Two-layer robust unit commitment method and component for power system

Abstract

The application provides a double-layer robust unit commitment method and assembly for a power system, obtains an output scene set of a new energy unit and a unit commitment value of a thermal power unit based on an inner-layer two-stage robust unit commitment model; according to the output scene set of the new energy unit of the inner-layer model, the feasibility of the unit commitment value of the thermal power unit is verified based on an outer-layer multi-period sequential decision model; when the verification of the outer-layer multi-period sequential decision model is passed, the unit commitment value of the thermal power unit is taken as a final unit commitment value; when the verification of the outer-layer multi-period sequential decision model is not passed, the inner-layer two-stage robust unit commitment model is updated, and the determined step and the verified step are repeatedly iterated and executed according to the updated inner-layer two-stage robust unit commitment model to make a decision on the unit commitment of the power system, so that a unit commitment scheme capable of guaranteeing the real-time operation feasibility of the power system is obtained.

Description

Technical Field

[0001] This invention relates to the field of power system dispatching and operation technology, and in particular to a two-layer robust unit combination method and components for power systems. Background Technology

[0002] Faced with the pressures of resource scarcity and environmental degradation, vigorously developing renewable energy has become a broad consensus in the power industry. While the large-scale grid connection of renewable energy sources, such as wind and solar power, promotes the "green" development of the power system, the uncertainty of their output poses a severe challenge to the dispatch and operation of the power system. To address this challenge, the power system must utilize more generating units and reserve resources to ensure that grid dispatch meets operational constraints under random renewable energy output scenarios.

[0003] Robust unit combination refers to the decision-making process for the start-up and shutdown of thermal power units, considering the range of variation of random variables in the power system. Its aim is to ensure that the power system meets operational constraints under the worst-case scenario of these random variables. In power systems with a high proportion of renewable energy, renewable energy output is an important type of random variable. The unique characteristic of the robust unit combination problem lies in the need for unpredictability in unit start-up and shutdown decisions. This means that, constrained by predictive capabilities, the values ​​of renewable energy output are revealed sequentially over time periods. The start-up and shutdown decisions for the current period can only rely on the renewable energy output values ​​known up to this time period and the values ​​or ranges of renewable energy output for a finite future period. Failure to consider unpredictability in dispatching results may lead to infeasible intraday dispatching of the power system. Therefore, how to consider the unpredictability of decisions in robust unit combination remains a crucial issue. Summary of the Invention

[0004] This invention provides a two-layer robust unit combination method and components for power systems, which addresses the shortcomings of existing technologies that fail to consider unforeseen dispatch results, which may lead to the infeasibility of intraday actual dispatch of the power system. It fully considers the interval uncertainty of renewable energy output and the unforeseen conditions that need to be met before dispatch decisions and renewable energy output, and makes decisions on the unit combination of the power system, thereby obtaining a unit combination scheme that can ensure the real-time operational feasibility of the power system.

[0005] This invention provides a two-layer robust unit combination method for a power system, wherein the power system includes at least thermal power units and renewable energy units. The method includes: determining a set of output scenarios for renewable energy units and a unit combination value for the thermal power units based on an inner-layer two-stage robust unit combination model; verifying the feasibility of the unit combination value for the thermal power units based on an outer-layer multi-time-period sequential decision model according to the set of output scenarios for the renewable energy units in the inner-layer model; if the verification by the outer-layer multi-time-period sequential decision model is passed, using the unit combination value for the thermal power units as the final unit combination value; if the verification by the outer-layer multi-time-period sequential decision model is not passed, updating the inner-layer two-stage robust unit combination model, and repeatedly iteratively executing the determined steps and the verification steps according to the updated inner-layer two-stage robust unit combination model until the verification by the outer-layer multi-time-period sequential decision model is passed, and using the unit combination value of the thermal power units verified by the outer-layer multi-time-period sequential decision model as the final unit combination value.

[0006] According to the present invention, a two-layer robust unit combination method for a power system, the step of determining the set of output scenarios for new energy units and the unit combination value of the thermal power units obtained from the inner-layer two-stage robust unit combination model includes: acquiring power system parameters and initializing the set of output scenarios for new energy units, the set of indexes for the scenario sequences of new energy units, the iteration index between the inner and outer layers, the iteration index of the inner layer model stages, and the set of unexpected cutting plane constraints for the inner-layer pre-scheduled unit combination; and obtaining the pre-scheduled unit combination unexpected cutting plane constraint based on the inner-layer pre-scheduled unit combination optimization model based on the power system parameters, the set of output scenarios for new energy units, the set of indexes for the scenario sequences of new energy units, the iteration index between the inner and outer layers, the iteration index of the inner layer model stages, and the set of unexpected cutting plane constraints for the inner-layer pre-scheduled unit combination. The system establishes a set of output scenarios and unit combination values ​​for new energy generating units in the pre-schedule phase. Based on the set of output scenarios for new energy generating units in the pre-schedule phase, it verifies the unit combination values ​​in the pre-schedule phase using an inner-layer rescheduling phase optimization model. If the verification of the inner-layer rescheduling phase optimization model is passed, the unit combination values ​​in the pre-schedule phase are used as the unit combination values ​​for the thermal power units. If the verification of the inner-layer rescheduling phase optimization model is not passed, the inner-layer pre-schedule phase optimization model is updated, and the steps based on the parameters of the power system are iteratively executed repeatedly according to the updated inner-layer pre-schedule phase optimization model until the verification of the inner-layer rescheduling phase optimization model is passed. In this case, the unit combination values ​​in the pre-schedule phase verified by the inner-layer rescheduling phase optimization model are used as the unit combination values ​​for the thermal power units.

[0007] According to a two-layer robust unit combination method for a power system provided by the present invention, the step of verifying the unit combination value of the pre-schedule stage based on the output scenario set of the new energy units in the pre-schedule stage and an optimization model of the inner-layer rescheduling stage includes: obtaining the optimal solution of the intraday highest operating cost auxiliary variable of the inner-layer pre-schedule stage optimization model based on the output scenario set of the new energy units in the pre-schedule stage; when the optimal solution of the intraday highest operating cost auxiliary variable and the objective function value of the inner-layer rescheduling stage optimization model satisfy a first preset formula, the unit combination value of the pre-schedule stage passes the verification of the inner-layer rescheduling stage optimization model; when the optimal solution of the intraday highest operating cost auxiliary variable and the objective function value of the inner-layer rescheduling stage optimization model do not satisfy the first preset formula, the unit combination value of the pre-schedule stage fails the verification of the inner-layer rescheduling stage optimization model; the first preset formula is:

[0008]

[0009] in, The optimal solution for the intraday highest operating cost auxiliary variable η of the inner pre-scheduling phase optimization model; ε is the objective function value of the optimized model in the inner rescheduling phase; m is the iteration index of the inner model phase; n is the iteration index between the inner and outer models; wait is the pre-scheduling phase; and see is the rescheduling phase.

[0010] According to a two-layer robust unit combination method for a power system provided by the present invention, when the inner layer rescheduling stage optimization model fails the verification, updating the inner layer pre-scheduling stage optimization model includes: when the optimal solution of the highest intraday operating cost auxiliary variable and the objective function value of the inner layer rescheduling stage optimization model do not satisfy the first preset formula, updating the inner layer model stage iteration index, updating the output scenario set of the new energy units and the index set of the scenario sequence of the new energy units, so as to update the inner layer pre-scheduling stage optimization model according to the updated inner layer model stage iteration index, the output scenario set of the new energy units, and the index set of the scenario sequence of the new energy units.

[0011] According to a two-layer robust unit combination method for a power system provided by the present invention, the feasibility of the unit combination value of the thermal power unit based on the output scenario set of the new energy units in the inner layer model and the outer layer multi-time-period sequential decision model is verified. This includes: determining the decision solution of the outer layer multi-time-period sequential decision model based on the output scenario set of the new energy units in the inner layer model; calculating the slack variable penalty cost based on the decision solution; when the slack variable penalty cost satisfies a second preset formula, the unit combination value of the thermal power unit passes the verification of the outer layer multi-time-period sequential decision model; when the slack variable penalty cost does not satisfy the second preset formula, the unit combination value of the thermal power unit fails the verification of the outer layer multi-time-period sequential decision model. The second preset formula is:

[0012]

[0013] in, The slack variable is the penalty cost, y is the index of the new energy scenario sequence, n is the iteration index between the inner and outer layer models, and punish is the penalty.

[0014] According to a two-layer robust unit combination method for a power system provided by the present invention, determining the decision solution of the outer multi-time-period sequential decision model includes: obtaining the decision solution based on the outer multi-time-period sequential decision model when the scheduling period meets a preset time-period threshold; updating the scheduling period when the scheduling period does not meet the preset time-period threshold, and obtaining the decision solution based on the outer multi-time-period sequential decision model according to the updated scheduling period.

[0015] According to a two-layer robust unit combination method for a power system provided by the present invention, when the outer multi-time-period sequential decision model fails the verification, updating the inner two-stage robust unit combination model includes: when the slack variable penalty cost does not satisfy the second preset formula, updating the iteration index between the inner and outer models, and updating the inner pre-scheduled unit combination unexpected plane constraint set, so as to update the inner two-stage robust unit combination model according to the updated iteration index between the inner and outer models and the inner pre-scheduled unit combination unexpected plane constraint set.

[0016] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the two-layer robust unit combination method of the power system as described above.

[0017] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the two-layer robust unit combination method of the power system as described above.

[0018] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements a two-layer robust unit combination method for a power system as described above.

[0019] This invention provides a two-layer robust unit combination method and components for a power system, wherein the power system includes at least thermal power units and renewable energy units. The method includes: determining the output scenario set of renewable energy units and the unit combination value of thermal power units in the inner layer model, obtained based on an inner-layer two-stage robust unit combination model; verifying the feasibility of the unit combination value of thermal power units based on the output scenario set of renewable energy units in the inner layer model and an outer-layer multi-time-period sequential decision model; if the feasibility of the feasibility of the feasibility of the thermal power unit combination value is passed by the outer-layer multi-time-period sequential decision model, the unit combination value of the thermal power units is taken as the final unit combination value; if the feasibility of the feasibility of the feasibility of the feasibility of the outer-layer multi-time-period sequential decision model is not passed, the unit combination value of the thermal power units is taken as the final unit combination value; if ... outer-layer multi-time-period sequential decision model is not passed, the unit combination value of the thermal power units is taken as the final unit combination value; if the feasibility of the feasibility of the outer-layer multi-time-period sequential decision model is not passed, the unit combination value of the thermal power units is taken as the final unit combination value; if the feasibility of the feasibility of the outer-layer multi-time-period sequential decision model is not passed, the unit combination value of the thermal power units is taken as the final unit combination value. During the verification of the decision model, the inner two-stage robust unit combination model is updated. Based on the updated inner two-stage robust unit combination model, the determined steps and verification steps are repeatedly executed iteratively until the outer multi-time sequential decision model is verified. When the verification of the outer multi-time sequential decision model is passed, the unit combination value of the thermal power unit verified by the outer multi-time sequential decision model is taken as the final unit combination value. The unit combination decision of the power system is made by fully considering the interval uncertainty of renewable energy output and the unpredictability that dispatch decisions and renewable energy output need to meet before they can be implemented, thereby obtaining a unit combination scheme that can ensure the real-time operation feasibility of the power system. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating a two-layer robust unit combination method for a power system provided by the present invention;

[0022] Figure 2 This is a schematic diagram of the specific process of a two-layer robust unit combination method for a power system provided by the present invention;

[0023] Figure 3 This is the IEEE 39-node system topology diagram provided by the present invention;

[0024] Figure 4 The present invention provides the load curve, predicted center value of total output power of wind farm, and confidence interval diagram of the example.

[0025] Figure 5 This is a diagram showing the unit combination results provided by existing technology;

[0026] Figure 6 This is a graph showing the summation of slack variables provided by existing technology;

[0027] Figure 7 This is a diagram showing the unit combination results provided by the present invention;

[0028] Figure 8 This is a graph showing the summation of relaxed variables during inner and outer layer iterations, provided by the present invention.

[0029] Figure 9 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0031] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a two-layer robust unit combination method for a power system provided by the present invention.

[0032] Please refer to Figure 2 , Figure 2 A schematic diagram illustrating the specific process of a two-layer robust unit combination method for a power system provided by the present invention.

[0033] This invention provides a two-tier robust unit combination method for a power system, wherein the power system includes at least thermal power units and new energy units, and the method includes:

[0034] 11: Determine the set of output scenarios for new energy units and the unit combination values ​​for thermal power units based on the inner-layer two-stage robust unit combination model obtained from the inner-layer model;

[0035] As a preferred embodiment, the set of output scenarios for new energy units and the unit combination values ​​for thermal power units are determined based on the inner-layer two-stage robust unit combination model, including:

[0036] 101: Obtain the parameters of the power system and initialize the output scenario set of new energy units, the index set of scenario sequences of new energy units, the iteration index between inner and outer layer models, the stage iteration index of the inner layer model, and the set of unexpected cutting plane constraints for the combination of inner layer pre-scheduled units;

[0037] Specifically, the parameters of thermal power units include: the output range of thermal power unit i. Uphill speed Downhill speed Minimum runtime T i on Minimum downtime T i off Start-up costs Shutdown costs Unit power output cost and the power transfer factor of thermal power unit i to line l Let the set of thermal power units be denoted as The number of thermal power units is N G .

[0038] Load parameters include: the predicted value of load d in time period t. Generation transfer factor of load d to line l Let the set of loads be . The number of loads is N D .

[0039] The parameters of the new energy generating unit include: the predicted central value of the new energy generating unit j in time period t. The predicted power output range of the new energy unit j in time period t The power transfer factor of new energy unit j to line l Let the collection of new energy units be denoted as The number of new energy generating units is N R .

[0040] The parameters of an energy storage power station include: the upper limit of the charging and discharging power of energy storage power station k. Maximum capacity Capacity lower limit Charging efficiency Discharge efficiency The initial energy value of energy storage power station k Cost per unit energy storage charging and discharging power of energy storage power station k Generation transfer factor of energy storage power station k to line l Let the collection of energy storage power stations be denoted as The number of energy storage power stations is N E .

[0041] Other parameters: Let the maximum power transmission capacity of line l be... Let the number of transmission lines be N. L The transmission line set is Let the number of nodes be N. B Let the set of nodes be denoted as . Let N be the number of scheduling periods.T The set of scheduling periods is

[0042] Initialize new energy power output scenario The index n of the iteration count for the inner and outer layers, and the iteration index m for the two stages within the inner layer.

[0043] Specifically, the set of power output scenarios for the new energy generator units used in this invention is denoted as follows: New energy power output scenarios in The tag is y, where y is the index of the new energy scenario sequence. It is a set The number of elements contained. Let the index set of the scenario sequence for new energy generating units be denoted as . have Initial new energy power output scenario Set as

[0044] Given the limited number of English letters and the large number of parameters in the text, some parameters are represented by two letters to distinguish them. The subscript letters have no specific meaning and are only used to facilitate differentiation from other parameters.

[0045] In this invention, the index of the iteration count for the inner and outer layers is set to n, and the iteration index for the two stages within the inner layer is set to m. Initialize n = 1 and m = 1.

[0046] Please refer to Figure 3 , Figure 3 The IEEE 39-node system topology diagram provided by this invention.

[0047] Please refer to Figure 4 , Figure 4 The diagram provided for this invention shows the load curve, predicted center value of total wind farm output power, and confidence interval. The horizontal axis represents the dispatch period (in hours), and the vertical axis represents the load and wind farm power values ​​(in megawatts). The solid black line represents the load power curve, the dashed black line represents the predicted center value of total wind farm output power, and the gray area represents the confidence interval of total wind farm output power. The total output power of the wind farm will fluctuate within the confidence interval during the day.

[0048] 102: Based on the parameters of the power system, the set of output scenarios of new energy units, the set of indexes of scenario sequences of new energy units, the iteration index between inner and outer layer models, the iteration index of inner layer model stages, and the set of unexpected cutting plane constraints for inner layer pre-scheduled unit combinations, the set of output scenarios and unit combination values ​​of new energy units in the pre-scheduled stage are obtained based on the inner layer pre-scheduled stage optimization model.

[0049] Specifically, the inner layer of this invention is a two-stage robust unit combination model, which has two stages: pre-scheduling and rescheduling.

[0050] During the pre-scheduling phase, the scheduling center needs to make decisions on the start and stop of the units, which requires constructing and training an optimization model with constraints (1)-(17). The decision variables and specific meanings of constraints (1)-(17) will be explained in detail below.

[0051] Construct inner-layer pre-scheduled thermal power unit start-up and shutdown constraints (1)-(3): u i (t), All are zero-one variables: u it u is the start-up and shutdown variable for thermal power units. it =0 / 1 indicates that thermal power unit i is in a shutdown / operation state during time period t; For unit start-up indicator variables, This indicates whether thermal power unit i is not started or is started during time period t; This is a variable indicating unit shutdown. This indicates that thermal power unit i is either not shut down or shut down during time period t; The variable represents the operating time of thermal power unit i during time period t. Let be the shutdown duration variable of thermal power unit i, representing the shutdown duration of thermal power unit i in time period t. Constraint (1) is the association constraint between the unit start-up, shutdown variables and the state variables; constraint (2) is the minimum operating time constraint of the unit; constraint (3) is the minimum shutdown time constraint of the unit.

[0052]

[0053]

[0054]

[0055] Constructing feasible cut constraints for the inner layer pre-scheduled robust unit combination: Since the inner layer uses alternating iterations of pre-scheduling and re-scheduling to achieve convergence, the set of scenarios where re-scheduling feeds back to pre-scheduling is considered. Need to be based on Each scene in Construct feasible cuts for robust unit combinations (4)-(16). Among them, variables This represents the active power output of thermal power unit i under scenario y and time period t; variables To represent the input active power of energy storage power station k in scenario y and time period t, the variable E represents the output active power of energy storage power station k under scenario y and time period t. k,y (t) represents the stored electricity of energy storage station k in scenario y and time period t; variable This represents the active power output of the new energy power plant j under scenario y and time period t; variables Let y be the power balance constraint relaxation variable under scenario y and time period t. Let be the slack variable for the capacity constraint of transmission line l under scenario y; η is the auxiliary variable for the highest daily operating cost, which represents the system rescheduling cost considering the worst renewable energy scenario during the pre-scheduling phase, where the coefficient c p The penalty cost per unit of slack variable.

[0056]

[0057]

[0058]

[0059]

[0060]

[0061]

[0062]

[0063]

[0064]

[0065]

[0066]

[0067]

[0068]

[0069] In constraints (4)-(16), constraint (4) is the upper and lower limit constraint of the output of thermal power unit, constraint (5) and constraint (6) are the upper and lower ramp constraints of thermal power unit respectively, constraint (7) and constraint (8) are the charging and discharging power limits of energy storage respectively, constraint (9) is the time period coupling constraint of the stored energy, constraint (10) is the upper and lower limit constraint of energy storage capacity; constraint (11) is the numerical constraint of the energy storage power at the beginning and end of the day, constraint (12) is the output power constraint of typical new energy, constraint (13) and constraint (14) are the system power balance constraint and line capacity constraint respectively, constraint (15) is the non-negativity constraint of slack variable, and constraint (16) is the constraint representing the highest operating cost within the day under typical scenarios.

[0070] Construct and add inner-layer pre-scheduled unit combination unexpected cutting plane constraints (17): In the inner and outer layer iteration process of this invention, the outer layer will pass unit combination unexpected cutting plane constraints to the inner layer. Therefore, the start-up and shutdown variables of thermal power units need to satisfy constraint (17). Among them, It is the set of unexpected cutting plane constraints for unit combination: when n=1, Indicate u i (t) The entire feasible region; when n≥2, The expression is detailed in equation (59).

[0071]

[0072] Construct the inner pre-scheduling objective function (18): The objective function of the pre-scheduling stage is to minimize the system scheduling cost. The expression of the objective function is shown in equation (18).

[0073]

[0074] Construct and train the inner pre-schedule optimization model (19): The objective function (18) includes the start-up cost of the thermal power unit, the shutdown cost of the thermal power unit, and the maximum operating cost in a typical scenario. The expression of the optimization model in the pre-schedule stage is shown in equation (19).

[0075]

[0076] st(1)~(17)

[0077] Training the inner pre-scheduling optimization model (19) will yield the total system operating cost, the decision value of the auxiliary variable η for the highest daily operating cost during the pre-scheduling phase, and the unit start-up and shutdown variables. The decision value. Since η represents the system rescheduling cost considering the worst renewable energy scenario in the pre-scheduling phase, the optimal solution of the auxiliary variable η of the highest intraday operating cost in (19) is denoted as remember The decision value Here, footnote n is the index of the number of iterations of the inner and outer models in this invention; footnote m is the index of the number of pre-scheduling and re-scheduling iterations in the inner layer. Then, and Optimize the model for transmission to the rescheduling phase.

[0078] Train the inner layer's rescheduling phase optimization model for the m-th iteration to obtain the rescheduling objective function value. In rescheduling cost and rescheduling objective function value satisfy At that time, output the inner layer unit combination decision value. That is, the unit combination value in the pre-scheduling phase and the updated set of output scenarios for new energy units. The set of indexes for scenario sequences of new energy units To the outermost multi-time sequential decision model.

[0079] 103: Based on the set of output scenarios of new energy units in the pre-schedule phase, the unit combination value in the pre-schedule phase is verified based on the inner reschedule phase optimization model;

[0080] 104: When verifying the optimization model through the inner rescheduling stage, the unit combination value of the pre-scheduling stage shall be used as the unit combination value of the thermal power unit;

[0081] 105: If the inner rescheduling stage optimization model fails to pass the verification, update the inner pre-scheduling stage optimization model. Repeatedly iterate the steps based on the power system parameters according to the updated inner pre-scheduling stage optimization model until the inner rescheduling stage optimization model passes the verification. Then, use the unit combination value of the pre-scheduling stage that has passed the inner rescheduling stage optimization model verification as the unit combination value of the thermal power unit.

[0082] As a preferred embodiment, based on the set of output scenarios for new energy units in the pre-schedule phase, the unit combination value in the pre-schedule phase is verified using the inner-layer rescheduling phase optimization model. This includes: obtaining the optimal solution of the intraday highest operating cost auxiliary variable of the inner-layer pre-schedule phase optimization model based on the set of output scenarios for new energy units in the pre-schedule phase; when the optimal solution of the intraday highest operating cost auxiliary variable and the objective function value of the inner-layer rescheduling phase optimization model satisfy a first preset formula, the unit combination value in the pre-schedule phase passes the verification of the inner-layer rescheduling phase optimization model; when the optimal solution of the intraday highest operating cost auxiliary variable and the objective function value of the inner-layer rescheduling phase optimization model do not satisfy the first preset formula, the unit combination value in the pre-schedule phase fails the verification of the inner-layer rescheduling phase optimization model; the first preset formula is:

[0083]

[0084] in, The optimal solution for the auxiliary variable η, representing the highest intraday operating cost, in the inner pre-scheduling phase optimization model; ε is the objective function value of the model optimized in the inner rescheduling phase; m is the iteration index of the inner model phase; n is the iteration index between the inner and outer models; wait is the pre-scheduling phase; and see is the rescheduling phase.

[0085] As a preferred embodiment, when the inner layer rescheduling stage optimization model fails to pass the verification, the inner layer pre-scheduling stage optimization model is updated, including: when the optimal solution of the auxiliary variable of the highest operating cost within the day does not satisfy the first preset formula with the objective function value of the inner layer rescheduling stage optimization model, the inner layer model stage iteration index is updated, the output scenario set of the new energy units and the index set of the scenario sequence of the new energy units are updated, so as to update the inner layer pre-scheduling stage optimization model according to the updated inner layer model stage iteration index, the output scenario set of the new energy units and the index set of the scenario sequence of the new energy units.

[0086] Specifically, to facilitate model training, the rescheduling phase of the two-stage robust unit combination assumes that the renewable energy output values ​​for all time periods are revealed simultaneously.

[0087] During the rescheduling phase, the power system needs to construct and train an optimization model with constraints (20)-(34). The decision variables of the constraints and the specific meanings of constraints (20)-(34) will be described in detail below.

[0088] Constructing inner-layer rescheduling of thermal power unit output and ramp-up constraints:

[0089] In constraints (20)-(22), the variables This represents the active power output of thermal power unit i during time period t. Constraint (20) is the upper and lower limit constraint of the thermal power unit's output, and constraints (21) and (22) are the uphill and downhill constraints of the thermal power unit, respectively.

[0090]

[0091]

[0092]

[0093] In constraints (23)-(27), the variables The variable represents the input active power of energy storage power station k in time period t. E represents the output active power of energy storage power station k in time period t. k (t) represents the stored electricity of energy storage station k in time period t. Constraints (23) and (24) are the energy storage charging and discharging power limits, respectively. Constraint (25) is the time period coupling constraint of the stored electricity state. Constraint (26) is the upper and lower limit constraint of energy storage capacity. Constraint (27) is the numerical constraint of the stored electricity at the beginning and end of the day.

[0094]

[0095]

[0096]

[0097]

[0098]

[0099] Construct inner-layer rescheduling constraints for new energy power plant operation (28)-(31): In constraints (28)-(31), Let y be the active power output variable of the new energy power plant j under scenario y and time period t; For zero-one variables, This indicates that the active power output of the new energy power plant j in time period t is at the upper limit of the output prediction range; This indicates that the active power output of the new energy power plant j in time period t is at the lower bound of the predicted output range. Γ s Γ represents the spatial clustering coefficient of new energy power stations. T is the time smoothing coefficient of the new energy power station. Constraint (28) is the constraint on the active power output variable of new energy. Here, it is set that the new energy output in the adverse scenario may appear at the prediction center value and the prediction boundary; Constraint (29) indicates that the new energy output in the same period cannot be at both the upper and lower bounds; Constraint (30) and Constraint (31) are the spatial cluster effect constraint of multiple new energy power stations and the time smoothing effect constraint of a single new energy power station, respectively.

[0100]

[0101]

[0102]

[0103]

[0104] Constructing system power balance and transmission line capacity constraints for inner-layer rescheduling (32)-(34): In constraints (32)-(34), s 1 (t),s 2 (t) represents the slack variable of the power balance constraint. Let l be the slack variable for the transmission line capacity constraint. Constraint (32) is the system power balance constraint, constraint (33) is the line capacity constraint, and constraint (34) is the slack variable non-negativity constraint.

[0105]

[0106]

[0107]

[0108] Construct the inner rescheduling objective function (35): The objective function of the rescheduling stage is to minimize the system operating cost under the worst new energy output scenario. The expression of the objective function is shown in equation (35).

[0109]

[0110] Construct and train the inner rescheduling optimization model (36): The objective function (35) includes the operating cost of thermal power units, the charging and discharging cost of energy storage, and the slack variable penalty cost. The expression of the rescheduling stage optimization model is shown in (36).

[0111]

[0112] The rescheduling objective function η can be obtained by training the inner rescheduling optimization model (36). see numerical value and the active power output variables of new energy units The value of

[0113] Constructing an inner-layer iterative convergence criterion (37): The numerical value of the rescheduling cost... The value of the auxiliary variable for the highest intraday operating cost obtained during the pre-scheduling phase Substitute into equation (37), where ε is the preset allowable deviation.

[0114]

[0115] If equation (37) holds, the inner layer ends and transitions to the “outer layer multi-period sequential decision-making model”; otherwise, the newly obtained adverse new energy output scenario in the rescheduling phase will be re-scheduled. Include in set Let m = m + 1, return to the "pre-scheduling phase optimization model", and continue iterating until equation (37) is true.

[0116] After the pre-scheduling and re-scheduling iterations are completed, the inner layer outputs the decision values ​​for starting and stopping the thermal power units. The set of power output scenarios for new energy units obtained from inner-layer iterations Will and Once transmitted to the outer layer, the iteration between the inner and outer layers of this invention can begin.

[0117] 12: Based on the set of output scenarios of new energy units in the inner layer model, the feasibility of the unit combination value of thermal power units is verified based on the outer layer multi-time sequential decision model;

[0118] 13: When verifying the outer multi-time sequential decision model, the unit combination value of the thermal power unit is taken as the final unit combination value;

[0119] 14: If the outer multi-stage sequential decision model fails to pass the verification, update the inner two-stage robust unit combination model. Repeatedly iterate and execute the determined steps and verification steps according to the updated inner two-stage robust unit combination model until the outer multi-stage sequential decision model passes the verification. Then, the unit combination value of the thermal power unit that has passed the verification of the outer multi-stage sequential decision model is taken as the final unit combination value.

[0120] As a preferred embodiment, based on the set of output scenarios for new energy units in the inner layer model, the feasibility of the unit combination value of thermal power units is verified using the outer layer multi-time-period sequential decision model. This includes: determining the decision solution of the outer layer multi-time-period sequential decision model based on the set of output scenarios for new energy units in the inner layer model; calculating the slack variable penalty cost based on the decision solution; when the slack variable penalty cost satisfies a second preset formula, the unit combination value of the thermal power unit passes the verification of the outer layer multi-time-period sequential decision model; when the slack variable penalty cost does not satisfy the second preset formula, the unit combination value of the thermal power unit fails the verification of the outer layer multi-time-period sequential decision model. The second preset formula is:

[0121]

[0122] in, The penalty cost is the slack variable, y is the index of the new energy scenario sequence, n is the iteration index between the inner and outer models, and punish is the penalty.

[0123] As a preferred embodiment, determining the decision solution of the outer multi-time-period sequential decision model includes: obtaining the decision solution based on the outer multi-time-period sequential decision model when the scheduling period meets the preset time-period threshold; updating the scheduling period when the scheduling period does not meet the preset time-period threshold, and obtaining the decision solution based on the outer multi-time-period sequential decision model according to the updated scheduling period.

[0124] As a preferred embodiment, when the outer multi-time sequential decision model fails the verification, the inner two-stage robust unit combination model is updated, including: when the slack variable penalty cost does not meet the second preset formula, updating the iteration index between the inner and outer models, and updating the inner pre-scheduled unit combination unexpected plane constraint set, so as to update the inner two-stage robust unit combination model according to the updated iteration index between the inner and outer models and the inner pre-scheduled unit combination unexpected plane constraint set.

[0125] Specifically, the outer layer multi-time sequential decision-making model: the inner layer and After being transmitted to the outer layer, it can then enter the outer layer decision model of this invention. As described above, The footnote 'm' in the text only represents the inner iteration process. For the sake of brevity and without causing confusion, the outer layer will use... express In this context, footnote n is the index of the number of iterations of the inner and outer layers of this invention.

[0126] In actual dispatching, due to limitations in forecasting technology, the power system in time period t0 can only predict the renewable energy output from time period t0 to time period t0+ΔT with relatively high accuracy, where ΔT is the number of time periods in which renewable energy output can be accurately predicted. For Each scene in At time period t0, the predictable output value of new energy sources in the power system is Considering the time-varying coupling characteristics of decision variables for thermal power units and energy storage power stations, the unit output value in time period t0-1 And the stored electricity of energy storage power stations This will affect the decision variables in time period t0.

[0127] t1 = min(N) T ,t0+ΔT) (38)

[0128] Based on the above characteristics, the unit output decision variables from t0 to t1 need to be considered during the scheduling period t0, where t1 satisfies equation (38). The power system needs to construct constraints (39)-(49) to establish an optimization model for the scheduling period t0.

[0129] Construction of the outer thermal power unit output range and ramping constraints (39)-(41): Under constraints (39)-(41), Let represent the active power output variable of thermal power unit i under scenario y and time period t. Constraint (39) is the upper and lower limit constraint of thermal power unit output, and constraints (40) and (41) are the uphill and downhill constraints of thermal power unit, respectively.

[0130]

[0131]

[0132]

[0133] Construction of outer layer energy storage power station operation constraints (42)-(46): Under constraints (42)-(44), This represents the active power input variable of energy storage power station k under scenario y and time period t. E represents the active power output variable of energy storage power station k under scenario y and time period t. k,y (t) represents the stored electricity of energy storage station k in scenario y and time period t. Constraints (42) and (43) are the charging and discharging power limits of the energy storage station, respectively. Constraint (44) is the time period coupling constraint of energy storage capacity, and constraint (45) is the upper and lower limit constraint of energy storage capacity.

[0134]

[0135]

[0136]

[0137]

[0138] Additionally, during the scheduling period t0 = 1, there is During the end-of-day scheduling period, it is necessary to consider the numerical constraint of the stored energy, namely: when t0 + ΔT ≥ N T When doing so, constraints need to be considered (46).

[0139]

[0140] Constructing outer-layer new energy predicted output constraints (47): Constraint (47) is a new energy predicted output constraint, in which, It is the active power output variable of the new energy power plant j under scenario y and time period t.

[0141]

[0142] Constructing outer layer power balance and transmission line capacity constraints (48)-(50): Under constraints (48)-(50), and These are the power balance constraint relaxation variables in scenario y. and Let l be the slack variable for the capacity constraint of transmission line l under scenario y. Constraints (48) and (49) are the system power balance constraint and the line capacity constraint, respectively, and constraint (50) is the non-negativity constraint of the slack variable.

[0143]

[0144]

[0145]

[0146] Constructing the outer objective function (51): For the set of new energy scenarios For each scenario, the objective function of scheduling period t0 is to minimize the system operating cost from t0 to t1, and the expression of the objective function is shown in equation (51). Where c p The penalty cost per unit of slack variable.

[0147]

[0148] Construct and train the outer multi-time-period sequential decision model (52): The objective function (51) includes the active power output cost of the thermal power unit, the energy storage charging and discharging cost, and the slack variable penalty cost from time period t0 to t1. Therefore, the optimization model expression (52) for scheduling time period t0 is shown.

[0149]

[0150] Training the outer multi-time-period sequential decision model (52) at time period t0 yields the decision values ​​of each variable from time period t0 to time period t1. The decision value at t = t0 is selected as the sequential decision solution for the current time period. For all time periods t0 = 1,...,N T After training the outer multi-time sequential decision model (52), the first... New energy scenarios Decision values ​​for unit start-up and shutdown Multi-time sequential decision solution

[0151] exist middle, Decision variables The sequential decision solution, These are decision variables The sequential decision solution, These are decision variables The sequential decision solution, with footnote n being the index of the number of iterations of the inner and outer layers of this invention.

[0152] Establish termination criteria for inner and outer iterations: In The feasibility of the sequential decision scheduling solution is characterized: if for any have If the condition is met, the iterative process between the inner and outer layers ends, and the current unit start-up / shutdown decision value is output. The unit combination value of thermal power units is used as the decision result of this method, i.e., the final unit combination value; otherwise, the decision value is based on slack variables. Construct unexpected cutting plane constraints for unit combination, then let n = n + 1, update the constraints of the inner pre-scheduling stage (17), and start the next round of inner and outer iterations.

[0153] The following describes the method for constructing unintended cutting plane constraints for unit combinations.

[0154] Establish a method for constructing unintended cutting plane constraints for unit combination (53)-(59): Define the scenario in the nth iteration. The cost of slack variable penalty is Its expression (53) is shown.

[0155]

[0156] In the optimization model (52) for time period t0, the scene y is represented and contains The constraints are shown in (54)-(56), where, and These are the dual variables of constraints (54), (55), and (56), respectively.

[0157]

[0158]

[0159]

[0160] After the optimized model (52) for all time periods has been trained, record and These are the dual variables. and The value of .

[0161] Let the set of unexpected cutting plane constraints for unit combination be . If the relaxation cost of scenario y is If non-zero, then the set of constraints in the unexpected cutting plane of the unit combination. Add constraints (57) to the list.

[0162]

[0163] In constraint (57), It is a weighting coefficient that can improve the correction effect of the unit combination sequential evolution cutting plane on the upper-level optimization model and improve the feasibility of the n+1 iteration sequential evolution results. In this invention, The specific expression is shown in equation (58).

[0164]

[0165] Then, after the nth iteration of this invention, The expression for is shown in equation (59).

[0166]

[0167] The following example illustrates the effects of the present invention.

[0168] Simulation System: Considering improvements to the IEEE 39-node system, the system topology is as follows. Figure 3 As shown in the table. The system has 10 thermal power units; the parameters of the thermal power units are detailed in Table 1. The system load distribution at the nodes is as follows. Figure 3As shown, the peak and trough daily loads are 5003.4MW and 2952MW, respectively, and the load proportion of each node to the total load is the same as in the standard IEEE 39-node system. In this example, wind farms with an installed capacity of 500MW (equivalent to 250 standard wind turbines) and energy storage power stations with a power cap of 200MW and a capacity cap of 200MWh are installed at nodes 8, 14, and 29, respectively. The predicted center value and confidence interval of the total wind power output of the three wind farms are shown below. Figure 4 As shown in Table 2, the three energy storage power stations have identical parameters. The system has 46 lines, with the same transmission capacity as the IEEE 39-node system. Scheduling occurs once per hour during the day, with N scheduling periods. T =24.

[0169] Table 1 Parameters of Thermal Power Units

[0170]

[0171] Table 2. Relevant parameters of energy storage batteries

[0172]

[0173] During initialization, the predicted center value of the total output power of the wind farm is placed into the output scenario set of the new energy units. Set the number of time periods during which the output of new energy sources can be accurately predicted to be ΔT = 4. Set ε in equation (37) to 0.01, and set the penalty cost per unit slack variable to c. p =5×10 4 $ / MW. The actual dispatching strategies for each time period of the power system during the day are the same as those used in an outer multi-time-period sequential decision model. Initialize n=1, m=1, then present the results and compare the examples.

[0174] Please refer to Figure 5 , Figure 5 A diagram showing the unit combination results provided for existing technologies.

[0175] Please refer to Figure 6 , Figure 6 A graph showing the summation of slack variables provided for existing technologies.

[0176] Please refer to Figure 7 , Figure 7 The diagram shows the unit combination results provided by this invention.

[0177] Please refer to Figure 8 , Figure 8 The graph shows the summation of relaxed variables during inner and outer layer iterations, as provided by this invention.

[0178] Under this real-time scheduling strategy for different time periods within the day, the unit combination results obtained by existing technology show that the sum of slack variables is greater than 0 in time periods 22, 23, and 24, indicating a power imbalance in the power system. During these three time periods, the feasibility of actual intraday power system scheduling is difficult to meet.

[0179] In this invention, the unit combination result was obtained after six "inner-outer layer" iterations. The sum of the slack variables for real-time scheduling in each time period of the day, based on the unit combination results from the first to the sixth iteration, shows that the unit combination result in the sixth iteration enables the real-time scheduling strategy for each time period of the day to be feasible. Therefore, the unit combination result of this invention can satisfy the feasibility of actual daily scheduling of the power system.

[0180] Compared with the unit combination results of the prior art, the unit combination results obtained by the present invention add unit G5 in time period 20 to time period 23, and shut down unit G8 in time period 23 and time period 24. This adjustment makes the unit combination results of the present invention meet the feasibility of the actual intraday dispatch of the power system.

[0181] In summary, this invention addresses the day-ahead unit combination and scheduling problem in power systems by designing a two-layer robust unit combination method for power systems. This method fully considers the range uncertainty of renewable energy output and the unpredictability that scheduling decisions and renewable energy output must meet beforehand, and makes decisions on the unit combination of the power system to obtain a unit combination scheme that can ensure the real-time operational feasibility of the power system.

[0182] The beneficial effects of this invention are:

[0183] First, this invention establishes a multi-period sequential decision-making model, which can make multi-period sequential decisions for a given set of new energy scenarios. This model is closer to the actual dispatching of the power system.

[0184] Second, based on the scheduling solution of the multi-period sequential decision model, this invention proposes a method for constructing unexpected cutting plane constraints for unit combination;

[0185] Third, by adding this unexpected cutting plane constraint to the two-layer robust unit combination model, the unit combination results can be corrected, the unexpectedness of the unit combination results can be improved, and the feasibility of the unit combination results in real-time scheduling can be significantly improved.

[0186] Fourth, the method proposed in this invention is general and applicable to various types of integer variable optimization decisions with unpredictable decision-making requirements. Furthermore, this method exhibits good compatibility with widely used two-stage robust unit combination methods, high computational efficiency, and strong engineering practicality.

[0187] Figure 9 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 9 As shown, the electronic device may include: a processor 901, a communication interface 902, a memory 903, and a communication bus 904, wherein the processor 901, the communication interface 902, and the memory 903 communicate with each other through the communication bus 904. The processor 901 can call logic instructions in the memory 903 to execute a two-layer robust unit combination method for a power system. The power system includes at least thermal power units and renewable energy units. The method includes: determining the set of output scenarios for renewable energy units and the unit combination value for thermal power units based on the inner-layer two-stage robust unit combination model; verifying the feasibility of the unit combination value for thermal power units based on the set of output scenarios for renewable energy units in the inner-layer model and the outer-layer multi-time-period sequential decision model; if the verification of the outer-layer multi-time-period sequential decision model is passed, using the unit combination value for thermal power units as the final unit combination value; if the verification of the outer-layer multi-time-period sequential decision model is not passed, updating the inner-layer two-stage robust unit combination model, and repeatedly iteratively executing the determined steps and verification steps according to the updated inner-layer two-stage robust unit combination model until the verification of the outer-layer multi-time-period sequential decision model is passed, and using the unit combination value of thermal power units verified by the outer-layer multi-time-period sequential decision model as the final unit combination value.

[0188] Furthermore, the logical instructions in the aforementioned memory 903 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0189] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the two-layer robust unit combination method for power systems provided by the above methods. The power system includes at least thermal power units and new energy units. The method includes: determining the set of output scenarios for new energy units and the unit combination value for thermal power units in the inner layer model obtained based on the inner layer two-stage robust unit combination model; and, based on the set of output scenarios for new energy units in the inner layer model, determining the unit combination value for thermal power units. The multi-stage sequential decision model verifies the feasibility of the unit combination value of thermal power units. When the verification of the outer multi-stage sequential decision model is passed, the unit combination value of the thermal power units is taken as the final unit combination value. When the verification of the outer multi-stage sequential decision model is not passed, the inner two-stage robust unit combination model is updated. The determined steps and verification steps are repeatedly executed iteratively according to the updated inner two-stage robust unit combination model until the verification of the outer multi-stage sequential decision model is passed. When the verification of the outer multi-stage sequential decision model is passed, the unit combination value of the thermal power units verified by the outer multi-stage sequential decision model is taken as the final unit combination value.

[0190] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the two-layer robust unit combination method for a power system provided by the methods described above. The power system includes at least thermal power units and renewable energy units. The method includes: determining the set of output scenarios for renewable energy units and the unit combination value for thermal power units in the inner layer model obtained based on the inner layer two-stage robust unit combination model; and, based on the set of output scenarios for renewable energy units in the inner layer model, performing a two-layer robust unit combination method for a power system in the outer layer multi-time sequential decision model. The feasibility of the unit combination value of the power generation units is verified. If the verification of the outer multi-time sequential decision model is passed, the unit combination value of the thermal power units is taken as the final unit combination value. If the verification of the outer multi-time sequential decision model is not passed, the inner two-stage robust unit combination model is updated. The determined steps and verification steps are repeatedly executed iteratively according to the updated inner two-stage robust unit combination model until the verification of the outer multi-time sequential decision model is passed. If the verification of the outer multi-time sequential decision model is passed, the unit combination value of the thermal power units verified by the outer multi-time sequential decision model is taken as the final unit combination value.

[0191] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0192] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0193] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for combining two layers of robust generating units in a power system, characterized in that, The power system includes at least thermal power units and new energy power units, and the method includes: Determining the set of output scenarios for new energy generating units and the unit combination value for thermal power units based on the inner-layer two-stage robust unit combination model includes: acquiring power system parameters and initializing the set of output scenarios for new energy generating units, the index set of scenario sequences for new energy generating units, the iteration index between the inner and outer layers, the stage iteration index of the inner layer model, and the set of unexpected cutting plane constraints for the inner-layer pre-scheduled unit combination; obtaining the set of output scenarios for new energy generating units and the unit combination value for the pre-scheduled stage based on the inner-layer pre-scheduled stage optimization model based on the power system parameters, the set of output scenarios for new energy generating units, the index set of scenario sequences for new energy generating units, the iteration index between the inner and outer layers, the stage iteration index of the inner layer model, and the set of unexpected cutting plane constraints for the inner-layer pre-scheduled unit combination based on the inner-layer pre-scheduled stage optimization model; and verifying the unit combination value for the pre-scheduled stage based on the set of output scenarios for new energy generating units in the pre-scheduled stage based on the inner-layer re-scheduled stage optimization model based on the inner-layer re-scheduled stage optimization model. Based on the set of output scenarios for new energy units in the inner-layer model, the feasibility of the unit combination value of the thermal power unit is verified using the outer-layer multi-time-period sequential decision model. This includes: determining the decision solution of the outer-layer multi-time-period sequential decision model based on the set of output scenarios for new energy units in the inner-layer model; calculating the slack variable penalty cost based on the decision solution; when the slack variable penalty cost satisfies a second preset formula, the unit combination value of the thermal power unit passes the verification by the outer-layer multi-time-period sequential decision model; when the slack variable penalty cost does not satisfy the second preset formula, the unit combination value of the thermal power unit fails the verification by the outer-layer multi-time-period sequential decision model. The second preset formula is: ， in, The cost of the slack variable is the penalty. This serves as an index for the new energy scenario sequence. This serves as the iteration index between the inner and outer layer models. As punishment; When verifying the outer multi-time sequential decision model, the unit combination value of the thermal power unit is taken as the final unit combination value. If the outer multi-period sequential decision model fails the verification, the inner two-stage robust unit combination model is updated. The determined steps and the verification steps are repeatedly iterated and executed according to the updated inner two-stage robust unit combination model until the outer multi-period sequential decision model passes the verification. Then, the unit combination value of the thermal power unit that has passed the verification of the outer multi-period sequential decision model is taken as the final unit combination value. The step of updating the inner two-stage robust unit combination model when the outer multi-time-period sequential decision model fails the verification includes: When the slack variable penalty cost does not satisfy the second preset formula, the iteration index between the inner and outer layer models is updated, and the set of unexpected cut plane constraints for the inner layer pre-scheduled unit combination is updated, so as to update the inner layer two-stage robust unit combination model according to the updated iteration index between the inner and outer layer models and the set of unexpected cut plane constraints for the inner layer pre-scheduled unit combination.

2. The method for combining two robust generator units in a power system according to claim 1, characterized in that, The determination of the set of output scenarios for new energy units and the unit combination value of the thermal power units, obtained from the inner-layer two-stage robust unit combination model, further includes: When verifying the optimization model through the inner rescheduling stage, the unit combination value of the pre-scheduling stage is used as the unit combination value of the thermal power unit. If the inner-layer rescheduling stage optimization model fails the verification, the inner-layer pre-scheduling stage optimization model is updated. The steps of obtaining the set of output scenarios and unit combination values ​​of new energy units in the pre-scheduling stage based on the updated inner-layer pre-scheduling stage optimization model are iteratively executed, using the parameters of the power system, the set of output scenarios of the new energy units, the set of indexes of the scenario sequences of the new energy units, the iteration index between the inner and outer layers, the stage iteration index of the inner-layer model, and the set of unexpected cutting plane constraints for the inner-layer pre-scheduling unit combination. This process continues until the inner-layer rescheduling stage optimization model passes the verification. Then, the unit combination value of the pre-scheduling stage that passes the verification is used as the unit combination value of the thermal power unit.

3. The method for combining two robust generator units in a power system according to claim 2, characterized in that, The step of verifying the unit combination value in the pre-scheduling phase based on the set of output scenarios of the new energy units in the pre-scheduling phase and the inner-layer rescheduling phase optimization model includes: The optimal solution for the highest daily operating cost auxiliary variable of the inner pre-schedule stage optimization model is obtained based on the set of output scenarios of the new energy units in the pre-schedule stage. When the optimal solution of the auxiliary variable of the highest operating cost within the day and the objective function value of the inner rescheduling stage optimization model satisfy the first preset formula, the unit combination value of the pre-scheduling stage passes the verification of the inner rescheduling stage optimization model. When the optimal solution of the auxiliary variable of the highest operating cost within the day does not satisfy the first preset formula with the objective function value of the inner rescheduling stage optimization model, the unit combination value of the pre-scheduling stage fails the verification of the inner rescheduling stage optimization model. The first preset formula is: ， in, The auxiliary variable for the highest intraday operating cost of the optimization model in the inner pre-scheduling phase. The optimal solution; The objective function value of the optimization model in the inner rescheduling phase; This is the preset allowable deviation. This serves as the iteration index for the inner model stage. This serves as the iteration index between the inner and outer layer models. For the pre-scheduling phase, This is the rescheduling phase.

4. The method for combining two robust generator units in a power system according to claim 3, characterized in that, The step of updating the inner pre-scheduling phase optimization model when it fails the verification of the inner rescheduling phase optimization model includes: When the optimal solution of the auxiliary variable of the highest operating cost within the day does not satisfy the first preset formula with the objective function value of the inner rescheduling stage optimization model, the inner model stage iteration index is updated, the output scenario set of the new energy unit and the index set of the scenario sequence of the new energy unit are updated, so as to update the inner pre-scheduling stage optimization model according to the updated inner model stage iteration index, the output scenario set of the new energy unit and the index set of the scenario sequence of the new energy unit.

5. The method for combining two robust generator units in a power system according to claim 1, characterized in that, Determining the decision solution of the outer multi-time-period sequential decision model includes: When the scheduling period meets the preset time period threshold, the decision solution is obtained based on the outer multi-time period sequential decision model; When the scheduled time period does not meet the preset time period threshold, the scheduled time period is updated, and the decision solution is obtained based on the updated scheduled time period and the outer multi-time period sequential decision model.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the two-layer robust unit combination method for the power system as described in any one of claims 1 to 5.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the two-layer robust unit combination method for the power system as described in any one of claims 1 to 5.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the two-layer robust unit combination method for the power system as described in any one of claims 1 to 5.

Images