relaxed solutions, thereby constructing a search information dataset;

[0109] 3) Evaluate the search information dataset to obtain the upper bound of the optimal solution of the electrical-electric interconnection system scheduling model;

[0110] 4) Based on the upper bound of the optimal solution of the electric-gas interconnection system scheduling model, the electric-gas interconnection system scheduling model is subsequently solved to obtain the optimal solution of the electric-gas interconnection system scheduling model.

[0111] Example 2:

[0112] The method for accelerating the scheduling of electrical interconnected systems based on branch-bound search information is the same as in Embodiment 1. Furthermore, the scheduling model of the electrical-gas interconnected system includes a power system scheduling model and a natural gas system scheduling model.

[0113] Example 3:

[0114] The method for accelerating the scheduling of interconnected electrical systems based on branch-bound search information is the same as any one of Embodiments 1-2. Further, the objective function of the power system scheduling model is as follows:

[0115]

[0116] In the formula, Represent the set of scheduling time periods and the set of scheduling units, respectively; constants These represent the output cost coefficient and start-up cost coefficient of each unit, respectively; continuous decision variable p g,t This represents the output of each unit at different times; discrete decision variable x g,t This flag indicates whether unit g is started during time period t; t represents any scheduling time period; g represents any scheduling unit.

[0117] Example 4:

[0118] The method for accelerating the scheduling of electrical interconnection systems based on branch-bound search information is the same as any one of Examples 1-3. Furthermore, the constraints of the power system scheduling model include load balance constraints, line flow constraints, logical constraints between discrete decision variables of generating units, generating unit capacity constraints, and generating unit ramping constraints.

[0119] The load balancing constraints are as follows:

[0120]

[0121] In the formula, d represents the set of system loads; d represents any system load; the constant D d,t This indicates the load of each load node at different times.

[0122] The power flow constraints for the lines are as follows:

[0123]

[0124] In the formula, b represents the set of network lines; b represents any network line; the constant F b Represents the maximum capacity of each line; constant PTDF g,b PTDF constant d,b Indicates the power transfer distribution factor;

[0125] The logical constraints between the discrete decision variables of the unit are shown below:

[0126]

[0127]

[0128]

[0129] In the formula, discrete decision variable y g,t ,z g,t These represent the operating status of unit g during time period t and whether it is shut down; constants. These represent the minimum start-up and shutdown times for each unit; y g,t-1 This indicates the operating status of unit g during the time period t-1; x g,j Discrete decision variable x g,t This indicates whether unit g is started during time period j; z g,j This indicates whether unit g is shut down during time period j;

[0130] The unit capacity constraints are as follows:

[0131]

[0132] In the formula, constant These represent the minimum and maximum output of each unit, respectively.

[0133] The climbing constraints are as follows:

[0134]

[0135]

[0136] In the formula, constant These represent the ramp-up / ramp-down rates and start-up / shutdown rates of each unit, respectively.

[0137] Example 5:

[0138] The method for accelerating the scheduling of electrical interconnected systems based on branch-bound search information is the same as any one of embodiments 1-4. Further, the objective function of the natural gas system scheduling model is as follows:

[0139]

[0140] In the formula, Let represent the gas source and the pipeline set, respectively; s represents any gas source; m and n represent the beginning and end of the pipeline, respectively; constants. These represent the gas source cost coefficient and the pipeline storage cost coefficient, respectively; continuous decision variable f s,t ,L m,n,t These represent the outflow rate of each gas source and the pipe inventory of each pipeline at different time periods;

[0141] Example 6:

[0142] The method for accelerating the scheduling of electrical interconnected systems based on branch and boundary search information is the same as any one of embodiments 1-5. Furthermore, the constraints of the natural gas system scheduling model include gas source output constraints, compressor constraints, gas node constraints, and pipeline constraints.

[0143] The gas source outlet constraints are as follows:

[0144]

[0145] In the formula, constant F s ,constant These represent the minimum and maximum air output of the gas source, respectively;

[0146] The compressor constraints include compressor flow rate upper limit constraints and compressor start and end pressure constraints.

[0147] The compressor flow rate upper limit constraint is as follows:

[0148]

[0149] In the formula, the set This refers to a compressor; the subscript 'v' indicates flow into the compressor, and the subscript 'w' indicates flow out of the compressor; constants. Represents the upper limit of compressor capacity; continuous decision variable f v,w,t This indicates the airflow rate through the compressor at different times.

[0150] The compressor's starting and ending pressure constraints are shown below:

[0151]

[0152] In the formula, constant These represent the lower and upper limits of the compression coefficient for each compressor, respectively; the continuous decision variable π v,t ,π w,t These represent the initial and final pressures of the compressor at different time periods;

[0153] The gas node constraints include node pressure constraints and node airflow balance constraints;

[0154] The nodal pressure constraints are shown below:

[0155]

[0156] In the formula, n' represents the set of gas network nodes; n' represents the set of gas network nodes; constant These represent the upper and lower limits of pressure at each node; the continuous decision variable π. n,t This indicates the pressure at each node during different time periods;

[0157] The nodal airflow balance constraints are shown below:

[0158]

[0159] In the formula, Represent the gas turbine unit set and the gas load set, respectively; constant D d,t ,η v,w These represent the nodal gas load and the compressor gas consumption coefficient, respectively; continuous decision variable f m,t ,f n,t These represent the inflow and outflow airflow at the beginning and end of the pipe, respectively; continuous decision variable f s,t This indicates the outflow rate of each gas source at different time periods; f v,t f w,t These represent the air flow rates into and out of the compressor, respectively; the subscript v indicates the direction of airflow into the compressor, and the subscript w indicates the direction of airflow out of the compressor; the subscripts v and w are collectively referred to as the two cases above.

[0160] Gas consumption f of gas turbine unit g,t As shown below:

[0161]

[0162] In the formula, constant Indicates the gas consumption coefficient of the gas turbine unit;

[0163] Pipe constraints include pipe capacity constraints, pipe inventory constraints, and the Weymouth equation;

[0164] Pipeline capacity constraints are shown below:

[0165]

[0166] In the formula, constant Indicates the maximum capacity of each pipe; f m,n,t This represents the pipeline capacity during time period t;

[0167] The storage constraints are as follows:

[0168]

[0169]

[0170] In the formula, constant L represents the pipe storage coefficient for each pipeline; m,n,t-1 These represent the pipe inventory of each pipe during time period t-1;

[0171] The Weymouth equation is shown below:

[0172]

[0173] In the formula, K m,n It is a constant;

[0174] The linearized Weymouth equation is shown below:

[0175] h(x)=∑ j∈{1,...,P} (A j δ j +B j u j ) (twenty one)

[0176] u j X j ≤δ j ≤u j X j+1 (twenty two)

[0177] ∑ j∈{1,...,P} u j ≤1 (23)

[0178] ∑ j∈{1,...,P} δ j =x (24)

[0179] In the formula, A j B j Represents a linear function within each segment; The segment in which the independent variable is located; p represents the number of segments; auxiliary continuous decision variables. represents the values ​​of the independent variable after linearization and the auxiliary discrete decision variable; h(x) represents the piecewise linear function of the Weymouth equation.

[0180] Example 7:

[0181] The method for accelerating the scheduling of electrical interconnection systems based on branch and bound search information is the same as any one of embodiments 1-6. Furthermore, the method for solving the scheduling model of electrical interconnection systems includes a branch and bound algorithm.

[0182] The elements of the search information dataset include sample inputs and corresponding labels; the sample input is the objective function value obj of the relaxation solution, and the label corresponding to the sample input is denoted as . The mapping relationship between sample input and labels is as follows:

[0183] Among them, the label As shown below:

[0184]

[0185] In the formula, j represents the piecewise variable u j The index is obtained during modeling, as shown in Equation 21.

[0186] Example 8:

[0187] The method for accelerating the scheduling of electrical interconnection systems based on branch-bound search information has the same technical content as any one of embodiments 1-7. Further, the step of evaluating the search information dataset includes:

[0188] 1) Calculate the distance d between all samples in the search information dataset and the test target; the test target is the current lower bound of the optimal solution;

[0189] The distance d between the sample and the test target is shown below:

[0190]

[0191] In the formula, q belongs to d(obj,obj) q The smallest set of k samples; obj q To determine the value of the objective function for the sample;

[0192] 2) Calculate the effective range of values ​​for the piecewise auxiliary discrete decision variables. and

[0193] The parameter Δ is shown below:

[0194]

[0195] In the formula, the parameters

[0196] 3) Based on the valid value range and Establish a small-scale auxiliary MILP model;

[0197] 4) Solve the small-scale auxiliary MILP model to obtain the upper bound of the optimal solution.

[0198] Example 9:

[0199] The method for accelerating scheduling of electrical interconnected systems based on branch-bound search information is the same as any one of embodiments 1-8. Further, a small-scale auxiliary MILP model is shown below:

[0200]

[0201]

[0202] Example 10:

[0203] A system applying the electrical interconnection system scheduling acceleration method based on branch and bound search information as described in any one of Examples 1-9 includes:

[0204] Model building unit: Constructing a scheduling model for an electrical-pneumatic interconnected system;

[0205] Search information dataset construction unit: Solve the scheduling model of the electric-electric interconnection system to obtain N relaxed solutions, thereby constructing the search information dataset;

[0206] Optimization Unit: Evaluates the search information dataset to obtain the upper bound of the optimal solution for the scheduling model of the electric-electric interconnected system;

[0207] Solution Unit: Based on the upper bound of the optimal solution of the electric-gas interconnected system scheduling model, the electric-gas interconnected system scheduling model is solved to obtain the optimal solution of the electric-gas interconnected system scheduling model.

[0208] When the system is in operation, it performs the steps of the method described in any one of Examples 1-9.

[0209] Example 11:

[0210] The system applying the electrical interconnection system scheduling acceleration method based on branch delimitation search information described in any one of Examples 1-10 further includes a storage unit;

[0211] The storage unit is used to store data from the model building unit, the search information dataset building unit, the optimization unit, and the solution unit.

[0212] Example 12:

[0213] The system applying the electrical interconnection system scheduling acceleration method based on branch and bound search information described in any one of Examples 1-10 has the same technical content as Example 11. Furthermore, when the system is working, it performs the following steps:

[0214] 1) Construct a scheduling model for the electrical-pneumatic interconnection system;

[0215] 2) Perform preliminary calculations on the scheduling model of the electrical-pneumatic interconnection system to obtain N relaxed solutions, and construct a search information dataset;

[0216] 3) Evaluate the search information dataset to obtain the upper bound of the optimal solution of the electrical-electric interconnection system scheduling model;

[0217] 4) Based on the upper bound of the optimal solution of the electric-gas interconnection system scheduling model, the electric-gas interconnection system scheduling model is subsequently solved to obtain the optimal solution of the electric-gas interconnection system scheduling model.

[0218] 2. The method for accelerating the scheduling of an electrical interconnection system according to claim 1, wherein the electrical-gas interconnection system scheduling model includes a power system scheduling model and a natural gas system scheduling model.

[0219] Example 13:

[0220] The system applying the electrical interconnection system scheduling acceleration method based on branch-bound search information described in any one of Embodiments 1-10 has the same technical content as any one of Embodiments 11-12. Furthermore, the objective function of the power system scheduling model is as follows:

[0221]

[0222] In the formula, Represent the set of scheduling time periods and the set of scheduling units, respectively; constants These represent the output cost coefficient and start-up cost coefficient of each unit, respectively; continuous decision variable p g,t This represents the output of each unit at different times; discrete decision variable x g,t This flag indicates whether unit g is started during time period t; t represents any scheduling time period; g represents any scheduling unit.

[0223] The constraints of the power system dispatch model include load balance constraints, line power flow constraints, logical constraints between discrete decision variables of generating units, generating unit capacity constraints, and generating unit ramping constraints.

[0224] The load balancing constraints are as follows:

[0225]

[0226] In the formula, d represents the set of system loads; d represents any system load; the constant D d,t This indicates the load of each load node at different times.

[0227] The power flow constraints for the lines are as follows:

[0228]

[0229] In the formula, b represents the set of network lines; b represents any network line; constants PTDF,F b These represent the power transfer distribution factor and the maximum capacity of each line, respectively.

[0230] The logical constraints between the discrete decision variables of the unit are shown below:

[0231]

[0232]

[0233]

[0234] In the formula, discrete decision variable y g,t ,z g,t These represent the operating status of unit g during time period t and whether it is shut down; constants. These represent the minimum start-up and shutdown times for each unit; y g,t-1 This indicates the operating status of unit g during the time period t-1; x g,j Discrete decision variable x g,t This indicates whether unit g is started during time period j; z g,j This indicates whether unit g is shut down during time period j;

[0235] The unit capacity constraints are as follows:

[0236]

[0237] In the formula, constant These represent the minimum and maximum output of each unit, respectively.

[0238] The climbing constraints are as follows:

[0239]

[0240]

[0241] In the formula, constant These represent the ramp-up / ramp-down rates and start-up / shutdown rates of each unit, respectively.

[0242] Example 14:

[0243] The system applying the electrical interconnection system scheduling acceleration method based on branch and boundary search information described in any one of Examples 1-10 has the same technical content as any one of Examples 11-13. Further, the objective function of the natural gas system scheduling model is as follows:

[0244]

[0245] In the formula, Let represent the gas source and the pipeline set, respectively; s represents any gas source; m and n represent the beginning and end of the pipeline, respectively; constants. These represent the gas source cost coefficient and the pipeline storage cost coefficient, respectively; continuous decision variable f s,t,L m,n,t These represent the outflow rate of each gas source and the pipe inventory of each pipeline at different time periods;

[0246] The constraints of the natural gas system scheduling model include gas source output constraints, compressor constraints, gas node constraints, and pipeline constraints.

[0247] The gas source outlet constraints are as follows:

[0248]

[0249] In the formula, constant F s These represent the minimum and maximum air output of the gas source, respectively;

[0250] The compressor constraints include compressor flow rate upper limit constraints and compressor start and end pressure constraints.

[0251] The compressor flow rate upper limit constraint is as follows:

[0252]

[0253] In the formula, the set This refers to a compressor; the subscript 'v' indicates flow into the compressor, and the subscript 'w' indicates flow out of the compressor; constants. Represents the upper limit of compressor capacity; continuous decision variable f v,w,t This indicates the airflow rate through the compressor at different times.

[0254] The compressor's starting and ending pressure constraints are shown below:

[0255]

[0256] In the formula, constant These represent the lower and upper limits of the compression coefficient for each compressor, respectively; the continuous decision variable π v,t ,π w,t These represent the initial and final pressures of the compressor at different time periods;

[0257] The gas node constraints include node pressure constraints and node airflow balance constraints;

[0258] The nodal pressure constraints are shown below:

[0259]

[0260] In the formula, n' represents the set of gas network nodes; n' represents the set of gas network nodes; constant These represent the upper and lower limits of pressure at each node; the continuous decision variable π. n,t This indicates the pressure at each node during different time periods;

[0261] The nodal airflow balance constraints are shown below:

[0262]

[0263] In the formula, Represent the gas turbine unit set and the gas load set, respectively; constant D d,t ,η v,w These represent the nodal gas load and the compressor gas consumption coefficient, respectively; continuous decision variable f m,t ,f n,t These represent the inflow and outflow airflow at the beginning and end of the pipe, respectively; continuous decision variable f s,t This indicates the outflow rate of each gas source at different times;

[0264] Gas consumption f of gas turbine unit g,t As shown below:

[0265]

[0266] In the formula, constant Indicates the gas consumption coefficient of the gas turbine unit;

[0267] Pipe constraints include pipe capacity constraints, pipe inventory constraints, and the Weymouth equation;

[0268] Pipeline capacity constraints are shown below:

[0269]

[0270] In the formula, constant Indicates the maximum capacity of each pipe; f m,n,t This represents the pipeline capacity during time period t;

[0271] The storage constraints are as follows:

[0272]

[0273]

[0274] In the formula, constant L represents the pipe storage coefficient for each pipeline; m,n,t-1 These represent the pipe inventory of each pipe during time period t-1;

[0275] The Weymouth equation is shown below:

[0276]

[0277] The linearized Weymouth equation is shown below:

[0278] h(x)=∑ j∈{1,...,P} (Aj δ j +B j u j (50)

[0279] u j X j ≤δ j ≤u j X j+1 (51)

[0280] ∑ j∈{1,...,P} u j ≤1 (52)

[0281] ∑ j∈{1,...,P} δ j =x (53)

[0282] In the formula, A j B j Represents a linear function within each segment; The segment in which the independent variable is located; p represents the number of segments; auxiliary continuous decision variables. represents the values ​​of the independent variable after linearization and the auxiliary discrete decision variable; h(x) represents the piecewise linear function of the Weymouth equation.

[0283] Example 15:

[0284] The system applying the electrical interconnection system scheduling acceleration method based on branch and bound search information described in any one of Embodiments 1-10 has the same technical content as any one of Embodiments 11-14. Furthermore, the method for solving the electrical interconnection system scheduling model includes a branch and bound algorithm.

[0285] The elements of the search information dataset include sample inputs and corresponding labels; the sample input is the objective function value obj of the relaxation solution, and the label corresponding to the sample input is denoted as . The mapping relationship between sample input and labels is as follows:

[0286] Among them, the label As shown below:

[0287]

[0288] In the formula, j represents the piecewise variable u j The index.

[0289] Example 16:

[0290] The system applying the electrical interconnection system scheduling acceleration method based on branch-bound search information described in any one of Embodiments 1-10 has the same technical content as any one of Embodiments 11-15. Further, the step of evaluating the search information dataset includes:

[0291] 1) Calculate the distance d between all samples in the search information dataset and the test target; the test target is the current lower bound of the optimal solution;

[0292] The distance d between the sample and the test target is shown below:

[0293]

[0294] In the formula, q belongs to d(obj,obj) q The smallest set of k samples;

[0295] 2) Calculate the effective range of values ​​for the piecewise auxiliary discrete decision variables. and

[0296] The parameter Δ is shown below:

[0297]

[0298] 3) Based on the valid value range and Establish a small-scale auxiliary MILP model;

[0299] 4) Solve the small-scale auxiliary MILP model to obtain the upper bound of the optimal solution.

[0300] Example 17:

[0301] The system applying the electrical interconnection system scheduling acceleration method based on branch and bound search information described in any one of Examples 1-10 has the same technical content as any one of Examples 11-16. Further, a small-scale auxiliary MILP model is shown below:

[0302]

[0303]

[0304] Example 18:

[0305] An electronic device includes: a memory, a processor, and a data bus, wherein the memory stores all machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the data bus, and the machine-readable instructions are executed by the processor to perform the method described in any one of embodiments 1-9.

[0306] Example 19:

[0307] A computer-readable storage medium storing a computer program that, when executed by a processor, performs the method described in any one of embodiments 1-9.

[0308] Example 20:

[0309] A method for accelerating scheduling of electrically interconnected systems based on branch and bound search information is described below:

[0310] First, a search information dataset is constructed. Based on the large amount of relaxed solution information generated in the initial search phase of branch and bound, a search information dataset is constructed to help evaluate the values ​​of piecewise auxiliary discrete decision variables in the optimal solution.

[0311] Secondly, the piecewise auxiliary discrete decision variables are evaluated. The search information dataset is fitted using the k-nearest neighbor regression (KNN) algorithm to obtain the evaluation results of the effective range of values ​​for the piecewise auxiliary discrete decision variables.

[0312] Then, a small-scale auxiliary MILP model is constructed. Based on the evaluation results, a small-scale auxiliary MILP model is constructed to reduce a large number of piecewise auxiliary discrete decision variables. Through efficient solution and feasibility repair strategies, a high-quality feasible solution to the scheduling decision problem of the electric-gas interconnection system is obtained, and it is used to assist the branch and bound algorithm to remove redundant search space, thereby accelerating convergence while ensuring optimality.

[0313] Finally, simulation results from 30 examples of RTS-GMLC power and natural gas systems under different load levels and linear segmentation numbers demonstrate that, compared to directly using commercial MILP solvers, the proposed method can achieve an average speedup of 4.20 times without sacrificing optimality, thus verifying the effectiveness of the proposed method.

[0314] The specific method steps of this invention are as follows:

[0315] (1) Constructing a scheduling model for an electrical-electric interconnection system

[0316] 1.1) Establish a power system dispatch model

[0317] The dispatching decision-making cost on the power system side mainly consists of two parts: unit output cost and unit start-up cost.

[0318]

[0319] The set Representing the scheduling period and the scheduling unit respectively; constant These represent the output cost coefficient and start-up cost coefficient of each unit, respectively; continuous decision variable pg,t This represents the output of each unit at different times; discrete decision variable x g,t This indicates whether each unit is started at any given time (i.e., from shutdown to startup).

[0320] System-side constraints include load balance constraints and line power flow constraints. Load balance constraints require that the sum of unit outputs in each time period equals the system load.

[0321]

[0322] The set Represents system load; constant D d,t This represents the load at each load node in each time period. Line power flow constraints typically use DC power flow instead of nonlinear AC power flow, requiring that the branch power flow of each line in each time period does not exceed the line capacity.

[0323]

[0324] The set Represents network lines; constants PTDF,F b These represent the Power Transfer Distribution Factor (PTDF) and the maximum capacity of each line, respectively.

[0325] The constraints that each unit needs to satisfy include logical constraints between discrete decision variables, capacity constraints, and ramp-up constraints. Logical constraints connect the discrete decision variables (all of which are 0-1 variables) through minimum start-up and shutdown times, etc.

[0326]

[0327]

[0328]

[0329] Among them, discrete decision variable y g,t ,z g,t These represent the operating status (on / off) and shutdown status of each unit at different times (i.e., from startup to shutdown); constants These represent the minimum start-up and shutdown times for each unit. Capacity constraints limit the upper and lower limits of unit output:

[0330]

[0331] Where the constant These represent the minimum and maximum output of each unit, respectively. The ramp constraint limits the rate of change of unit output between adjacent time periods:

[0332]

[0333]

[0334] Where the constant These represent the ramp-up / ramp-down rates and start-up / shutdown rates of each unit, respectively.

[0335] 1.2) Establish a natural gas system scheduling model

[0336] A natural gas system includes equipment such as gas sources, pipelines, compressors, and gas engines, each of which must meet its own physical operating constraints.

[0337] The scheduling decision-making cost on the natural gas system side mainly consists of two parts: gas source cost and pipeline storage cost.

[0338]

[0339] The set Let m and n represent the gas source and pipeline, respectively, and m and n represent the beginning and end of the pipeline, respectively; constant. These represent the gas source cost coefficient and the pipeline storage cost coefficient, respectively; continuous decision variable f s,t ,L m,n,t These represent the outflow rate of each gas source and the storage capacity of each pipeline at different times.

[0340] The gas supply must meet upper and lower limit constraints:

[0341]

[0342] Where the constant These represent the minimum and maximum air output of the gas source, respectively.

[0343] Each compressor needs to meet constraints including upper flow rate constraints and start / end pressure constraints. The upper flow rate constraint limits the gas flow rate through the compressor:

[0344]

[0345] The set The compressor is represented by v and w, which represent the beginning and end points of the compressor, respectively; constants. Represents the upper limit of compressor capacity; continuous decision variable f v,w,t This represents the gas flow rate through the compressor at various time points. The start-end pressure constraints limit the pressure at the gas network nodes at the start and end of the compressor.

[0346]

[0347] Where the constant These represent the lower and upper limits of the compression coefficient for each compressor, respectively; the continuous decision variable π v,t ,π w,t These represent the initial and final pressures of the compressor at different time periods.

[0348] Gas node constraints include nodal pressure constraints and nodal airflow balance constraints. Nodal pressure constraints limit the pressure at each gas node:

[0349]

[0350] The set Represents a gas network node; constant These represent the upper and lower limits of pressure at each node; the continuous decision variable π. n,t This represents the pressure at each node during different time periods. The node airflow balance constraint requires that the total inflow and outflow of air at each node be equal during different time periods. The inflow airflow to the node includes the gas source outlet, the outlet airflow at the end of the pipeline, and the outlet airflow at the end of the compressor. The outflow airflow to the node or the consumed airflow includes gas consumption by the gas turbine unit, gas load, inflow airflow at the beginning of the pipeline, inflow airflow at the beginning of the compressor, and gas consumption by the compressor.

[0351]

[0352] The set Representing the gas turbine unit and gas load respectively; constant D d,t ,η v,w These represent the nodal gas load and the compressor gas consumption coefficient, respectively; continuous decision variable f m,t ,f n,t f represents the inflow and outflow of air at the beginning and end of the pipe, respectively. g,t Indicates the gas consumption of the gas turbine unit:

[0353]

[0354] Where the constant This indicates the gas consumption coefficient of the gas turbine unit.

[0355] The constraints that each pipeline needs to satisfy include pipeline capacity constraints, pipeline storage constraints, and the Weymouth equation. Pipeline capacity constraints limit the pipeline flow rate:

[0356]

[0357] Where the constant This represents the maximum capacity of each pipe. Pipe inventory constraints limit the pressure at the beginning and end of the pipe and the varying pipe inventory over time.

[0358]

[0359]

[0360] Where the constant This represents the pipe storage coefficient for each pipe. The Weymouth equation uses a quadratic function to establish the relationship between pipe flow rate and pressure at the beginning and end points of the pipe:

[0361]

[0362] 1.3) Piecewise linearization of the Weymouth equation

[0363] For a nonlinear function f(x) whose independent variable takes values ​​in the range X, an auxiliary continuous decision variable is introduced. This represents the values ​​of the independent variable after linearization, and the auxiliary discrete decision variables (all 0-1 variables). Let f(x) be a segment containing the independent variable, and let it be replaced by a linear function h(x) segmented into P segments as follows:

[0364] h(x)=∑ j∈{1,...,P} (A j δ j +B j u j ) (twenty one)

[0365] Where A j B j Let represent a linear function of the form consisting of two parameters (A and B) within each segment, satisfying the following constraints:

[0366] u j X j ≤δ j ≤u j X j+1 (twenty two)

[0367] ∑ j∈{1,...,P} u j ≤1 (23)

[0368] ∑ j∈{1,...,P} δ j =x (24)

[0369] By applying piecewise linearization to both pipeline flow rate and node pressure, the Weymouth equation can be linearized, transforming the scheduling decision problem of the electro-pneumatic interconnected system into a MILP form.

[0370] (2) Constructing a search information dataset

[0371] Each relaxation solution represents a sample in the search information dataset, with the total number of samples collected in the initial search phase of branch and bound being N. Each sample consists of an input and a label. The sample input represents the information that distinguishes this relaxation solution from other relaxation solutions; therefore, the objective function value obj of the relaxation solution is used as the input. The effective range of values ​​for the piecewise auxiliary discrete decision variables provided by the relaxation solution has three primitive forms: u j =1, u j =0 or 0 < u j <1, to determine the optimal solution u j =1 possible segment set, using u j =1 and 0<u j The average value of j in <1 is used as the label, that is:

[0372]

[0373] This holds true for all nonlinear functions that require piecewise linearization. Therefore, the search information dataset establishes a mapping between the input and the label:

[0374] (3) Evaluation of piecewise auxiliary discrete decision variables

[0375] KNN has a simple structure; its basic principle is to calculate the average of the labels of its k nearest neighbors. Typically, Euclidean distance is used to calculate the distance between the test sample and each training sample.

[0376]

[0377] Since the sample input in the search information dataset is only one-dimensional, this distance is equivalent to the absolute difference between the values ​​of the objective function. Using the current lower bound objective function value as the input to KNN, the output of KNN for the test sample is:

[0378]

[0379] Where q belongs to d(obj,obj) q The smallest set of k samples. Therefore, the output of KNN can be directly adjusted by the number of neighborhoods k. Generally speaking, the larger k is set, the more relaxation solution information is considered, and the larger the effective range of values ​​of the piecewise auxiliary discrete decision variables is.

[0380] (4) Constructing a small-scale auxiliary MILP model

[0381] make These are piecewise auxiliary discrete decision variables. The upper and lower bounds of the valid values ​​(i.e., values ​​of 1). Modify the piecewise linear constraint (23) as follows:

[0382]

[0383]

[0384] when P =1, When the condition is met, equations (28)-(29) are equivalent to the original piecewise linear constraint (23); otherwise, equations (28)-(29) represent reducing the number of segments from P to P.

[0385] For simplicity, assume that the effective range of values ​​for the auxiliary discrete decision variable is related to j. * Symmetric, and for j respectively * Rounding up and down yields the effective range of values ​​for the piecewise auxiliary discrete decision variables. and in:

[0386]

[0387] It can be seen that the larger the Δ is set, the larger the effective range of values ​​for the piecewise auxiliary discrete decision variables will be.

[0388] Because the number of segments is significantly reduced in the auxiliary MILP model, if the effective range of the piecewise auxiliary discrete variables is evaluated too narrowly, the auxiliary MILP model may be infeasible, thus failing to accelerate the branch and bound search in the main process. On the other hand, if the range is evaluated too wide, the solution time of the auxiliary MILP model will be long, with limited improvement on the branch and bound search efficiency in the main process. To balance the feasibility and solution efficiency of the auxiliary MILP model, the parameters k and Δ can jointly determine the range of segment values, with Δ having a greater impact. Therefore, a set of progressively increasing k and Δ parameters is set:

[0389]

[0390] Equation (31) controls the size of the discrete decision variables of the auxiliary MILP model from small to large, so as to quickly find a set of parameters within a limited time to make the auxiliary MILP model feasible, without excessively losing the solution efficiency.

[0391] (5) Parallel embedded solver

[0392] like Figure 1As shown, this is achieved through parallel interaction between the main process and child processes. The main process is the MILP solver, which provides information obtained during the branch-and-bound algorithm search to the child processes and receives feasible solutions from the child processes to accelerate the search. The child processes collect search information from the main process to evaluate the effective range of values ​​for the piecewise auxiliary discrete decision variables, thereby constructing a small-scale auxiliary MILP model with a reduced number of pieces. Finally, the model is solved using a feasibility repair method to obtain a high-quality feasible solution. Since this auxiliary MILP model does not change other physical constraints, the obtained feasible solution is also a feasible solution of the original MILP model and can be used to improve the upper bound in the main process.

[0393] The pseudocode is shown in Table 1.

[0394] Table 1 shows the pseudocode flow of the proposed method.

[0395]

[0396] Example 14:

[0397] The verification experiment of the scheduling acceleration method for electrical interconnection systems based on branch and bound search information is as follows:

[0398] This section verifies the effectiveness of the invention through a test case combining an RTS-GMLC power system and a natural gas system. The power system has 73 nodes, 158 generating units (including 2 gas turbine units), and 120 lines; the natural gas system has 10 nodes, 2 gas sources, 3 pipelines, and 3 compressors; the scheduling period is 24 hours. The test equipment consists of an Intel i5-9300H CPU and 32GB of RAM. The MILP solver used is CPLEX 12.9, accessed via a Python interface. The solution termination condition is when the MILPgap reaches 0.01% or the solution time reaches 10,000 seconds.

[0399] The benchmark for comparing the algorithm's effectiveness is the direct solution using CPLEX with default parameter settings and 16 threads in parallel. The proposed method sets up a main process and child processes to call single-threaded computations, with the threshold for collecting branch-bound relaxation solutions set to a total number greater than N=300. Feasibility improvement parameters are selected sequentially as follows:

[0400] (k,Δ)∈{(100,0),(200,0),(300,0),(100,1),(200,1),(300,1)}(1)

[0401] 1. Validation of different numbers of segments

[0402] In the same electro-electric interconnected system, MILP models are constructed considering different numbers of piecewise linear segments. The solution efficiency of CPLEX and the proposed method is compared, and the results are as follows: Figure 2 As shown in the figure, numerical results demonstrate that the proposed method has a shorter solution time than CPLEX for all segments ranging from 15 to 20, verifying the effectiveness of the proposed method under different segment numbers. Specifically, the CPLEX solution time ranges from 1245.69 to 10000 seconds (the case with 19 segments was terminated by the termination condition), while the proposed method's solution time ranges from 270.98 to 1188.97 seconds. The speedup ratio (CPLEX solution time / proposed method solution time) is 1.60 to over 36.90 times (the speedup ratio for the case with 19 segments is greater than 36.90 times), with an average speedup of 11.28 times.

[0403] also, Figure 2 The results show that as the number of segments increases, directly calling CPLEX to solve the problem will lead to a sharp increase in computational burden. The proposed method can significantly alleviate this bottleneck, thereby improving the accuracy of piecewise linearization modeling and the scheduling decision efficiency of the electro-pneumatic interconnection system.

[0404] 2. Validation of effectiveness under different load conditions

[0405] Thirty scheduling decision models for the electrical-electrical interconnected system were constructed using different load curves and different numbers of segments. The solution efficiency of CPLEX and the proposed method was compared. Figure 3 As shown, the vertical axis is the exponential axis.

[0406] Numerical results show that the proposed method has a shorter solution time than CPLEX in different examples, verifying the effectiveness of the proposed method under different load conditions. Specifically, the solution time of CPLEX ranges from 216.38 to 5671.70 seconds, while the solution time of the proposed method ranges from 72.33 to 734.36 seconds, with a speedup of 1.64 to 13.52 times and an average speedup of 4.20 times.

[0407] In summary, this invention proposes an accelerated scheduling algorithm for electrical interconnection systems based on branch-and-bound search information. It utilizes information from the initial branch-and-bound search process to identify discrete decision variables that require optimization in subsequent stages, thereby constructing and solving a corresponding small-scale integer programming model. This model is then embedded into existing commercial solvers, guiding the branch-and-bound algorithm to prune more redundant search space and accelerating the convergence of electrical interconnection system scheduling solutions while ensuring optimality. Case studies demonstrate that this invention is more suitable for the computational needs of electrical interconnection system scheduling.

[0408] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0409] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0410] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0411] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0412] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0413] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for accelerating scheduling of an interconnected power system based on branch-and-bound search information, characterized by, include: Construct a scheduling model for an electrical-electric interconnected system; The scheduling model of the electrical-electric interconnection system is initially solved to obtain N relaxed solutions, and a search information dataset is constructed. The search information dataset is evaluated to obtain an upper bound for the optimal solution of the electrical-electric interconnection system scheduling model; Based on the upper bound of the optimal solution of the electric-gas interconnected system scheduling model, the electric-gas interconnected system scheduling model is subsequently solved to obtain the optimal solution of the electric-gas interconnected system scheduling model; Methods for solving the scheduling model of an electrical-electric interconnected system include the branch and bound algorithm; The elements of the search information dataset include sample inputs and corresponding labels; the sample inputs are the values ​​of the objective function of the relaxation solution. The label corresponding to this sample input is denoted as The mapping relationship between sample input and labels is as follows: ; wherein the tag as follows: （25） where j denotes the index of the segment variable ; The steps for evaluating a search information dataset include: Step 1) Calculate the distance d between all samples in the search information dataset and the test target; the test target is the current lower bound of the optimal solution; The distance d between the sample and the test target is shown below: （26）； where q belongs to the smallest k sample set; for the sample objective function Step 2) calculating the effective value range of the piecewise auxiliary discrete decision variable and ; Parameters As follows: （27）； In the formula, the parameters ; Step 3) Establishing a small-scale auxiliary MILP model based on the effective value range and ; Step 4) Solve the small-scale auxiliary MILP model to obtain the upper bound of the optimal solution.

2. The branch-and-bound search information based electrical interconnection system scheduling acceleration method of claim 1, wherein, The power-gas interconnection system scheduling model includes a power system scheduling model and a natural gas system scheduling model.

3. The branch-and-bound search information based electrical interconnection system scheduling acceleration method of claim 2, wherein, The objective function of the power system dispatch model is as follows: （1） In the formula, respectively represent a dispatch period set and a dispatch unit set; constants respectively represent an output cost coefficient and a startup cost coefficient of each unit; Continuous decision variables representing the output of each unit at each time period; Discrete decision variables a flag representing whether the unit g is started at the time period t; t represents any scheduling time period; g represents any scheduling unit; The constraints of the power system dispatch model include load balance constraints, line power flow constraints, logical constraints between discrete decision variables of generating units, generating unit capacity constraints, and generating unit ramping constraints. The load balancing constraints are as follows: （2） wherein denotes the set of system loads; d denotes an arbitrary system load; the constant denotes the load of each load node at each time period; The power flow constraints for the lines are as follows: （3） wherein denotes the set of network lines; b denotes an arbitrary network line; the constant denotes the maximum capacity of each line; the constant , the constant denotes the power transfer distribution factor; The logical constraints between the discrete decision variables of the unit are shown below: （4） （5） （6） In the formula, discrete decision variables These represent the operating status of unit g during time period t and whether it is shut down; constants. These represent the minimum start-up and shutdown times for each unit; This indicates the operating status of unit g during the time period t-1; Discrete decision variables This indicates whether unit g is started during time period j; This indicates whether unit g is shut down during time period j; The unit capacity constraints are as follows: （7） wherein the constant Pmin and Pmax represent the minimum and maximum power output of each unit, respectively. The climbing constraints are as follows: （8） （9） wherein the constant respectively represent the up / down ramp rate and the start / stop rate of each unit.

4. The branch-and-bound search information based electrical interconnection system scheduling acceleration method of claim 3, wherein, The objective function of the natural gas system scheduling model is shown below: （10） wherein, representing a gas source and a collection of pipes, respectively; s represents any gas source; representing the beginning and end of a pipe, respectively; constant representing a gas source cost coefficient and a pipe cost coefficient, respectively; Continuous decision variable Qs and Qp represent the flow rate of each gas source and the pipe storage of each pipe at each time interval, respectively. The constraints of the natural gas system scheduling model include gas source output constraints, compressor constraints, gas node constraints, and pipeline constraints. The gas source outlet constraints are as follows: （11） wherein the constants , the constants respectively represent the minimum and maximum air output of the air source; The compressor constraints include compressor flow rate upper limit constraints and compressor start and end pressure constraints. The compressor flow rate upper limit constraint is as follows: （12） In the formula, the set This refers to a compressor; the subscript 'v' indicates flow into the compressor, and the subscript 'w' indicates flow out of the compressor; constants. Represents the upper limit of compressor capacity; a continuous decision variable. This indicates the airflow rate through the compressor at different times. The compressor's starting and ending pressure constraints are shown below: （13） wherein the constant respectively represent lower and upper limits of the compression factor of each compressor; continuous decision variables respectively represent the start and end pressures of the compressor in each time period; The gas node constraints include node pressure constraints and node airflow balance constraints; The nodal pressure constraints are shown below: （14） wherein denotes a set of gas network nodes; denotes a set of gas network nodes; constant denotes the upper and lower pressure limit of each node, respectively; continuous decision variable denotes the pressure of each node at each time period; The nodal airflow balance constraints are shown below: （15） In the formula, Represent the set of gas turbine units and the set of gas loads, respectively; constants Representing the nodal gas load and compressor gas consumption coefficient, respectively; continuous decision variables. These represent the inflow and outflow airflow at the beginning and end of the pipe, respectively; continuous decision variables. This indicates the outflow rate of each gas source at different times; , These represent the air flow rates flowing into and out of the compressor, respectively. Gas consumption of gas turbine units As shown below: （16） wherein the constant represents the gas consumption coefficient of the gas turbine unit; Pipe constraints include pipe capacity constraints, pipe inventory constraints, and the Weymouth equation; Pipeline capacity constraints are shown below: （17） wherein the constant represents the maximum capacity of each pipe; represents the pipe capacity at time t; The storage constraints are as follows: （18） （19） wherein the constant represents the pipe storage coefficient of each pipe; respectively represents the pipe storage of each pipe at the t-1 period. The Weymouth equation is shown below: （20） wherein is a constant; The linearized Weymouth equation is shown below: （21） （22） （23） （24） wherein represents a linear function within each segment; represents the segment in which the independent variable lies; p represents the number of segments; auxiliary continuous decision variable represents the linearized independent variable value, auxiliary discrete decision variable; represents a piecewise linear function of the Weymouth equation.

5. The branch-and-bound search information based electrical interconnection system scheduling acceleration method of claim 1, wherein, The small-scale auxiliary MILP model is shown below: （28） （29）。 6. An electrical interconnection system dispatch acceleration system based on the method of any of claims 1-5, characterized in that, include: Model building unit: Constructing a scheduling model for an electrical-pneumatic interconnected system; Search information dataset construction unit: Solve the scheduling model of the electric-electric interconnection system to obtain N relaxed solutions, thereby constructing the search information dataset; Optimization Unit: Evaluates the search information dataset to obtain the upper bound of the optimal solution for the scheduling model of the electric-electric interconnected system; Solution Unit: Based on the upper bound of the optimal solution of the electric-gas interconnected system scheduling model, the electric-gas interconnected system scheduling model is solved to obtain the optimal solution of the electric-gas interconnected system scheduling model.

7. An electronic device, comprising: include: The device includes a memory, a processor, and a data bus, wherein the memory stores all machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the data bus, and the machine-readable instructions are executed by the processor to perform the method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the method described in any one of claims 1 to 5.

Images