A robust economic dispatch method for distribution-level electric-gas integrated energy system

Abstract

The application discloses to belong to the technical field of integrated energy system operation optimization, and particularly relates to a robust economic dispatching method for an electric-gas integrated energy system at a distribution network level. In a pre-dispatching stage, a robust economic dispatching model is constructed based on wind power prediction output. In a re-dispatching stage, an extreme wind power output scenario is identified, the feasibility of the robust economic dispatching model to a wind power uncertainty set is verified, and an optimized robust economic dispatching model is obtained. A second-order cone relaxation method is used to simplify the nonlinear constraint in the Distflow model. A piecewise linearization method is used to simplify the nonlinear constraint in the approximate dynamic model. The optimized robust economic dispatching model is solved by using a nested C&CG algorithm, and a robust dispatching strategy is obtained. It is proved by calculation examples that the inner and outer loops can converge in a small number of times under the condition that the accuracy is guaranteed, and the convergence of the solving algorithm is good.

Description

Technical Field

[0001] This invention belongs to the field of integrated energy system operation optimization technology, and specifically relates to a robust economic dispatch method for a distribution network-level integrated electricity-gas energy system. Background Technology

[0002] In recent years, with the large-scale deployment of gas turbine units and the rapid development of power-to-gas technology, especially on the distribution network side, the coupling between the power system and the natural gas system has gradually increased. Meanwhile, due to the large-scale integration of renewable energy sources such as wind power into the grid, wind power penetration in some areas has reached 100%. Because wind power output is highly volatile, the optimal scheduling and safe operation of the distribution network-level integrated power-gas energy system face challenges; therefore, developing flexible and reliable economic dispatch strategies is essential.

[0003] Generally, solving the two-stage robust scheduling problem relies on the strong duality theorem and the C&CG algorithm (column-and constraint generation), requiring the constructed model to be a linear convex model. However, both power systems and natural gas systems are non-convex nonlinear models. Given the structural and parametric characteristics of distribution networks, the Distflow model, as an AC power flow model, is more suitable for characterizing the operating characteristics of distribution networks compared to DC power flow models. However, in robust optimization studies of integrated power and gas distribution network systems, the Distflow model is often simplified or its nonlinear constraints are ignored to improve the model's solvability, leading to accuracy issues.

[0004] For natural gas system models, the accuracy of their pipeline equations also suffers. As a medium-pressure gas network, the Polyflo equations more accurately characterize the gas flow characteristics of the distribution network. However, current research has largely neglected practical application scenarios. Algorithmically, the use of piecewise linearization to handle nonlinear pipeline equations introduces auxiliary integer variables into the model, preventing the second-stage max-min problem from being equivalently transformed into a single-layer problem using the strong duality theorem.

[0005] Therefore, there is an urgent need for a robust economic dispatch method for distribution network-level integrated power-gas energy systems to solve the problem of solving the two-stage robust dispatch problem in general for non-convex nonlinear models of power systems and natural gas systems, and to achieve good convergence while ensuring accuracy. Summary of the Invention

[0006] The purpose of this invention is to provide a robust and economical dispatch method for a distribution network-level integrated electricity-gas energy system, characterized by the following steps:

[0007] In the pre-schedule phase, based on the predicted wind power output, and considering the constraints of the distribution network operation, the gas distribution network operation, and the constraints of the coupling components, a robust economic dispatch model is constructed.

[0008] In the robust economic dispatch model, the distribution network section adopts the Distflow model, while the gas distribution network section adopts an approximate dynamic model that considers pipeline inventory.

[0009] During the rescheduling phase, considering the deviation between the actual and predicted wind power output, the robust economic dispatch model is verified for the feasibility of the robust economic dispatch model for the set of wind power uncertainties by identifying extreme wind power output scenarios, and an optimized robust economic dispatch model is obtained.

[0010] The nonlinear constraints in the Distflow model are simplified using a second-order cone relaxation method; the nonlinear constraints in the approximate dynamic model are simplified using a piecewise linearization method; thus enabling the optimized robust economic scheduling model to be solved using a commercial solver.

[0011] A nested C&CG algorithm is used to solve the optimized robust economic scheduling model, thereby obtaining a robust scheduling strategy. The robust economic scheduling model is as follows:

[0012] Objective function for the pre-scheduling phase:

[0013]

[0014]

[0015] In the formula, t and This represents the optimization period and its set; and These represent the power generation cost of conventional generating units and the gas production cost of gas wells, respectively. and c s These represent the cost coefficients for conventional generator sets and gas wells, respectively. and f st Representing time period t and conventional unit g respectively n Active power and gas well production;

[0016] The distribution network operation constraints include at least one of the following: node power balance constraints, unit active and reactive power output constraints, positive and negative ramping constraints for conventional units and gas turbine units, line current carrying capacity constraints, node voltage constraints, no backflow of active and reactive power flow, line voltage drop constraints, and constraints on the relationship between line current, voltage and power.

[0017] The gas distribution network operation constraints include at least one of the following: node flow balance constraints, natural gas source production constraints, natural gas system node pressure constraints, node gas pressure and pipeline inventory relationship constraints, pipeline gas inventory conservation constraints, natural gas pipeline average flow, and pipeline equation constraints.

[0018] The constraints of the coupling element include: the operating constraints of the gas turbine;

[0019] The set of decision variables in the min problem during the pre-scheduling phase is:

[0020]

[0021] in, Indicates time period t, conventional unit g n Those who have made meritorious contributions; Indicates time period t, conventional unit g n Unproductive efforts; Indicates time period t, gas turbine unit g g Those who have made meritorious contributions; Indicates time period t, gas turbine unit g g Unproductive output; p lt q represents the active power flow through line l during time period t; lt This represents the reactive power flow through line l during time period t; v et i represents the square of the voltage at node e during time period t; et f represents the square of the current flowing through line e during time period t; st This represents the gas production of well s during time period t; This represents the inflow / outflow rate of pipe p during time period t; π represents the average flow rate of pipe p during time period t; gt Indicates the pressure at node g of the natural gas system during time period t; m pt This indicates the amount of gas stored in pipe p during time period t.

[0022] The optimized robust economic scheduling model is as follows:

[0023] The objective function for the rescheduling phase is:

[0024]

[0025]

[0026] in, Indicates conventional unit g n / Adjustment cost of gas wells; C pw / C pd This indicates the penalty cost for wind curtailment / load shedding; Indicates conventional unit g n Adjust cost coefficients upwards / downwards; This indicates that the cost coefficient for gas well s can be adjusted upwards or downwards. Indicates time period t, conventional unit g n Adjust the power up / down; Indicates the upward / downward adjustment of gas production rate by gas well s during time period t; δ w / δ d Indicates the wind curtailment / load shedding penalty factor; Δp wt / Δp dt This represents the active power generated during time period t (wind curtailment / load shedding).

[0027] The logical constraint for uncertain wind power output is:

[0028]

[0029]

[0030] In the formula, P wt This represents the predicted wind power output. P represents wt Variables in the real-time phase, This indicates the upper / lower boundaries of the predicted wind power output. Setting 1 indicates the upper boundary of the predicted range for actual wind power output. Setting it to 0 indicates the predicted value within the predicted range of actual wind power output. Setting 1 indicates the lower boundary of the predicted range for actual wind power output. Setting it to 0 indicates the predicted value within the predicted range of actual wind power output;

[0031] The constraint on the number of times different wind turbines at the same moment obtain the wind power output boundary value is:

[0032]

[0033] The constraint on the number of times the wind power output boundary value is obtained at different times for the same wind power is:

[0034]

[0035] In the formula, Δ W Indicates a spatially uncertain budget, Δ T Indicates an uncertain timeframe for the budget;

[0036] The decision variables for the max problem in the rescheduling phase are:

[0037]

[0038] The decision variables for the min problem in the rescheduling phase are:

[0039]

[0040] Where, Δp wt Δp represents the amount of wind curtailment during time period t. dt / Δq dt Indicates active / reactive load shedding. Indicates time period t, conventional unit g n Adjusting reactive power upwards / downwards Indicates time period t, conventional unit g n Adjust the power up / down; Indicates time period t, gas turbine unit g g Adjusting reactive power upwards / downwards Indicates time period t, gas turbine unit g g Adjust the power up / down; This indicates the upward / downward adjustment of gas production from well s during time period t; for p lt ,q lt ,v et i et ,P wt , m pt Variables in the real-time phase.

[0041] The simplification of the nonlinear constraints in the Distflow model using the second-order cone relaxation method includes:

[0042] The branch flow constraints of the Distflow model in the distribution network are relaxed to a standard second-order cone constraint, as shown in Equation (31):

[0043]

[0044] In the formula, p lt q represents the active power flow through line l during time period t; lt This represents the reactive power flow through line l during time period t; Indicates the voltage at the beginning / end of line l; i lt The square of the current flowing through line l during time period t is expressed as t.

[0045] The simplification of nonlinear constraints in the approximate dynamic model using a piecewise linearization method includes:

[0046] Using two substitution variables F pt and Π pt Nonlinear terms in the Polyflo equation and By substitution, we obtain equation (32).

[0047]

[0048] In the formula, F pt , as well as ψ represents a substitution variable used to replace nonlinear terms in the Polyflo equation. p This represents the constant coefficients in the pipeline equation.

[0049] The steps for solving the optimized robust economic scheduling model using the nested C&CG algorithm include:

[0050] S41: Use the inner C&CG algorithm to solve the two-layer max-min problem and obtain the extreme wind power output scenario;

[0051] S42: Add feasible cut constraints corresponding to extreme wind power output scenarios to the pre-scheduling master problem, and use the outer C&CG algorithm to solve the pre-scheduling master problem.

[0052] The specific steps of the inner C&CG algorithm are as follows:

[0053] S411: Set the iteration count n = 1, the upper bound UB = +∞, the lower bound LB = -∞, and let the initial wind power uncertainty vector v n =0;

[0054] S412: Based on the uncertainty vector v n Solving equation (33) yields the optimal solution y for the decision variables in the rescheduling phase. n The optimal solution z for integer variables n :

[0055]

[0056] Where λ represents the dual variable of the linear constraint, δ,ω represent the dual variable of the second-order cone constraint, x represents the decision variable of the pre-schedule stage, y / z represent the continuous / integer variables in the decision variable of the reschedule stage, v represents the wind power uncertainty vector, and B,K,C,m,G are coefficients derived from equations (4)-(22), F lt h lt The coefficient is derived from equation (14), β T The coefficients are derived from equations (2)-(3);

[0057] S413: Update the lower bound LB = max{LB,β} T y n}, fix the resulting integer variable z n By transforming the min problem into its dual form, we can obtain:

[0058]

[0059] Where χ represents the objective function variable of the rescheduling phase;

[0060] S414: Obtain the optimal solution v * ,χ * Update the upper bound UB = min{UB,χ *};

[0061] S415: Determine if the algorithm has converged. If |UB-LB|≤ε, the algorithm has converged and terminates. Otherwise, let n = n+1 and return to S412.

[0062] The specific steps of the outer C&CG algorithm are as follows:

[0063] S421: Set the iteration count k = 0, the upper bound UB = +∞, and the lower bound LB = -∞;

[0064] S422: Solve equation (36) to obtain the solution x for the pre-scheduling decision variable. k ,η * Update the lower bound LB = max{LB, α T x * +η *};

[0065]

[0066] Where, α T The coefficients are derived from equations (2)-(3), where x and η are the pre-scheduling decision variables;

[0067] S423: Enter the inner C&CG loop to obtain the optimal solution v * ,χ * And update the upper bound UB = min{UB,α} T x k +χ *};

[0068] S424: Determine if the algorithm has converged. If |UB-LB|≤ξ, terminate the algorithm; otherwise, create variable y. k+1 and z k+1 And add the following constraint to equation (37);

[0069]

[0070] Among them, B,K,C,m are derived from equations (4)-(22), and F lt h lt It is derived from equation (14) that y k+1 and z k+1 β is an intermediate variable. T The coefficients are derived from equations (2)-(3);

[0071] Let k = k + 1, and return to S422.

[0072] A robust economic dispatch device for a distribution network-level integrated power-gas energy system, characterized in that it comprises:

[0073] Data acquisition module: Collects the actual value of wind power output;

[0074] The scheduling and control module: In the pre-scheduling phase, a robust economic scheduling model is constructed based on the predicted wind power output; in the rescheduling phase, extreme wind power output scenarios are identified, and the feasibility of the robust economic scheduling model for the set of wind power uncertainties is verified to obtain an optimized robust economic scheduling model; the second-order cone relaxation method is used to simplify the nonlinear constraints in the Distflow model; the piecewise linearization method is used to simplify the nonlinear constraints in the approximate dynamic model; and the nested C&CG algorithm is used to solve the optimized robust economic scheduling model to obtain a robust scheduling strategy.

[0075] The steps of the scheduling control module to simplify the nonlinear constraints in the approximate dynamic model using a piecewise linearization method include:

[0076] Using two substitution variables F pt and Π pt Nonlinear terms in the Polyflo equation and By substitution, we obtain equation (32).

[0077]

[0078] In the formula, F pt , as well as ψ represents a substitution variable used to replace nonlinear terms in the Polyflo equation. p This represents the constant coefficients in the pipeline equation.

[0079] The steps by which the scheduling control module solves the optimized robust economic scheduling model using a nested C&CG algorithm include:

[0080] S41: Use the inner C&CG algorithm to solve the two-layer max-min problem and obtain the extreme wind power output scenario;

[0081] S42: Add feasible cut constraints corresponding to extreme wind power output scenarios to the pre-scheduling master problem, and use the outer C&CG algorithm to solve the pre-scheduling master problem.

[0082] The beneficial effects of this invention are as follows:

[0083] This invention discloses a robust economic dispatch method for a distribution network-level integrated power-gas energy system. The model consists of two stages: pre-dispatch and re-dispatch. For the nonlinear constraints in the Distflow model and the natural gas system model, a second-order cone relaxation method and a piecewise linearization method are used to handle them, respectively. The C&CG algorithm is employed to solve the problem, thereby obtaining extreme wind power output scenarios. This invention solves the problem of solving the two-stage robust dispatch problem in general for non-convex nonlinear models such as power systems and natural gas systems. Numerical examples demonstrate that, while ensuring accuracy, the algorithm converges within a relatively small number of iterations for both inner and outer loops, exhibiting good convergence. Attached Figure Description

[0084] Figure 1 This is a flowchart illustrating a robust economic dispatch method for a distribution network-level integrated power-gas energy system according to the present invention.

[0085] Figure 2 This is a flowchart illustrating the nested C&CG algorithm of the present invention.

[0086] Figure 3 This is a topology diagram of a distribution network-level integrated electricity-gas energy system consisting of a 13-node distribution network and a 6-node gas distribution network verified by the present invention.

[0087] Figure 4 This is a schematic diagram of the daily load power and wind power prediction power studied in this invention;

[0088] Figure 5 This is a schematic diagram of the scene distribution under different scheduling strategies adopted in this invention;

[0089] Figure 6 This is a schematic diagram of the convergence process of the nested C&CG algorithm used in this invention. Detailed Implementation

[0090] This invention provides a robust and economical dispatch method for a distribution network-level integrated power-gas energy system. The invention will be further described in detail below with reference to the accompanying drawings.

[0091] The robust economic dispatch problem of the distribution network-level integrated power-gas energy system focuses on two stages: pre-dispatch and re-dispatch, aiming to ensure that the system meets economic dispatch requirements while minimizing the risks caused by wind power fluctuations. This problem can be expressed as solving a min-max-min optimization problem as shown in equation (1). The objective of the min problem is to minimize the operating cost DA of the entire system under pre-dispatch, while the objective of the inner max-min problem is to minimize the re-adjustment cost RT under the worst-case scenario.

[0092]

[0093] In the formula, x *and These are the scheduling strategies for the pre-scheduling and rescheduling phases, respectively. W and W represent the uncertain variables and their set during the rescheduling phase, respectively.

[0094] like Figure 1 The embodiment of the present invention disclosed presents a robust economic dispatch method for a distribution network-level integrated power-gas energy system. The specific steps of the method are as follows:

[0095] In the pre-schedule phase, based on the predicted wind power output, and considering the constraints of the distribution network, gas distribution network, and coupling components, a robust economic dispatch model is constructed.

[0096] In the robust economic dispatch model, the distribution network section adopts the Distflow model, while the gas distribution network section adopts an approximate dynamic model that considers pipeline inventory.

[0097] The objective of the pre-scheduling phase is to minimize the system operating cost DA. The constraints include distribution network operation constraints, gas distribution network constraints considering dynamic pipeline storage, and operation constraints with gas turbines as the coupling unit.

[0098] The robust economic scheduling model is as follows:

[0099] Objective function for the pre-scheduling phase:

[0100]

[0101]

[0102] In the formula, t and This represents the optimization period and its set; and These represent the power generation cost of conventional generating units and the gas production cost of gas wells, respectively. and c s These represent the cost coefficients for conventional generator sets and gas wells, respectively. and f st Representing time period t and conventional unit g respectively n Active power and gas well production;

[0103] Distribution network operation constraints include at least one of the following:

[0104] 1) Node power balance constraints:

[0105]

[0106]

[0107] Equations (4) and (5) represent the active and reactive power balance constraints, respectively; where, Indicates time period t, conventional unit gn Those who have made meritorious contributions; Indicates time period t, gas turbine unit g g The meritorious contribution; P wt p represents the predicted wind power output for time period t. lt P represents the active power flow through line l during time period t; dt Indicates the active load during time period t; i lt V represents the square of the current flowing through line l during time period t; et G represents the square of the voltage at node e during time period t; e Represents the equivalent conductance of node e; r l This represents the equivalent resistance of line l; This represents the set of conventional generator units / gas turbine units / wind power / electric loads connected to node e; This represents the set of beginning and end points of the lines connected to node e. Indicates time period t, conventional unit g n Unproductive efforts; Indicates time period t, gas turbine unit g g Unproductive output; q lt Q represents the reactive power flow through line l during time period t; dt B represents the reactive load during time period t; e x represents the equivalent susceptance of node e; l This represents the equivalent reactance of line l;

[0108] 2) Constraints on active and reactive power output of the generator unit:

[0109]

[0110]

[0111] In the formula, Indicates conventional unit g n / Gas turbine unit g g Upper limit of active power; Indicates conventional unit g n / Gas turbine unit g g Lower limit of active power; Indicates conventional unit g n / Gas turbine unit g g Reactive power limit; Indicates conventional unit g n / Gas turbine unit g g Lower limit of reactive power;

[0112] 3) Positive and negative ramp constraints for conventional and gas turbine units:

[0113]

[0114]

[0115] In the formula, Indicates conventional unit g n Positive / negative ramp limits; Indicates gas turbine unit g g Positive / negative ramp limits;

[0116] 4) Line current carrying capacity constraints:

[0117]

[0118] In the formula, Indicates the maximum current carrying capacity of the power line;

[0119] 5) Node voltage constraints:

[0120]

[0121] In the formula, Indicates the upper / lower limit of the node voltage;

[0122] 6) No return flow for both active and reactive currents:

[0123]

[0124] In the formula, q lt This represents the reactive power flow through line l during time period t;

[0125] 7) Line voltage drop constraints:

[0126]

[0127] In the formula, This indicates the voltage at the beginning and end of line l;

[0128] 8) Constraints on the relationship between line current, voltage, and power:

[0129]

[0130] In the formula, This indicates the voltage at the beginning and end of line l;

[0131] Gas distribution network operation constraints include at least one of the following:

[0132] 1) Node traffic balancing constraints:

[0133]

[0134] In the formula, f st This represents the gas flow rate from well s during time period t. This represents the inflow / outflow rate of pipe p during time period t. Indicates natural gas load. Indicates time period t, gas turbine g g Gas consumption, This represents the set of gas wells and gas loads connected to node g. This represents the set of beginning / end lines connected to node g. This represents the set of gas turbine units connected to node g;

[0135] 2) Natural gas source production constraints:

[0136]

[0137] In the formula, Indicates the upper / lower limit of gas production from gas well s;

[0138] 3) Pressure constraints at natural gas system nodes:

[0139]

[0140] In the formula, π gt This represents the pressure at node g of the natural gas system during time period t; Indicates the upper / lower limit of the pressure at node g;

[0141] 4) Constraints on the relationship between node gas pressure and pipe inventory:

[0142]

[0143] In the formula, m pt φ represents the gas storage in pipeline p during time period t; p Indicates the gas storage capacity coefficient in the pipeline; This represents the gas pressure at both ends of the natural gas pipeline p during time period t.

[0144] 5) Pipeline gas storage conservation constraint:

[0145]

[0146] 6) Average flow rate of natural gas pipelines

[0147]

[0148] In the formula, This represents the average flow rate of natural gas in pipeline p at time t;

[0149] 7) Pipeline equation constraints

[0150]

[0151] Equation (21) is the Polyflo equation describing the relationship between pipeline flow rate and pressure at both ends of the pipeline. Its power exponent for the flow rate term is 1.848. It is mainly used in gas distribution networks with pressures ranging from 0.75 to 7 bar. p The constant coefficients of the pipeline equation are represented by sgn(), which represents the sign function. hour, Take 1, when hour, Take -1;

[0152] Coupling element constraints include:

[0153] Operating constraints of gas turbines;

[0154]

[0155] In the formula, K represents the calorific value of natural gas; Indicates gas turbine g g The gas-to-electric conversion efficiency;

[0156] The set of decision variables in the min problem during the pre-scheduling phase is:

[0157]

[0158] in, Indicates time period t, conventional unit g n Those who have made meritorious contributions; Indicates time period t, conventional unit g n Unproductive efforts; Indicates time period t, gas turbine unit g g Those who have made meritorious contributions; Indicates time period t, gas turbine unit g g Unproductive output; p lt q represents the active power flow through line l during time period t; lt This represents the reactive power flow through line l during time period t; v et i represents the square of the voltage at node e during time period t; et f represents the square of the current flowing through line e during time period t; st This represents the gas production of well s during time period t; This represents the inflow / outflow rate of pipe p during time period t; π represents the average flow rate of pipe p during time period t; gt Indicates the pressure at node g of the natural gas system during time period t; m pt This indicates the amount of gas stored in pipe p during time period t.

[0159] The power distribution network adopted the Distflow model, which includes power balance constraints, unit output ramp-up constraints, line current carrying capacity constraints, node pressure constraints, non-return power constraints, line voltage drop constraints, and branch power flow constraints. The natural gas system adopted an approximate dynamic model that considers pipeline storage, and used the Polyflo equation to characterize the gas flow characteristics in the pipeline, including gas well production constraints, node flow balance constraints, pipeline gas storage constraints, pipeline gas storage conservation constraints, and pipeline equation constraints.

[0160] During the rescheduling phase, considering the deviation between the actual and predicted wind power output, the robust economic dispatch model is verified for the feasibility of the robust economic dispatch model for the set of wind power uncertainties by identifying extreme wind power output scenarios, and an optimized robust economic dispatch model is obtained.

[0161] During the rescheduling phase, a deviation occurs between the actual and predicted wind power output. Consequently, adjustments will be made to conventional turbines, gas turbines, and gas wells. Simultaneously, the coupled system may experience wind curtailment and load shedding. The adjustment costs for conventional turbines and gas wells, as well as the wind curtailment and load shedding penalty costs of the coupled system, are calculated in the objective function. Here, the adjustment cost of gas wells is considered to include the adjustment cost of the gas turbines. The objective of the rescheduling phase is to minimize the rescheduling cost RT when facing extreme wind power output scenarios. The objective functions are shown in equations (24) and (25).

[0162] The optimized robust economic scheduling model is as follows:

[0163] The objective function for the rescheduling phase is:

[0164]

[0165]

[0166] in, Indicates conventional unit g n / Adjustment cost of gas wells; C pw / C pd This indicates the penalty cost for wind curtailment / load shedding; Indicates conventional unit g n Adjust the cost coefficient upwards / downwards; This indicates that the cost coefficient for gas well s can be adjusted upwards or downwards. Indicates time period t, conventional unit g n Adjust the power up / down; Indicates the upward / downward adjustment of gas production rate by gas well s during time period t; δ w / δ d Indicates the wind curtailment / load shedding penalty factor; Δp wt / Δp dt This represents the active power generated during time period t (wind curtailment / load shedding).

[0167] For wind power uncertainty, a box-type uncertainty set is used for modeling. Equation (26) represents the relationship between the actual and predicted wind power output, and Equation (27) represents the logical constraints of wind power output uncertainty, i.e. They cannot both be 1.

[0168] The logical constraint for uncertain wind power output is:

[0169]

[0170]

[0171] In the formula, P wt This represents the predicted wind power output. P represents wt Variables in the real-time phase, This indicates the upper / lower boundaries of the predicted wind power output. Setting 1 indicates the upper boundary of the predicted range for actual wind power output. Setting it to 0 indicates the predicted value within the predicted range of actual wind power output. Setting 1 indicates the lower boundary of the predicted range for actual wind power output. A value of 0 represents the predicted value within the predicted range of actual wind power output.

[0172] According to the central limit theorem, if the probability of the same parameter reaching the boundary value of wind power output under all time periods and different parameters at the same time is very small, the scheduling result will be too conservative if the uncertain parameter is not restricted. An uncertain budget parameter is introduced to limit the number of times the uncertain wind power output reaches the boundary, thereby adjusting the conservatism of the scheduling scheme, as expressed by equations (28) and (29).

[0173] The constraint on the number of times different wind turbines at the same moment obtain the wind power output boundary value is:

[0174]

[0175] The constraint on the number of times the wind power output boundary value is obtained at different times for the same wind power is:

[0176]

[0177] Equations (28) and (29) respectively constrain the number of times the wind power output boundary value is obtained for different wind power sources at the same time and for the same wind power source at different times. Δ W Indicates a spatially uncertain budget, Δ T Indicates an uncertain timeframe for the budget;

[0178] Uncertain budgets define the maximum number of times an uncertain variable can reach its boundary within a scheduling cycle. Taking a time-uncertain budget as an example, when Δ...T When Δ = 0, the wind power during the rescheduling phase can only be taken as the predicted value. T When the value is 12, it means that the wind power output can be any value between the upper and lower boundaries during any 12 time periods, while the wind power output can only be the predicted value during the remaining 12 time periods. It can be seen that as the uncertain budget increases, the conservatism of the dispatching scheme increases, and system decision-makers can adjust the conservatism of the decision-making scheme by adjusting the uncertain budget.

[0179] The rescheduling phase model is a two-level max-min problem, where,

[0180] The decision variables for the max problem in the rescheduling phase are:

[0181]

[0182] The decision variables for the min problem in the rescheduling phase are:

[0183]

[0184] Where, Δp wt Δp represents the amount of wind curtailment during time period t. dt / Δq dt Indicates active / reactive load shedding. Indicates time period t, conventional unit g n Adjusting reactive power upwards / downwards Indicates time period t, conventional unit g n Adjust the power up / down; Indicates time period t, gas turbine unit g g Adjusting reactive power upwards / downwards Indicates time period t, gas turbine unit g g Adjust the power up / down; This indicates the upward / downward adjustment of gas production from well s during time period t; for p lt ,q lt ,v et i et ,P wt , m pt Variables in the real-time phase.

[0185] The nonlinear constraints in the Distflow model are simplified using a second-order cone relaxation method; the nonlinear constraints in the approximate dynamic model are simplified using a piecewise linearization method; thus enabling the optimized robust economic scheduling model to be solved using a commercial solver.

[0186] In this step, the robust economic dispatch model is a two-stage robust optimization problem. However, due to the existence of nonlinear constraints, the model cannot be solved directly by commercial solvers such as Gurobi. The following is a method for handling the power flow constraints of distribution network branches and the Polyflo equation.

[0187] 1) Methods for handling power flow constraints in distribution network branches.

[0188] The simplification of the nonlinear constraints in the Distflow model using the second-order cone relaxation method includes:

[0189] The power flow constraint of the distribution network branch can be relaxed to a standard second-order cone constraint, as shown in equation (31).

[0190]

[0191] In the formula, p lt q represents the active power flow through line l during time period t; lt This represents the reactive power flow through line l during time period t; Indicates the voltage at the beginning / end of line l; i lt The square of the current flowing through line l during time period t is expressed as t.

[0192] 2) Methods for handling Polyflo equations in pipeline systems.

[0193] The simplification of nonlinear constraints in the approximate dynamic model using a piecewise linearization method includes:

[0194] For the pipe equations, a piecewise linearization method was adopted, using two substitution variables F. pt and Π pt Nonlinear terms in the Polyflo equation and By substitution, we obtain equation (32).

[0195]

[0196] In the formula, F pt , as well as ψ represents a substitution variable used to replace nonlinear terms in the Polyflo equation. p This represents the constant coefficients in the pipeline equation.

[0197] The nested C&CG algorithm is used to solve the optimized robust economic scheduling model and obtain a robust scheduling strategy. Since the max-min problem in the rescheduling phase introduces auxiliary integer variables, it cannot be transformed into a single-layer problem using the strong duality theorem. Therefore, the C&CG algorithm is used to solve it, thereby obtaining extreme wind power output scenarios.

[0198] Generally, two-stage robust optimization models can be solved using the strong duality theorem and the C&CG algorithm. However, due to the introduction of integer variables in the piecewise linearization algorithm, the max-min problem in the rescheduling stage cannot be transformed using the strong duality theorem or KKT conditions, and the entire model cannot be directly solved using the C&CG algorithm. In fact, the max-min problem with integer variables can be solved by iteratively fixing the integer variables. In the rescheduling problem, the identification of the wind power uncertainty vector requires fixing the integer variables, and solving for the integer variables requires parameterizing the wind power uncertainty vector. Thus, the identification of the wind power uncertainty vector and the solution for the integer variables can be viewed as a two-stage decision-making process, for which the C&CG algorithm is introduced. Therefore, the entire model is solved by nesting the C&CG algorithm, specifically including:

[0199] S41: Using the inner C&CG algorithm, solve the two-layer max-min problem to obtain extreme wind power output scenarios; this part is the rescheduling stage, which identifies extreme wind power output scenarios by solving the two-layer max-min problem.

[0200] S42: Add feasible cut constraints corresponding to extreme wind power output scenarios to the pre-scheduling master problem, and use the outer C&CG algorithm to solve the pre-scheduling master problem.

[0201] The solution process for the inner and outer C&CG algorithms is given below.

[0202] In S41, the specific steps of the inner C&CG algorithm are as follows:

[0203] S411: Set the iteration count n = 1, the upper bound UB = +∞, the lower bound LB = -∞, and let the initial wind power uncertainty vector v n =0;

[0204] S412: Based on the uncertainty vector v n Solving equation (33) yields the optimal solution y for the decision variables in the rescheduling phase. n The optimal solution z for integer variables n :

[0205]

[0206] Where λ represents the dual variable of the linear constraint, δ,ω represent the dual variable of the second-order cone constraint, x represents the decision variable of the pre-schedule stage, y / z represent the continuous / integer variables in the decision variable of the reschedule stage, v represents the wind power uncertainty vector, and B,K,C,m,G are coefficients derived from equations (4)-(22), F lt h lt The coefficient is derived from equation (14), βT The coefficients are derived from equations (2)-(3);

[0207] S413: Update the lower bound LB = max{LB,β} T y n}, fix the resulting integer variable z n By transforming the min problem into its dual form, we can obtain:

[0208]

[0209] Where χ represents the objective function variable of the rescheduling phase;

[0210] S414: Obtain the optimal solution v * ,χ * Update the upper bound UB = min{UB,χ *};

[0211] S415: Determine if the algorithm has converged. If |UB-LB|≤ε, the algorithm has converged and is terminated; otherwise, let n = n+1 and return to S412.

[0212] The outer loop mainly targets the pre-scheduling master problem, solves the robust scheduling strategy, and adds the feasible cut constraints corresponding to the extreme wind power output scenarios identified by the inner C&CG algorithm to the pre-scheduling master problem.

[0213] In S42, the specific steps of the outer C&CG algorithm are as follows:

[0214] S421: Set the iteration count k = 0, the upper bound UB = +∞, and the lower bound LB = -∞;

[0215] S422: Solve equation (36) to obtain the solution x for the pre-scheduling decision variable. k ,η * Update the lower bound LB = max{LB, α T x * +η *};

[0216]

[0217] Where, α T The coefficients are derived from equations (2)-(3), where x and η are the pre-scheduling decision variables;

[0218] S423: Enter the inner C&CG loop to obtain the optimal solution v * ,χ * And update the upper bound UB = min{UB,α} T x k +χ *};

[0219] S424: Determine if the algorithm has converged. If |UB-LB|≤ξ, terminate the algorithm; otherwise, create variable y. k+1 and z k+1 And add the following constraint to equation (37);

[0220]

[0221] Among them, B,K,C,m are derived from equations (4)-(22), and F lt h lt It is derived from equation (14) that y k+1 and z k+1 β is an intermediate variable. T The coefficients are derived from equations (2)-(3);

[0222] Let k = k + 1, and return to S422.

[0223] The solution algorithm of this invention is a nested C&CG algorithm, which can solve the two-stage robust economic scheduling problem. The entire algorithm flow is as follows: Figure 2 As shown.

[0224] The following section verifies the proposed robust economic dispatch method for a distribution network-level integrated power-gas energy system. To test the robustness of the economic dispatch strategy for the distribution network-level integrated power-gas energy system, as well as the accuracy and computational efficiency of the algorithm, the following steps are taken: Figure 3 As shown, a coupled system of a 13-node distribution network and a 6-node gas distribution network was used for testing. The test platform was Matlab, and the optimization problem was solved using the Yalmip toolbox and Gurobi 8.0.1.

[0225] Table 1 Generator Parameter Table

[0226]

[0227] Table 2 Gas Well Parameters

[0228]

[0229] The power distribution network includes one conventional unit G1, two gas turbine units G2 and G3 (detailed parameters are shown in Table 1), gas turbines G2 and G3 are connected to natural gas system nodes 1 and 3 respectively, one wind power unit W1, and eight electrical loads PL1-PL8; the gas distribution network includes two gas sources GW1 and GW2 (detailed parameters are shown in Table 2), five natural gas pipelines, and three gas loads GL1-GL2.

[0230] The active power load curve of the distribution network, the predicted output curve of wind power and its upper and lower prediction boundaries, and the load curve of the gas distribution network are as follows: Figure 4As shown, the trend of electricity load changes is consistent with the electricity consumption habits of residents and industrial parks. The upper and lower boundaries of the wind power forecast curve are ±10% of the predicted output.

[0231] Robustness refers to the ability to withstand risks. A robust scheduling strategy means that when wind power output fluctuates, the strategy can meet the feasibility requirements of uncertain wind power sets, thus reducing the risk of system losses in adverse scenarios. To verify the robustness of the proposed scheduling strategy, a comparative test was conducted between deterministic optimization scheduling results and robust scheduling results.

[0232] In wind power scenario design, to best reflect actual conditions, 1000 scenarios are randomly generated within a ±10% fluctuation range based on predicted wind power output. By solving the rescheduling subproblems corresponding to these 1000 scenarios, the rescheduling cost of the system under these scenarios can be obtained, including the adjustment costs of conventional units and gas wells, and the penalty cost for wind curtailment and load shedding. Pre-scheduling cost and rescheduling cost constitute the total cost of the entire economic dispatch. By statistically analyzing the number of scenarios distributed within different cost ranges corresponding to these two dispatch strategies, the scenario distribution under the robust dispatch strategy and the deterministic dispatch strategy can be obtained, such as... Figure 5 As shown.

[0233] It can be seen that under the deterministic scheduling strategy, the scenario distribution is relatively dispersed, while under the robust scheduling strategy, the scenario distribution is more concentrated. In 1000 scenarios, the total scheduling cost of the robust scheduling strategy does not exceed $825.8, with no scenarios exhibiting excessively high total scheduling costs. This effectively avoids the risks that might arise with the deterministic scheduling strategy, demonstrating the robust scheduling strategy's ability to withstand risks. However, in 1000 scenarios, the total cost of robust scheduling is greater than $819.6, reflecting the conservative nature of the robust scheduling strategy. In scenarios where the actual wind power output is not significantly different from the predicted output, the rescheduling cost is relatively low. In this case, due to the higher pre-scheduling cost of the robust scheduling strategy, the total cost of the robust scheduling strategy will also be greater than the total cost of the deterministic scheduling strategy.

[0234] For the piecewise linearization method, the algorithm accuracy is greatly affected by the number of segments. Here, we conducted tests for different numbers of segments, and the specific results are shown in Table 3.

[0235] The table shows the iteration count, scheduling strategy cost, solution time of each subproblem, and relative error of the inner and outer C&CG algorithms for 20, 50, 70, 100, and 200 segments. The entire economic scheduling process can be divided into three subproblems: solving the pre-scheduling strategy, solving for integer variables, and identifying extreme scenarios. It can be seen that as the number of segments increases, the solution time of each subproblem also increases. This is because with the increase in the number of segments, the number of integer variables and constraints also increases, thus increasing the overall solution cost of the algorithm. Therefore, the solution time of each subproblem, including the total solution time of the entire algorithm, also increases continuously. However, the increasing number of segments does not change the iteration count of the inner C&CG algorithm. The inner C&CG algorithm mainly iterates between the two subproblems of solving for integer variables and identifying extreme scenarios, continuously fixing the integer variables to solve the lower-level max problem and obtain the extreme wind power output scenario. Therefore, the number of iterations of the inner algorithm remains unchanged.

[0236] Table 3 Algorithm Accuracy and Computational Efficiency

[0237]

[0238]

[0239] In general, solving two-stage robust optimization problems, especially the second-stage max-min problem, mainly involves transforming the bi-level problem into a single-level problem using the strong duality theorem. However, in this invention, due to the use of piecewise linearization in the second stage, integer variables are introduced. During the solution of extreme scenarios, the wind power uncertainty vector is parameterized, and the results of these integer variables also affect the obtained extreme scenarios. Thus, solving for integer variables and identifying extreme power output scenarios constitute a two-stage decision-making process. The introduced inner-layer C&CG algorithm effectively solves this problem, so the number of inner-layer iterations does not change with the increase in the number of segments. Regarding the scheduling cost, it can be seen that as the number of segments increases, the scheduling cost gradually stabilizes, eventually settling at $823.4. This indicates that increasing the number of segments improves the algorithm's accuracy and makes the optimization objective more stable. Although increasing the number of segments increases the burden on the overall model solution and lengthens the solution time, the solution accuracy also improves. The relative error can be obtained by comparing the percentage of unbalance on both sides of the pipeline equation. It can be seen that as the number of segments increases, the relative error of the algorithm continuously decreases. When the number of segments reaches 200, the relative error is as low as 0.52%. However, when the number of segments increases to a certain value, the improvement in model accuracy is not significant. In this invention, it can be seen that when the number of segments reaches 100, the error is already as low as 0.85%, which basically meets the requirements. At the same time, the computational cost is within an acceptable range. Therefore, in other tests, a number of segments of 100 was used for testing. Table 3 can also provide a reference for system decision-makers in balancing accuracy and efficiency, and the results are of reference value.

[0240] like Figure 6 As shown, the convergence process of the entire algorithm with 100 segments includes the convergence process of the outer C&CG algorithm and the convergence process of the inner C&CG algorithm under each outer loop. It can be seen that the outer loop converges after the 4th iteration. Under each outer loop, the inner loop converges twice, demonstrating the effectiveness of the inner C&CG algorithm for solving bi-level problems with integer variables. For the outer C&CG loop, the convergence time is 4, ensuring algorithm accuracy, indicating the good convergence of the nested C&CG algorithm.

[0241] Embodiments of the present invention include a robust economic dispatch device for a distribution network-level integrated power-gas energy system, comprising:

[0242] Data acquisition module: Collects the actual value of wind power output;

[0243] The scheduling and control module: In the pre-scheduling phase, a robust economic scheduling model is constructed based on the predicted wind power output; in the rescheduling phase, extreme wind power output scenarios are identified, and the feasibility of the robust economic scheduling model for the set of wind power uncertainties is verified to obtain an optimized robust economic scheduling model; the second-order cone relaxation method is used to simplify the nonlinear constraints in the Distflow model; the piecewise linearization method is used to simplify the nonlinear constraints in the approximate dynamic model; and the nested C&CG algorithm is used to solve the optimized robust economic scheduling model to obtain a robust scheduling strategy.

[0244] The steps of the scheduling control module to simplify the nonlinear constraints in the approximate dynamic model using a piecewise linearization method include:

[0245] Using two substitution variables F pt and Π pt Nonlinear terms in the Polyflo equation and By substitution, we obtain equation (32).

[0246]

[0247] In the formula, F pt , as well as ψ represents a substitution variable used to replace nonlinear terms in the Polyflo equation. p This represents the constant coefficients in the pipeline equation.

[0248] The steps by which the scheduling control module solves the optimized robust economic scheduling model using a nested C&CG algorithm include:

[0249] S41: Use the inner C&CG algorithm to solve the two-layer max-min problem and obtain the extreme wind power output scenario;

[0250] S42: Add feasible cut constraints corresponding to extreme wind power output scenarios to the pre-scheduling master problem, and use the outer C&CG algorithm to solve the pre-scheduling master problem.

[0251] Embodiments of the present invention include an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement various steps in the robust economic dispatch method for a distribution network-level integrated power-gas energy system as described in the present invention.

[0252] Embodiments of the present invention also include a storage medium storing a computer program thereon, which, when executed by a processor, implements the various steps of the robust economic dispatch method for the distribution network-level integrated power-gas energy system as described in the present invention.

[0253] In summary, the two-stage robust optimization model could originally be solved using the strong duality theorem and the C&CG algorithm. However, due to the introduction of integer variables in the piecewise linearization algorithm, the max-min problem in the rescheduling stage cannot be transformed using the strong duality theorem or KKT conditions, making the entire model unsolvable using the C&CG algorithm. This invention discloses a robust economic dispatch method for a distribution network-level integrated power-gas energy system, which solves the max-min problem with integer variables by iteratively fixing these integer variables. In the rescheduling problem, identifying the wind power uncertainty vector requires fixing the integer variables, and solving for the integer variables requires parameterizing the wind power uncertainty vector. Thus, the identification of the wind power uncertainty vector and the solution for the integer variables can be viewed as a two-stage decision-making process, for which the C&CG algorithm is introduced. In this way, the entire model is solved by nesting C&CG algorithms, which solves the problem of solving the two-stage robust scheduling problem for non-convex nonlinear models such as power systems and natural gas systems under general conditions. Furthermore, the numerical examples show that the algorithm can converge within a small number of iterations while ensuring accuracy, demonstrating the good convergence of the solution algorithm.

[0254] 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.

[0255] 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.

[0256] 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.

[0257] 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.

[0258] 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.

[0259] 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 robust economic dispatch method for a distribution network-level integrated power-gas energy system, characterized in that, Includes the following steps: In the pre-schedule phase, based on the predicted wind power output, and considering the constraints of the distribution network operation, the gas distribution network operation, and the constraints of the coupling components, a robust economic dispatch model is constructed. In the robust economic dispatch model, the distribution network section adopts the Distflow model, while the gas distribution network section adopts an approximate dynamic model that considers pipeline inventory. During the rescheduling phase, considering the deviation between the actual and predicted wind power output, the robust economic dispatch model is verified for the feasibility of the robust economic dispatch model for the set of wind power uncertainties by identifying extreme wind power output scenarios, and an optimized robust economic dispatch model is obtained. The nonlinear constraints in the Distflow model are simplified using a second-order cone relaxation method; the nonlinear constraints in the approximate dynamic model are simplified using a piecewise linearization method; thus enabling the optimized robust economic scheduling model to be solved using a commercial solver. The nested C&CG algorithm is used to solve the optimized robust economic scheduling model and obtain a robust scheduling strategy. The robust economic scheduling model is as follows: Objective function for the pre-scheduling phase: ， ， In the formula, t and This represents the optimization period and its set; and These represent the power generation cost of conventional generating units and the gas production cost of gas wells, respectively. and These represent the cost coefficients for conventional generator sets and gas wells, respectively. and These represent time periods t for conventional units. Active power and gas wells Gas production; The distribution network operation constraints include at least one of the following: node power balance constraints, unit active and reactive power output constraints, positive and negative ramping constraints for conventional units and gas turbine units, line current carrying capacity constraints, node voltage constraints, no backflow of active and reactive power flow, line voltage drop constraints, and constraints on the relationship between line current, voltage and power. The gas distribution network operation constraints include at least one of the following: node flow balance constraints, natural gas source production constraints, natural gas system node pressure constraints, node gas pressure and pipeline inventory relationship constraints, pipeline gas inventory conservation constraints, natural gas pipeline average flow, and pipeline equation constraints. The constraints of the coupling element include: the operating constraints of the gas turbine; The set of decision variables in the min problem during the pre-scheduling phase is: ， in, Indicates time period t for conventional units Those who have made meritorious contributions; Indicates time period t for conventional units Unproductive effort; Indicates time period t for gas turbine units Those who have made meritorious contributions; Indicates time period t for gas turbine units Unproductive efforts; This represents the active power flow through line l during time period t; This represents the reactive power flow through line l during time period t; This represents the square of the voltage at node e during time period t; This represents the square of the current flowing through line e during time period t; Indicates time period t gas well Gas production; This represents the inflow / outflow rate of pipe p during time period t; This represents the average flow rate of pipe p during time period t; Indicates the natural gas system node during time period t The pressure; This represents the amount of gas stored in pipeline p during time period t. The simplification of the nonlinear constraints in the Distflow model using the second-order cone relaxation method includes: The branch power flow constraints of the Distflow model in the distribution network section are relaxed to a standard second-order cone constraint, as shown in equation [equation missing]. As shown: ， In the formula, This represents the active power flow through line l during time period t; This represents the reactive power flow through line l during time period t; This represents the voltage at the first node of line l; This represents the square of the current flowing through line l during time period t; The simplification of nonlinear constraints in the approximate dynamic model using a piecewise linearization method includes: Using two substitution variables and Nonlinear terms in the Polyflo equation and By substitution, we obtain the formula , ， In the formula, , This represents a substitution variable used to replace nonlinear terms in the Polyflo equation; This represents the constant coefficients in the pipeline equation.

2. The robust economic dispatch method for a distribution network-level integrated power-gas energy system according to claim 1, characterized in that, The optimized robust economic scheduling model is as follows: The objective function for the rescheduling phase is: ， ， in, Indicates conventional units / gas well Adjustment costs; This indicates the penalty cost for wind curtailment / load shedding; Indicates conventional units Adjust cost coefficients upwards / downwards; Indicates gas well Adjust cost coefficients upwards / downwards; Indicates time period t for conventional units Adjust the power up / down; Indicates time period t gas well Adjust the gas production rate upwards / downwards; Indicates the wind curtailment / load shedding penalty factor; This represents the active power generated during time period t (wind curtailment / load shedding). The logical constraint for uncertain wind power output is: ， ， In the formula, This represents the predicted wind power output. express Variables in the real-time phase, This indicates the upper / lower boundaries of the predicted wind power output. Setting 1 indicates the upper boundary of the predicted range for actual wind power output. Setting it to 0 indicates the predicted value within the predicted range of actual wind power output. Setting 1 indicates the lower boundary of the predicted range for actual wind power output. Setting it to 0 indicates the predicted value within the predicted range of actual wind power output; The constraint on the number of times different wind turbines at the same moment obtain the wind power output boundary value is: ， The constraint on the number of times the wind power output boundary value is obtained at different times for the same wind power is: ， In the formula, Indicates an uncertain budget for space. Indicates an uncertain timeframe for the budget; The decision variables for the max problem in the rescheduling phase are: ， The decision variables for the min problem in the rescheduling phase are: ， in, This represents the amount of wind curtailed during time period t. Indicates active / reactive load shedding. Indicates time period t for conventional units Adjusting reactive power upwards / downwards Indicates time period t for conventional units Adjust the power up / down; Indicates time period t for gas turbine units Adjusting reactive power upwards / downwards Indicates time period t for gas turbine units Adjust the power up / down; Indicates time period t gas well Adjust the gas production rate upwards / downwards; for Variables in the real-time phase.

3. The robust economic dispatch method for a distribution network-level integrated power-gas energy system according to claim 1, characterized in that, The steps for solving the optimized robust economic scheduling model using the nested C&CG algorithm include: S41: Use the inner C&CG algorithm to solve the two-layer max-min problem and obtain the extreme wind power output scenario; S42: Add feasible cut constraints corresponding to extreme wind power output scenarios to the pre-scheduling master problem, and use the outer C&CG algorithm to solve the pre-scheduling master problem.

4. The robust economic dispatch method for a distribution network-level integrated power-gas energy system according to claim 3, characterized in that, The specific steps of the inner C&CG algorithm are as follows: S411: Set the number of iterations Upper Realm The lower realm And let the initial wind power uncertainty vector ; S412: Based on the uncertainty vector Solving equation The optimal solution for the decision variables in the rescheduling phase is obtained. Optimal solution for integer variables : ， in, Represents the dual variable of the linear constraint. Denotes the dual variables of a second-order cone constraint. This represents the decision variables in the pre-scheduling phase. This represents the continuous / integer variables in the decision variables during the rescheduling phase. This represents the uncertainty vector of wind power. , , , , The coefficients are derived from equations (4)-(22). , The coefficient is derived from equation (14). The coefficients are derived from equations (2)-(3); S413: Update the Nether Fixed integer variable By transforming the min problem into its dual form, we can obtain: ， in, To represent the objective function variables in the rescheduling phase; S414: Obtain the optimal solution , Update the upper boundary ; S415: Determine if convergence has occurred. If the algorithm converges, the algorithm terminates; otherwise, let n = n+1 and return to S412. The specific steps of the outer C&CG algorithm are as follows: S421: Set the number of iterations Upper Realm The lower realm ; S422: Solution formula The solution for the pre-scheduling decision variables is obtained. , Update the Nether ; ， in, The coefficients are derived from equations (2)-(3). Pre-scheduling decision variables; S423: Enter the inner C&CG loop to obtain the optimal solution. , and update the upper bound. ; S424: Determine if the algorithm has converged. If the condition is met, the algorithm terminates; otherwise, a variable is created. and And add the following constraints to the equation ; ， in, , , , It is derived from equations (4)-(22). , It is derived from equation (14). and As an intermediate variable, The coefficients are derived from equations (2)-(3); make And return to S422.

5. A robust economic dispatching device for a distribution network-level integrated power and gas energy system, based on the robust economic dispatching method for any one of claims 1-4, characterized in that... include: Data acquisition module: Collects the actual value of wind power output; Dispatch and control module: In the pre-dispatch phase, a robust economic dispatch model is constructed based on the predicted wind power output; In the rescheduling phase, extreme wind power output scenarios are identified, and the feasibility of the robust economic dispatch model for the set of wind power uncertainties is verified to obtain an optimized robust economic dispatch model. The second-order cone relaxation method is used to simplify the nonlinear constraints in the Distflow model. The piecewise linearization method is used to simplify the nonlinear constraints in the approximate dynamic model. The nested C&CG algorithm is used to solve the optimized robust economic dispatch model to obtain the robust dispatch strategy.

6. A robust economic dispatch device for a distribution network-level integrated power-gas energy system according to claim 5, characterized in that, The steps of the scheduling control module to simplify the nonlinear constraints in the approximate dynamic model using a piecewise linearization method include: Using two substitution variables and Nonlinear terms in the Polyflo equation and By substitution, we obtain equation (32). （32） In the formula, , This represents a substitution variable used to replace nonlinear terms in the Polyflo equation; This represents the constant coefficients in the pipeline equation.

7. A robust economic dispatch device for a distribution network-level integrated power-gas energy system according to claim 5, characterized in that, The steps by which the scheduling control module solves the optimized robust economic scheduling model using a nested C&CG algorithm include: S41: Use the inner C&CG algorithm to solve the two-layer max-min problem and obtain the extreme wind power output scenario; S42: Add feasible cut constraints corresponding to extreme wind power output scenarios to the pre-scheduling master problem, and use the outer C&CG algorithm to solve the pre-scheduling master problem.

Images