Power grid two-stage robust optimization method considering wind power uncertainty

Abstract

The invention discloses a power grid two-stage robust optimization method considering wind power uncertainty, and the method comprises the steps: predicting wind power output and an average fluctuation value of wind power per hour according to the historical output data of a wind driven generator, and then carrying out the modeling of the wind power output uncertainty based on the prediction data of a fan, secondly, constructing a power grid two-stage robust optimization model, and adopting an improved Camp; and solving the model by a CG algorithm to obtain a target function value of the overall optimization model and an optimal scheduling scheme.

Description

Technical Field

[0001] This invention belongs to the field of power system technology, and more specifically, relates to a two-stage robust optimization method for power grids that considers the uncertainty of wind power. Background Technology

[0002] The long-term large-scale exploitation and use of fossil fuels by human society has triggered a series of serious problems, among which global warming is particularly significant. In the past decade, the international community has been pursuing energy reforms towards low-carbon emission reduction, with the transformation of electricity being a crucial area of ​​focus. Among various clean energy sources, wind power has lower costs, requires less land, and its technology is more mature in China, making it an important component of China's energy structure. With the continuous expansion of wind power generation, the scale of distributed wind power generation connected to the grid is also gradually increasing. Therefore, how to consider the randomness of wind power in the power system has become a key issue.

[0003] In the power system optimization scheduling problem considering wind power uncertainty, the key to solving the problem lies in simplifying the second-stage two-level optimization problem into a single level. Traditional solution methods mainly utilize duality theory, but these methods have some shortcomings in terms of solution accuracy, adaptability, and computational efficiency. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a two-stage robust optimization method for power grids that considers the uncertainty of wind power. First, the objective function and constraints of the first stage are established. Then, the optimization scheduling model of the second stage is established considering the randomness of wind power. Finally, the two are integrated into a compact overall optimization model to obtain the optimal scheduling strategy of the power grid.

[0005] To achieve the above-mentioned objective, this invention provides a two-stage robust optimization method for power grids considering wind power uncertainty, characterized by comprising the following steps:

[0006] (1) Collect the wind power output of each wind turbine at each sampling time at equal intervals, and reduce the wind power output data to obtain the predicted wind power output value and the average fluctuation value.

[0007] (2) Based on the predicted wind power output and the average fluctuation value, the uncertainty of wind power is modeled in MATLAB.

[0008] (3) Construct deterministic scheduling modeling in a two-stage robust optimization model of a power grid considering wind power uncertainty in MATLAB, namely the first-stage optimization model;

[0009] (4) Uncertainty scheduling modeling in a two-stage robust optimization model of a power grid considering wind power uncertainty in MATLAB, namely the second-stage optimization model;

[0010] (5) Integrate the first-stage optimization model and the second-stage optimization model into a compact overall optimization model, and then use the improved C&CG algorithm to solve the compact overall optimization model to obtain the objective function value and the optimal scheduling scheme of the overall optimization model.

[0011] The objective of this invention is achieved as follows:

[0012] This invention provides a two-stage robust optimization method for power grids that considers wind power uncertainty. First, wind power output and the average hourly fluctuation of wind power are predicted based on historical wind turbine output data. Then, the uncertainty of wind power output is modeled based on the predicted wind turbine data. Next, a two-stage robust optimization model for the power grid is constructed. The improved C&CG algorithm is used to solve the model to obtain the objective function value of the overall optimization model and the optimal scheduling scheme.

[0013] Meanwhile, the two-stage robust optimization method for power grid considering wind power uncertainty of the present invention also has the following beneficial effects:

[0014] (1) The two-stage robust optimization method of the present invention, with hierarchical modeling and unified optimization, can handle deterministic and uncertain variables more flexibly, so that the optimization model can take into account both robustness and flexibility, and solve the disadvantage of overly conservative solutions brought about by the robust optimization method.

[0015] (2) The bilinear term linear reconstruction method proposed in this invention has greater universality than traditional solution methods, and greatly improves computational efficiency and reduces the number of iterations;

[0016] (3) The two-stage robust optimization method of the power grid considering the uncertainty of wind power in this invention can promote wind power consumption and reduce wind power waste, while achieving better grid dispatch, lower grid operation cost and dispatch cost. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the node distribution of each wind turbine unit in this invention;

[0018] Figure 2 This is a flowchart of a two-stage robust optimization method for power grid considering wind power uncertainty according to the present invention.

[0019] Figure 3 The limit value Π of the wind turbine of this invention t =8, Π w Output diagrams of various types of generating units when the power output is equal to 2.

[0020] Figure 4 The limit value Π of the wind turbine of this invention t =8, Π w =2 Hourly standby power output diagram of thermal power units;

[0021] Figure 5 The limit value Π of the wind turbine of this invention t =8, Π w Solve the iteration count graph when = 2;

[0022] Figure 6 The limit value Π of the wind turbine of this invention t =8, Π w Output diagrams of various types of generating units when the power output is equal to 3.

[0023] Figure 7 The limit value Π of the wind turbine of this invention t =8, Π w =3 Calculation of standby power output of thermal power units (upper and lower standby output);

[0024] Figure 8 The limit value Π of the wind turbine of this invention t =8, Π w The number of iterations is plotted when the number of iterations is 3. Detailed Implementation

[0025] The specific embodiments of the present invention will now be described with reference to the accompanying drawings to enable those skilled in the art to better understand the invention. It should be particularly noted that in the following description, detailed descriptions of known functions and designs that might obscure the main content of the invention will be omitted here.

[0026] Example

[0027] In this embodiment, as Figure 1 As shown, in the 30-node example based on the MATLAB MATPOWER toolbox, the three added wind turbines are installed on nodes 3, 8, and 10 of the power system, while the thermal power units in the system are installed on nodes 1, 2, 13, 23, 27, and 22.

[0028] Below we combine Figure 1 This invention provides a detailed description of a two-stage robust optimization method for power grids that considers wind power uncertainty, as follows: Figure 2 As shown, the specific steps include:

[0029] (1) Collect the wind power output of each wind turbine at each sampling time at equal intervals, and reduce the wind power output data to obtain the predicted wind power output value and the average fluctuation value.

[0030] In this embodiment, the method for reducing wind power output data is as follows:

[0031] Let A be the wind power output of the w-th wind turbine at time t. w,t Where t = 1, 2, ..., N T w = 1, 2, ..., N W In this embodiment, N T =24 indicates the number of operating segments for the wind turbine, N W =3 indicates the number of wind turbines connected;

[0032] The traditional Latin hypersolution method is used to reduce the wind power output data to obtain the predicted wind power output of each wind turbine at time t. and the average output fluctuation of each wind turbine at time t.

[0033] (2) Based on the predicted wind power output and the average fluctuation value, the uncertainty of wind power is modeled in MATLAB.

[0034] In this embodiment, the method for modeling wind power uncertainty is as follows:

[0035] The actual wind power output of each wind turbine at each moment is represented as the superposition of the predicted wind power output and the fluctuation value of wind power. By restricting the number of turbines with fluctuations and the number of times fluctuations occur within the sampling time, the following wind power uncertainty model is obtained:

[0036]

[0037]

[0038] in, It is a 0 / 1 random integer variable used to characterize the fluctuations in the wind farm. A value of 0 indicates that the wind turbines have not fluctuated, and a value of 1 indicates that the wind turbines have fluctuated; Π t Π w These are the time uncertainty limits and the spatial uncertainty limits for wind power, respectively.

[0039] In this embodiment, Π is considered t =8, Π w =2 and Π t =8, Π w =3. Two possible solutions for the optimization model in this scenario.

[0040] (3) Construct deterministic scheduling modeling in a two-stage robust optimization model of a power grid considering wind power uncertainty in MATLAB, namely the first-stage optimization model;

[0041] In this embodiment, the process of establishing the first-stage optimization model is as follows:

[0042] (3.1) Establish the objective function of the first-stage optimization model;

[0043]

[0044] Where, N G P represents the number of thermal power units. g,t This is the output of the g-th thermal power unit at time t. and These represent the start-up cost and start-up / shutdown cost of a thermal power unit, respectively. g,t and v g,t These are 0-1 variables, representing the start-up and shutdown states of the thermal power unit, respectively. and These represent the upward and downward standby costs of thermal power units, respectively. and These represent the upward and downward reserve capacities of thermal power units, respectively, F. g (.) represents the operating cost function of a traditional thermal power unit, and its specific form is as follows:

[0045]

[0046] Among them, a 1,g a 2,g a 3,g These are the subsystem parameters of the thermal power unit, i g,t This is a 0 / 1 random integer variable used to represent the operating status of the thermal power unit. A value of 0 indicates that the unit is not started, and a value of 1 indicates that the unit is started.

[0047] In this embodiment, The units are all yuan / kW·h, a 1,g =[0.0211;0.07;0.079;0.0211;0.07;0.079],a 2,g =[21.05;23.9;21.63;21.05;23.9;21.63],a 3,g =[1313.6;471;480.29;1313.6;471;480.29], N G =6;

[0048] (3.2) Construct a power system balance model;

[0049]

[0050] Among them, G b and W bLet b represent the sets of thermal power units and wind farms connected to busbar b, where b = 1, 2, 3, ..., N. B l:θ1(l)=b and l:θ2(l)=b represent the power injection and power outflow from bus b, respectively; P w,t P represents the wind power absorption capacity of wind farm w at time t. l,t L represents the transmission power of line l at time t, with the direction indicated by the sign of the value. b,t This represents the load on bus b at time t;

[0051] (3.3) Add unit start-up and shutdown constraints and minimum start-up and shutdown time constraints;

[0052] Traditional thermal power unit start-stop constraints;

[0053]

[0054] Minimum start / stop time constraint;

[0055] i g,t =1, 1≤t≤U T,g

[0056]

[0057] i g,t =0,1≤t≤D T,g

[0058]

[0059] in, These are the initial start-up and shutdown time statistics for thermal power units, U T,g and D T,g These represent the minimum start-up and minimum shutdown times for thermal power units, respectively. and These are the minimum time periods during which thermal power units need to be continuously started and shut down, respectively.

[0060] In this embodiment, U T,g =2h,D T,g =2h,

[0061] (3.4) Power flow and capacity constraint model of the line;

[0062]

[0063] in, θ o2(l),t The voltage phase angles injected into and out of the transmission line l are x and y, respectively. l P is the reactance of transmission line l. l maxFor the maximum power flow of transmission line l, θ , θ represents the lower and upper limits of the phase angle of the voltage at bus b, respectively. b,t Let θ be the voltage phase angle of bus b at time t. ref,t Let t be the voltage phase angle of the balancing node at time t.

[0064] In this embodiment, θ = -180°, P l max =100kW;

[0065] (3.5) Modeling of output constraints for traditional thermal power units;

[0066]

[0067] in, These represent the maximum and minimum output limits of thermal power units, respectively.

[0068] In this embodiment, All units are in kW;

[0069] (3.6) Modeling of slope climbing constraints for thermal power units;

[0070]

[0071] Among them, R U,g R D,g These are the maximum limits for thermal power units climbing uphill and downhill, respectively.

[0072] In this embodiment, R U,g =[50,25,50,50,25,50], R D,g = [50,25,50,50,25,50], all in kW;

[0073] (3.7) Modeling of standby capacity constraints;

[0074]

[0075] in, These are the maximum upward and downward reserve capacities that a thermal power unit can provide, R. +,min R -,min These represent the minimum total upward and downward reserve capacity required by the power system, respectively.

[0076] In this embodiment, R -,min =2,R +,min =2, both units are kW;

[0077] (3.8) Wind farm output constraint modeling;

[0078]

[0079] in, The predicted wind power output of the wind turbine at time t.

[0080] (4) Uncertainty scheduling modeling in a two-stage robust optimization model of a power grid considering wind power uncertainty in MATLAB, namely the second-stage optimization model;

[0081] In this embodiment, the process of establishing the second-stage optimization model is as follows:

[0082] (4.1) Construct the objective function of the second-stage optimization model;

[0083]

[0084] in, Let A′ be the upward and downward reserve quantities called up by thermal power unit g at time t. w,t P′ represents the actual output value of the wind turbine w at time t in the second stage. w,t C represents the absorption capacity of wind turbine w at time t in the second stage. cur For the cost of wind curtailment penalties, ΔL b,t Let C be the emergency load shedding amount of bus b at time t. loss Penalty for load shedding;

[0085] In this embodiment, C cur =20, C loss =1000, all units are yuan / kW·h;

[0086] (4.2) Constructing power balance constraints for the second-stage optimized scheduling;

[0087]

[0088] Among them, P′ g,t , P′ w,t , P′ l,t These represent the output of thermal power unit g, the actual output of wind turbine w, and the transmission power of transmission line l at time t in the second stage, respectively.

[0089] (4.3) Constructing the second-stage line power flow and transmission capacity constraints;

[0090]

[0091] in, θ′ o2(l),tθ′ represents the voltage phase angles injected into and out of the bus of the second-stage transmission line l at time t. b,t Let θ′ be the voltage phase angle of bus b at time t in the second stage. ref,t The voltage phase angle at time t, the equilibrium node of the second stage;

[0092] (4.4) Constructing backup constraints for thermal power units;

[0093]

[0094] (4.5) Constructing wind power output constraints;

[0095]

[0096] (4.6) Construct emergency load shedding constraints;

[0097]

[0098] in, This represents the maximum emergency load shedding amount of bus b at time t.

[0099] In this embodiment,

[0100] (5) Integrate the first-stage optimization model with the second-stage optimization model into a compact overall optimization model:

[0101]

[0102] Q1x s +P1θ s =G1

[0103] Q2θ s ≤G2

[0104] Q3θ s ≤G3

[0105] Q4x s ≤G4

[0106] G5x s -T5Z≤G5

[0107] Q6x s ≤G6

[0108] B w Z≤Γ w

[0109] Where c,K,k are the coefficient matrices of the first-stage constraints and the objective function, and Q... i P1, G i T5, B w,Γ w Both are coefficient matrices, i = 1, 2, ..., 6; B s Z represents the cost of wind curtailment penalties, and C represents an uncertain variable. s It is the total penalty cost for wind curtailment, matrix matrix θ s For voltage phase angle variables, For the first stage control variables, including P g,t i g,t u g,t v g,t P w,t , L b,t P l,t θ b,t .

[0110] Then, the improved C&CG algorithm is used to solve the compact form of the global optimization model. The solution process is as follows:

[0111] (5.1) Initialization settings: Set the number of iterations to j = 1, and set the upper bound of the objective function of the overall optimization problem to U. B =+∞, with the lower bound set to L B =-∞, set the convergence error e=1e-5;

[0112] (5.2) Integrate the first-stage optimization model and the second-stage optimization model into a main problem-solving model:

[0113]

[0114] Q1x s +P1θ s =G1

[0115] Q2θ s ≤G2

[0116] Q3θ s ≤G3

[0117] Q4x s ≤G4

[0118] G5x s -T5Z≤G5

[0119] Q6x s ≤G6

[0120] B w Z≤Γ w

[0121] Where η is the objective function value to be optimized in the second stage, which is 0 in the first iteration. This represents the worst-case scenario value found in the second phase.

[0122] (5.3) Solve the main problem model to obtain the solution result, denoted as V′; then update the lower bound of the objective function in the overall optimization problem model to L. B =V′;

[0123] (5.4) Reconstruct the bilinear terms in the second-stage optimization model into a subproblem solution model:

[0124] (5.4.1) The second-stage optimization model is transformed into a compact form model as follows:

[0125]

[0126] stQ1x s +P1θ s =G1

[0127] Q2θ s ≤G2

[0128] Q3θ s ≤G3

[0129] Q4x s ≤G4

[0130] G5x s -T5Z≤G5

[0131] Q6x s ≤G6

[0132] B w Z≤Γ w

[0133] (5.4.2) Based on duality theory, the compact form model after the second stage transformation is transformed into a single-layer problem-solving model;

[0134]

[0135] stλ1 T Q1+λ4 T Q4+λ5 T Q5+λ6 T Q6≤A s

[0136] λ1 T P1+λ2 T Q2+λ3 T Q3≤0

[0137] λ i ≤0, i=1......6

[0138] B w Z≤Γ w

[0139] Where, λ i As dual variables;

[0140] (5.4.3) Separating the bilinear terms of the objective function in the single-layer problem-solving model yields the separated model:

[0141]

[0142] stλ1 T Q1+λ4 T Q4+λ5 T Q5+λ6 T Q6≤A s

[0143] λ1 T P1+λ2 T Q2+λ3 T Q3≤0

[0144] λ i ≤0, i=1......6

[0145] B w Z≤Γ w

[0146] (5.4.4) Transform the separated model into a dual problem model:

[0147]

[0148] λ7≥0

[0149] Where λ7 is the dual variable;

[0150] (5.4.5) The optimal solution set of the dual problem model in step (5.4.4) is equivalently represented as:

[0151]

[0152] Where u and λ8 are the equality constraint coefficients and the dual variables of the inequality constraints, respectively;

[0153] (5.4.6) Find the dual problem model again in step (5.4.4) to obtain a new dual problem model:

[0154]

[0155] stB w u+Γ w ≥0

[0156] (5.4.7) From the new dual problem model in step (5.4.6) and the model separated in step (5.4.3), we can draw the following conclusion:

[0157] u = -Z

[0158] (5.4.8) Obtain the subproblem model after linear reconstruction of the bilinear terms;

[0159] Substituting the optimal solution set of the dual problem model in step (5.4.5) and the conclusion obtained in step (5.4.7) into the model after separating the bilinear terms in step (5.4.3), we obtain the subproblem-solving model after linear reconstruction of the bilinear terms:

[0160]

[0161] stλ1 T Q1+λ4 T Q4+λ5 T Q5+λ6 T Q6≤A s

[0162] λ1 T P1+λ2 T Q2+λ3 T Q3≤0

[0163] λ i ≤0, i=1......6

[0164]

[0165] Γ w -B w Z-λ8=0

[0166] λ7,λ8≥0

[0167] (5.5) Solve the subproblem solution model after linear reconstruction of the bilinear terms to obtain the subproblem solution result V″ and the worst operating condition.

[0168] (5.6) Return the constraints in the subproblem-solving model after the bilinear term reconstruction to the main problem-solving model, and update the upper bound of the objective function of the overall optimization problem to...

[0169] (5.7) Convergence test: If |(U B -L B ) / L B If |≤e, then the problem has converged, and the iteration stops. The objective function value of the overall optimization model is U. BIf the optimal scheduling scheme is not found, continue iterating, j = j + 1, and return to step (5.2).

[0170] Ultimately, we can obtain the objective function value of the overall optimization model and the optimal scheduling scheme.

[0171] In this embodiment, the final optimized scheduling result is: within the wind turbine limit Π t =8, Π w When =2, the optimized scheduling results for various units are as follows: Figure 3 As shown, the optimal power output scheduling scheme for thermal power units and wind power units under the minimum operating cost in this scenario was obtained; the backup scheduling results for traditional thermal power units are as follows. Figure 4 As shown, the standby scheduling scheme for thermal power units in this scenario was obtained; the iterative curve is shown in the figure. Figure 5 As shown, convergence is achieved after the second iteration in this scenario; within the wind turbine limit Π t =8, Π w =3 Optimized scheduling results for various generating units are as follows Figure 6 As shown, the optimal power output scheduling scheme for thermal power units and wind power units under the minimum operating cost in this scenario was obtained; the backup scheduling results for traditional thermal power units are as follows. Figure 7 As shown, the standby scheduling scheme for thermal power units in this scenario was obtained; the iterative curve is shown in the figure. Figure 8 As shown, convergence is achieved after the second iteration in this scenario.

[0172] To verify the superiority of the proposed method, we compared it with the traditional C&CG algorithm. Table 1 shows the running results under different scenarios, and Table 2 shows the running results of the wind turbine.

[0173] Table 1 Results of running the program in different scenarios

[0174]

[0175]

[0176] Table 2 Fan Operation Results

[0177]

[0178] The improved C&CG solution method yields a better scheduling scheme, lower operating costs, lower backup costs, faster operating efficiency, and significantly increases the wind power absorption capacity of the grid. Traditional methods converge after approximately 5-6 iterations, while the improved method converges after approximately 2-3 iterations.

[0179] Although the illustrative specific embodiments of the present invention have been described above to enable those skilled in the art to understand the invention, it should be understood that the invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the invention as defined and determined by the appended claims, and all inventions utilizing the concept of the present invention are protected.

Claims

1. A two-stage robust optimization method for power grids considering wind power uncertainty, characterized in that, Includes the following steps: (1) Collect the wind power output of each wind turbine at each sampling time at equal intervals, and reduce the wind power output data to obtain the predicted wind power output value and the average fluctuation value. (2) Based on the predicted wind power output and the average fluctuation value, the uncertainty of wind power is modeled in MATLAB. (3) Construct deterministic scheduling modeling in a two-stage robust optimization model of a power grid considering wind power uncertainty in MATLAB, namely the first-stage optimization model; (4) Uncertainty scheduling modeling in a two-stage robust optimization model of a power grid considering wind power uncertainty in MATLAB, namely the second-stage optimization model; (5) Integrate the first-stage optimization model and the second-stage optimization model into a compact overall optimization model, and then use the improved C&CG algorithm to solve the compact overall optimization model to obtain the objective function value and the optimal scheduling scheme of the overall optimization model.

2. The two-stage robust optimization method for power grid considering wind power uncertainty according to claim 1, characterized in that, The method for reducing the wind power output data is as follows: Let A be the wind power output of the w-th wind turbine at time t. w,t Where t = 1, 2, ..., N T w = 1, 2, ..., N W N T N represents the number of operating segments connected to the wind turbine. W Indicates the number of wind turbines connected; The traditional Latin hypersolution method is used to reduce the wind power output data to obtain the predicted wind power output of each wind turbine at time t. and the average output fluctuation of each wind turbine at time t.

3. The two-stage robust optimization method for power grid considering wind power uncertainty according to claim 1, characterized in that, The method for modeling the uncertainty of wind power is as follows: The actual wind power output of each wind turbine at each moment is represented as the superposition of the predicted wind power output and the fluctuation value of wind power. By restricting the number of turbines with fluctuations and the number of times fluctuations occur within the sampling time, the following wind power uncertainty model is obtained: in, It is a 0 / 1 random integer variable used to characterize the fluctuations in the wind farm. A value of 0 indicates that the wind turbines have not fluctuated, and a value of 1 indicates that the wind turbines have fluctuated; Π t Π w These are the time uncertainty limits and the spatial uncertainty limits for wind power, respectively.

4. The two-stage robust optimization method for power grid considering wind power uncertainty according to claim 1, characterized in that, The process of establishing the first-stage optimization model is as follows: (4.1) Establish the objective function of the first-stage optimization model; Where, N G P represents the number of thermal power units. g,t This is the output of the g-th thermal power unit at time t. and These represent the start-up cost and start-up / shutdown cost of a thermal power unit, respectively. g,t and v g,t These are 0-1 variables, representing the start-up and shutdown states of the thermal power unit, respectively. and These represent the upward and downward standby costs of thermal power units, respectively. and These represent the upward and downward reserve capacities of thermal power units, respectively, F. g (.) represents the operating cost function of a traditional thermal power unit, and its specific form is as follows: Among them, a 1,g a 2,g a 3,g These are the subsystem parameters of the thermal power unit, i g,t This is a 0 / 1 random integer variable used to represent the operating status of the thermal power unit. A value of 0 indicates that the unit is not started, and a value of 1 indicates that the unit is started. (4.2) Construct a power system balance model; Among them, G b and W b Let b represent the sets of thermal power units and wind farms connected to busbar b, where b = 1, 2, 3, ..., N. B l:θ1(l)=b and l:θ2(l)=b represent the power injection and power outflow from bus b, respectively; P w,t P represents the wind power absorption capacity of wind farm w at time t. l,t L represents the transmission power of line l at time t, with the direction indicated by the sign of the value. b,t This represents the load on bus b at time t; (4.3) Add unit start-up and shutdown constraints and minimum start-up and shutdown time constraints; Traditional thermal power unit start-stop constraints; Minimum start / stop time constraint; i g,t =1,1≤t≤U T,g i g,t =0,1≤t≤D T,g in, These are the initial start-up and shutdown time statistics for thermal power units, U T,g and D T,g These represent the minimum start-up and minimum shutdown times for thermal power units, respectively. and These are the minimum time periods during which thermal power units need to be continuously started and shut down, respectively. (4.4) Power flow and capacity constraint model of the line; Where, θ o1(l),t θ o2(l),t These are the voltage phase angles injected into and out of the transmission line l, respectively, x l P is the reactance of transmission line l. l max For the maximum power flow of transmission line l, θ , θ represents the lower and upper limits of the phase angle of the voltage at bus b, respectively. b,t Let θ be the voltage phase angle of bus b at time t. ref,t Let be the voltage phase angle at the equilibrium node at time t; (4.5) Modeling of output constraints for traditional thermal power units; in, These represent the maximum and minimum output limits of thermal power units, respectively. (4.6) Modeling of slope climbing constraints for thermal power units; Among them, R U,g R D,g These are the maximum limits for thermal power units climbing upwards and downwards, respectively. (4.7) Modeling of standby capacity constraints; in, These are the maximum upward and downward reserve capacities that a thermal power unit can provide, R. +,min R +,min These are the minimum total upward and downward reserve capacity required by the power system, respectively. (4.8) Wind farm output constraint modeling; in, The predicted wind power output of the wind turbine at time t.

5. A two-stage robust optimization method for power grid considering wind power uncertainty according to claim 1, characterized in that, The process of establishing the second-stage optimization model is as follows: (5.1) Construct the objective function of the second-stage optimization model; in, Let A′ be the upward and downward reserve quantities called up by thermal power unit g at time t. w,t P′ represents the actual output value of the wind turbine w at time t in the second stage. w,t C represents the absorption capacity of wind turbine w at time t in the second stage. cur For the cost of wind curtailment penalties, ΔL b,t Let C be the emergency load shedding amount of bus b at time t. loss Penalty for load shedding; (5.2) Constructing power balance constraints for the second-stage optimized scheduling; Among them, P′ g,t , P′ w,t , P′ l,t These represent the output of thermal power unit g, the actual output of wind turbine w, and the transmission power of transmission line l at time t in the second stage, respectively. (5.3) Constructing the second-stage line power flow and transmission capacity constraints; in, θ′ o2(l),t θ′ represents the voltage phase angles injected into and out of the bus of the second-stage transmission line l at time t. b,t Let θ′ be the voltage phase angle of bus b at time t in the second stage. ref,t The voltage phase angle at time t, the equilibrium node of the second stage; (5.4) Constructing backup constraints for thermal power units; (5.5) Constructing wind power output constraints; (5.6) Construct emergency load shedding constraints; in, This represents the maximum emergency load shedding amount of bus b at time t.

6. The two-stage robust optimization method for power grid considering wind power uncertainty according to claim 1, characterized in that, The first-stage optimization model and the second-stage optimization model are integrated into a compact overall optimization model as follows: Q1x s +P1θ s =G1 Q2θ s ≤G2 Q3θ s ≤G3 Q4x s ≤G4 G5x s -T5Z≤G5 Q6x s ≤G6 B w Z≤Γ w Where c,K,k are the coefficient matrices of the first-stage constraints and the objective function, and Q... i P1, G i T5, B w ,Γ w Both are coefficient matrices, i = 1, 2, ..., 6; B s Z represents the cost of wind curtailment penalties, and C represents an uncertain variable. s It is the total penalty cost for wind curtailment, matrix matrix θ s For voltage phase angle variables, For the first stage control variables, including P g,t i g,t u g,t v g,t P w,t , L b,t P l,t θ b,t .

7. A two-stage robust optimization method for power grid considering wind power uncertainty according to claim 1, characterized in that, The process of solving the overall optimization model using the improved C&CG algorithm is as follows: (7.1) Initialization settings: Set the number of iterations to j = 1, and set the upper bound of the objective function of the overall optimization problem to U. B =+∞, with the lower bound set to L B =-∞, set the convergence error e; (7.2) Integrate the first-stage optimization model and the second-stage optimization model into a main problem-solving model: Q1x s +P1θ s =G1 Q2θ s ≤G2 Q3θ s ≤G3 Q4x s ≤G4 G5x s -T5Z≤G5 Q6x s ≤G6 B w Z≤Γ w Where η is the objective function value to be optimized in the second stage, which is 0 in the first iteration. This represents the worst-case scenario value found in the second phase. (7.3) Solve the main problem model to obtain the main problem solution result, denoted as V′; then update the lower bound of the objective function in the overall optimization problem solution model to L. B =V′; (7.4) Reconstruct the bilinear terms in the second-stage optimization model into a subproblem solution model: (7.4.1) The second-stage optimization model is transformed into a compact form model as follows: s.t.Q1x s +P1θ s =G1 Q2θ s ≤G2 Q3θ s ≤G3 Q4x s ≤G4 G5x s -T5Z≤G5 Q6x s ≤G6 B w Z≤Γ w (7.4.2) Based on duality theory, the compact form model after the second stage transformation is transformed into a single-layer problem-solving model; stλ1 T Q1+λ4 T Q4+λ5 T Q5+λ6 T Q6≤A s λ1 T P1+λ2 T Q2+λ3 T Q3≤0 l i ≤0,i=1......6 B w Z≤Γ w Where, λ i As dual variables; (7.4.3) Separating the bilinear terms of the objective function in the single-layer problem-solving model yields the separated model: stλ1 T Q1+λ4 T Q4+λ5 T Q5+λ6 T Q6≤A s λ1 T P1+λ2 T Q2+λ3 T Q3≤0 l i ≤0,i=1......6 B w Z≤Γ w (7.4.4) Transform the separated model into a dual problem model: λ7≥0 Where λ7 is the dual variable; (7.4.5) The optimal solution set of the dual problem model in step (7.4.4) is equivalently represented as: Where u and λ8 are the equality constraint coefficients and the dual variables of the inequality constraints, respectively; (7.4.6) Find the dual problem model again in step (7.4.4) to obtain a new dual problem model: s.t.B w u+Γ w ≥0 (7.4.7) From the new dual problem model in step (7.4.6) and the model separated in step (7.4.3), we can draw the following conclusion: u = -Z (7.4.8) Obtain the subproblem model after linear reconstruction of the bilinear terms; Substituting the optimal solution set of the dual problem model in step (7.4.5) and the conclusion obtained in step (7.4.7) into the model after separating the bilinear terms in step (7.4.3), we obtain the subproblem-solving model after linear reconstruction of the bilinear terms: stλ1 T Q1+λ4 T Q4+λ5 T Q5+λ6 T Q6≤A s λ1 T P1+λ2 T Q2+λ3 T Q3≤0 l i ≤0,i=1......6 C w -B w Z-λ8=0 λ7,λ8≥0 (7.5) Solve the subproblem solution model after linear reconstruction of the bilinear terms to obtain the subproblem solution result V″ and the worst operating condition. (7.6) Return the constraints in the subproblem-solving model after the bilinear term reconstruction to the main problem-solving model, and update the upper bound of the objective function of the overall optimization problem to... (7.7) Convergence test: If |(U B -L B ) / L B If |≤e, then the problem has converged, and the iteration stops. The objective function value of the overall optimization model is U. B If the optimal scheduling scheme is not found, continue iterating, j = j + 1, and return to step (7.2).

Images