A method for decentralized robust optimization scheduling of an ac-dc interconnected power system

Abstract

The application discloses a kind of AC-DC interconnected power system dispersion robust optimization scheduling method, comprising: according to DC tie-line operating characteristics, construction operation constraint;And construction upper coordination optimization model and lower robust optimization model: lower robust optimization model is based on target cascade analysis method and establishes double-layer iteration solving framework, then introduce column and constraint generation algorithm is processed and is solved to two-stage robust optimization model of each interconnected subsystem, and the active plan of virtual unit is fed back to upper coordination optimization model;Upper coordination optimization model is according to the active plan of virtual unit that lower robust optimization model optimization obtains, optimizes and updates DC tie-line transmission scheme, and the optimization result is passed back as the reference value of next optimization lower robust optimization model;Upper and lower double-layer model carries out iteration interaction solution, until meeting convergence standard, iteration stops.The application can realize the joint optimization of large-scale AC-DC interconnected system scheduling plan and standby scheme.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid control, in particular to a dispersed robust optimization scheduling method for an AC-DC interconnected power system. BACKGROUND

[0002] A clean power system with a high proportion of wind power has become the common vision of the industry. With the large-scale access of new energy, the local consumption capacity of some regions is limited and cannot completely consume local new energy generation. Through cross-regional transactions by means of cross-regional tie lines, promoting cross-regional consumption of new energy is an important measure to solve the problem of new energy consumption. With the increasing interconnection of different regional power grids and the rapid development of wind power, it is urgent to realize the coordinated optimization of interconnected regional power systems.

[0003] On the one hand, with the increase of interconnection scale, centralized joint scheduling faces the problems of large data dimension and complex and variable model, and is difficult to adapt to the actual operation framework of hierarchical partition and meet the information privacy requirements of different subjects in the market environment.

[0004] On the other hand, the current research on new energy DC cross-regional consumption mainly focuses on the optimization of DC line transmission components, the participation of DC lines in two-region peak shaving, the coordination of DC lines and two regions, etc. There is little research on how to coordinate the influence of multi-type flexible resources in interconnected regions on the cross-regional consumption of new energy in response to the uncertainty of new energy prediction. At the same time, the generation plan and power interaction plan of new energy uncertainty are ignored, which may waste the advantage of cross-regional consumption of new energy in actual operation. SUMMARY

[0005] The purpose of the present application is to provide a dispersed robust optimization scheduling method for an AC-DC interconnected power system, which can solve the above-mentioned problems of the prior art, consider the flexible supply characteristics of power system sources, loads and storages, and establish a power system source, load and storage operation flexibility model, which can effectively improve the system operation flexibility and new energy consumption capacity.

[0006] To achieve the above-mentioned purpose, the present application provides the following scheme:

[0007] A dispersed robust optimization scheduling method for an AC-DC interconnected power system, comprising the following steps:

[0008] S1, constructing operation constraints according to the operation characteristics of the DC tie line;

[0009] S2, constructing an upper-layer coordinated optimization model as a main problem:

[0010] constructing an upper-layer coordinated optimization model according to the operation characteristics of the DC tie line and the operation constraints;

[0011] S3, constructing a lower-layer robust optimization model as a sub-problem:

[0012] Based on the robust optimization theory, each interconnected subsystem establishes a corresponding two-stage robust optimization model considering multiple types of flexible resources. In the first stage, the system comprehensive operation constraints considering the wind power basic prediction scenario and the worst scenario of the second stage optimization feedback are considered, the start-stop plan of thermal power units, the active power plan of various flexible resources and the flexible reserve capacity scheme are jointly optimized, and the active power plan of virtual units is also optimized. In the second stage, based on the wind power uncertainty set and the optimization result of the first stage, the feasibility of the dispatching solution under the worst scenario is checked and analyzed, that is, the correction measures of flexible resources are used to ensure the existence of the real-time dispatching scheme under the worst case. If the check fails, the parameters of the corresponding worst scenario and the corresponding system operation constraint set are fed back to the main problem (i.e. the upper coordination optimization model) for updating, and the next iteration optimization is performed.

[0013] S4, the upper and lower double-layer model optimization solution:

[0014] The lower-layer robust optimization model establishes a double-layer iterative solution framework based on the target cascade analysis method, and then introduces column and constraint generation algorithms to transform and solve the two-stage robust optimization model of each interconnected subsystem, and feeds back the active power plan of the virtual unit to the upper-layer coordination optimization model;

[0015] The upper-layer coordination optimization model optimizes and updates the DC tie-line power transmission scheme according to the active power plan of the virtual unit obtained by the lower-layer robust optimization model, and transmits the optimization result back to the lower-layer robust optimization model as the reference value for the next optimization;

[0016] The upper and lower double-layer models are iteratively solved and interacted until the convergence criteria are met, and the iteration stops.

[0017] Further, in step S1, the operation constraints specifically include:

[0018] 1) Tie-line power interval constraint: the tie-line power transmission of the interconnected region must be operated within the set power interval and cannot exceed the upper and lower power limits, which is expressed as follows:

[0019] (1)

[0020] In the formula, , and are the maximum power transmission of the DC tie-line, the minimum power transmission and the power transmission of the time period t, respectively;

[0021] 2) Tie-line power interval constraint: the tie-line can only adjust in a single direction in each time period, and cannot be adjusted in the opposite direction in adjacent time periods, which is expressed as follows:

[0022] (2)

[0023] In the formula, , and respectively represent the state of the DC tie line adjusting power, increasing power and decreasing power at time period t, and respectively represent the state of the DC tie line increasing power and decreasing power at time period t+1, , , , and are 0-1 integer variables;

[0024] 3) The scheduling day regulation frequency constraint is expressed as follows:

[0025] (3)

[0026] In the formula, φ is the adjustment frequency limit value in the scheduling day, which is adjusted according to actual needs;

[0027] 4) Transmission power regulation rate constraint, the power change of adjacent time periods of the DC tie line is jointly constrained by the regulation state and the regulation rate, there is a coupling relationship between each time period, and the formula is expressed as follows:

[0028] (4)

[0029] In the formula, represents the transmission power at time period t+1, and are the upward and downward power regulation rates of the DC tie line; M is a constant term, which takes the maximum value;

[0030] 5) Transmission power step operation constraint, the DC tie line needs to meet the step constraint when operating, and needs to be stable for a certain period of time after power adjustment, and the formula is expressed as follows:

[0031] (5)

[0032] In the formula, is a state variable representing whether the tie line starts to adjust power, which is a 0-1 integer variable; is a state variable representing whether the tie line stops power adjustment, which is a 0-1 integer variable; is the interval time of adjustment action; T is the total number of scheduling time periods, and τ is an auxiliary variable;

[0033] 6) Transmission power deviation constraint, the total power delivered in each period of the day needs to meet the predetermined power range, and the power deviation coefficient is set to adjust the power, which is expressed as follows:

[0034] (6)

[0035] In the formula, is the transmission power signed in advance; and p is the power deviation coefficient.

[0036] Further, in step S2, the upper coordination optimization model is as follows:

[0037] (7)

[0038]

[0039] In the formula, and are the active power of the virtual unit m and n optimized by the lower robust optimization model, respectively; , are the penalty coefficients of the linear term and the quadratic term of the virtual unit m, respectively, for optimizing the tie-line power; and are the penalty coefficients of the linear term and the quadratic term of the virtual unit n, respectively, for optimizing the tie-line power; is the set of DC tie lines;

[0040] Further, in step S3, based on the robust optimization theory, each interconnected subsystem establishes a corresponding two-stage robust optimization model considering multiple types of flexible resources, the first stage considers the wind power basic prediction scenario and the second stage optimizes the system comprehensive operation constraint of the worst wind power scenario, jointly optimizes the start-stop plan of the thermal power unit, the active plan of each type of flexible resource and the flexible reserve capacity scheme, and optimizes the active plan of the virtual unit, specifically including:

[0041] S301, establishing an objective function: the objective function is as shown in formula (8), which contains four parts, wherein the first part, is the operation cost of the thermal power unit, including the power generation cost, the start-up cost, the shutdown cost and the flexible reserve capacity cost; the second part, is the operation cost of the energy storage system, including the energy storage charging and discharging cost and the flexible reserve capacity cost; the third part, is the flexible reserve capacity cost of demand response; and the fourth part, is the augmented Lagrange penalty function composed of the shared variables between interconnected systems, i.e., the tie-line power, which contains the tie-line power and the power variable of the virtual unit in the region obtained by the upper optimization;

[0042] (8)

[0043] where, , and are the planned output, start-up cost and shut-down cost of thermal generating unit i at time period t, respectively; and are the upward and downward spinning reserve capacity of conventional generating unit i reserved at time period t, respectively, and are the corresponding capacity cost, respectively; and are the charging and discharging power of energy storage system s, respectively, and are the corresponding operation cost, respectively; and are the upward and downward spinning reserve capacity of energy storage system s reserved at time period t, respectively; and are the capacity cost and interruptible reserve capacity of demand response k at time period t, respectively; T is the total number of scheduling time periods; are the sets of thermal generating units, energy storages and interruptible loads, respectively;

[0044] S302, constructing operation constraints under the basic scenario, including:

[0045] 1) thermal generating unit operation constraints, as shown in equations (9)-(20), wherein equations (9)-(11) are minimum operation duration constraints of thermal generating units, equations (12)-(14) are minimum shut-down duration constraints of thermal generating units, equations (15) and (16) represent start-up cost and shut-down cost of thermal generating units, respectively, equations (17) and (18) represent upper bound and lower bound constraints of active power output of thermal generating units, respectively, while considering the coupling relationship between spinning reserve capacity and active power output, equations (19) and (20) represent upward ramping rate constraint and downward ramping rate constraint of thermal generating units, respectively, while considering the adjacent time period restraint coupling relationship when the spinning reserve capacity is called;

[0046] (9)

[0047] (10)

[0048] (11)

[0049] (12)

[0050] (13)

[0051] (14)

[0052] (15)

[0053] (16)

[0054] (17)

[0055] (18)

[0056] (19)

[0057] (20)

[0058] In formula (9) to formula (20), is the operating state of the conventional unit i at time period t; and are the minimum start-up operating time and the minimum shutdown duration of the conventional unit i, respectively; and are the start-up cost and the shutdown cost of the conventional unit i, respectively; and are the minimum active power and the maximum active power of the conventional unit i, respectively; and are the upward ramping capability and the downward ramping capability of the conventional unit i, respectively;

[0059] 2) Energy storage system operating constraints, as shown in formula (21) to formula (35), wherein formula (21) represents the state of charge of the energy storage system at each time period, i.e., the storage capacity after charging and discharging at the time period, formula (22) is the upper and lower bound constraint of the storage capacity; formula (23) represents that the initial capacity of the energy storage system is consistent with the final capacity, formula (24) represents the power output of the energy storage, formula (25) and formula (26) represent the charging power and discharging power limit constraints of the energy storage system, respectively, formula (27) represents the operating state constraint of the energy storage system, i.e., charging and discharging cannot be performed at the same time at the same time period, formula (28) and formula (29) are the upward flexible reserve capacity supply capability constraints of the energy storage system in the charging state and the discharging state, respectively, formula (30) and formula (31) are the downward flexible reserve capacity supply capability constraints of the energy storage system in the charging state and the discharging state, respectively, formula (32) and formula (33) represent the total upward and downward flexible reserve capacity of the energy storage system, respectively, formula (34) and formula (35) are the energy lower bound and upper bound constraints of the energy storage system considering the flexible reserve capacity calling situation, respectively;

[0060] (21)

[0061] (22)

[0062] (23)

[0063] (24)

[0064] (25)

[0065] (26)

[0066] (27)

[0067] (28)

[0068] (29)

[0069] (30)

[0070] (31)

[0071] (32)

[0072] (33)

[0073] (34)

[0074] (35)

[0075] In formula (21) to formula (35), is a storage power value of the energy storage s at time period t; , and are an initial power value, an upper limit of power storage, a lower limit of power storage, a charging efficiency and a discharging efficiency of the energy storage s respectively; and are charging and discharging power limits of the energy storage s respectively; and are upward flexible reserve capacities of the energy storage s in charging and discharging states at time period t respectively; and are downward flexible reserve capacities of the energy storage s in charging and discharging states at time period t respectively;

[0076] 3) Demand response operation constraints, as shown in equations (36)-(40), wherein equation (36) is a load reduction capacity constraint of incentive-based demand response, equations (37) and (38) are minimum and maximum continuous time constraints of incentive-based demand response that can be continuously reduced, equation (39) is an interval time constraint of adjacent two load reduction actions, and equation (40) is a constraint of limiting the number of calls of scheduling intra-day demand response to avoid frequent calls affecting normal load power consumption;

[0077] (36)

[0078] (37)

[0079] (38)

[0080] (39)

[0081] (40)

[0082] In equations (36)-(40), is a state of interruptible load k participating in reserve supply in period t, and takes a value of 1 to indicate participation and a value of 0 to indicate non-participation; is the cumulative interruption time of interruptible load k in period t-1; , are the minimum and maximum interruption times of interruptible load k, respectively; is the cumulative non-interruption time of interruptible load k in period t-1; is the minimum interruption interval time of interruptible load k; is the maximum interruption number of interruptible load k;

[0083] 4) System operation constraints, including power balance constraints and power flow constraints, wherein equation (41) is a system power balance constraint, and equation (42) is a system line power flow constraint;

[0084] (41)

[0085] (42)

[0086] In equations (41) and (42), and are power flow distribution transfer factors of conventional unit i, wind farm w, energy storage s, node load j, and virtual unit m to grid line l, respectively; is the power transmission limit of grid line l;

[0087] S303, wind power uncertainty set: the wind power uncertainty in the region is characterized by the uncertainty set constructed by each region, as shown in equation (43);

[0088] (43)

[0089] wherein, and are the actual power and predicted power of wind farm w at time period t, respectively; and are the upper and lower bounds of the active power output prediction error of wind farm w at time period t, respectively; and are 0-1 variables representing the positive and negative prediction power deviation of wind farm w, and are optimization variables; is the uncertainty budget, which is used to adjust the robustness of the scheduling strategy, and can be determined according to the confidence level a according to equation (44);

[0090] (44)

[0091] S304, real-time operation constraints under uncertainty, the real-time stage rescheduling plan needs to ensure the feasibility under any wind power scenario by calling flexible resources, the rescheduling model in the real-time operation stage is shown in equations (45)-(50), wherein equation (45) is the second stage objective function, including the cost of curtailment and the cost of load shedding, equation (46) is the system power balance constraint considering the curtailment variable and the load shedding variable, equations (47) and (48) respectively represent the active power output constraints of thermal power units and energy storage systems considering the calling of flexible reserve capacity, equation (49) is the regulation capacity constraint of incentive demand response, and equation (50) is the line power flow constraint considering wind power uncertainty;

[0092] (45)

[0093] (46)

[0094] (47)

[0095] (48)

[0096] (49)

[0097] (50)

[0098] In equations (45)-(50), and are the curtailment and load shedding power at time period t under random error scenario, respectively; and respectively are the penalty cost of wind curtailment and load shedding.

[0099] Further, in step S4, the upper and lower double-layer model is iteratively solved until the convergence criterion is met, and the iteration stops, specifically including:

[0100] S401, convergence criterion and penalty coefficient updating principle:

[0101] The iteration convergence criterion of the distributed solution algorithm is shown in equation (51). When the optimization result meets the iteration convergence requirement, the scheduling plan of the interconnected region and the DC tie-line transmission plan are obtained;

[0102] (51)

[0103] In the formula, ε is the convergence threshold;

[0104] If the iteration optimization result does not meet the convergence criterion, the penalty coefficient is updated using equation (52) for the next iteration optimization,

[0105] (52)

[0106] In the formula, γ is a constant term;

[0107] S402, iteration solving process:

[0108] The optimization solving framework combining the target cascade analysis method, column and constraint generation is used for iteration solving, and the iteration interaction of the upper and lower double-layer optimization problems is realized by using the updating mechanism of the DC tie-line transmission plan. The specific solving process of the proposed distributed optimization algorithm is as follows:

[0109] 1) Initialize the iteration index , the penalty coefficient and the DC tie-line transmission power plan;

[0110] 2) Use the column and constraint generation algorithm to solve the lower optimization sub-problem of each regional power grid composed of equations (1)-(6) and equations (8)-(50), determine the regional power grid scheduling plan and the output plan of the virtual unit, and feed back the output information of the virtual unit to the upper master problem for coordinated optimization.

[0111] 3) Solve the upper master problem composed of equations (1)-(7) to optimize the DC tie-line transmission power plan, and send it as a reference value to each lower optimization sub-problem;

[0112] 4) If the optimization result of the upper and lower double-layer problem meets the convergence condition formula (51), the iteration is terminated and the scheduling plan of each region and the transmission plan of the direct current tie line are obtained, otherwise the penalty coefficient of each region is updated and the next iteration is returned to step 2).

[0113] Further, the method further comprises: step S5, based on the lower-layer robust optimization model, the upper-layer coordinated optimization model and the optimization solving algorithm proposed in step S4, solving by using a Yalmip tool box combined with an optimization software CPLEX to obtain the running state of the i th thermal power unit in the interconnected region at the t th time period , the planned active power output , the flexible reserve capacity , the charge and discharge state of the s th energy storage device at the t th time period and the charge and discharge power , the flexible reserve capacity , the calling state of the k th interruptible load at the t th time period and the capacity and the tie line power plan , realizing the decentralized robust optimization scheduling of the AC-DC interconnected power system.

[0114] According to the specific embodiments provided by the application, the following technical effects are disclosed: the AC-DC interconnected power system decentralized robust optimization scheduling method provided by the application can realize the joint optimization of the scheduling plan and the reserve scheme of the large-scale AC-DC interconnected system, the method decomposes the multi-region power system scheduling problem into an optimization problem with a double-layer structure by using a target cascade analysis method, wherein the upper-layer main problem is responsible for coordinating and updating the power transmission plan of the direct current tie line of the interconnected region power system, and the lower-layer regional sub-problem constructs a robust unit combination model to formulate the scheduling plan of each region, and the joint optimization of the active power plan, the flexible reserve scheme and the tie line plan of the multi-region system is realized through the iteration interaction of the upper and lower double layers. In order to improve the system operation flexibility to cope with wind power uncertainty, the flexible reserve supply characteristics of multiple types of flexible resources and their time coupling are considered, and a double-layer optimization solving method based on the target cascade analysis method is proposed, and the column and constraint generation algorithm is used to solve the lower-layer problem. The application considers the connection effect and flexible adjustment characteristics of the direct current tie line and the operation characteristics of the flexible resources (such as thermal power units, energy storage systems and demand response) of the interconnected region system, and can realize the coordinated optimization of the flexible resources from the entire system range, thereby providing more reasonable scheduling plans for the scheduling and operation personnel. BRIEF DESCRIPTION OF DRAWINGS

[0115] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0116] Figure 1 This is a flowchart illustrating the distributed robust optimization scheduling method for AC / DC interconnected power systems of the present invention.

[0117] Figure 2 This is a flowchart illustrating the iterative solution process of the distributed optimization algorithm in an embodiment of the present invention. Detailed Implementation

[0118] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0119] The purpose of this invention is to provide a decentralized robust optimization scheduling method for AC / DC interconnected power systems. Considering the flexible supply characteristics of power system sources, loads, and storage, a power system source-load-storage operation flexibility model is established, which can effectively improve the system's operational flexibility and renewable energy absorption capacity.

[0120] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0121] like Figure 1 As shown in the figure, the distributed robust optimization scheduling method for AC / DC interconnected power systems provided in this embodiment of the invention includes the following steps:

[0122] S1, construct operating constraints based on the operating characteristics of the DC tie line;

[0123] S2, Constructing a higher-level coordination optimization model as the main problem:

[0124] Based on the operating characteristics and constraints of the DC tie line, an upper-level coordination and optimization model is constructed.

[0125] S3, constructing a lower-level robust optimization model as a sub-problem:

[0126] Based on the robust optimization theory, the two-stage robust optimization model considering multiple types of flexible resources is established for each interconnected subsystem. In the first stage, the system comprehensive operation constraints considering the wind power basic prediction scenario and the worst scenario of the second stage optimization feedback are considered, the joint optimization of the start-stop plan of thermal power units, the active power plan of various flexible resources and the flexible reserve capacity scheme is carried out, and the active power plan of the virtual unit is optimized; in the second stage, based on the wind power uncertainty set and the optimization result of the first stage, the feasibility of the dispatching solution under the worst scenario is checked and analyzed, that is, the correction measures of the flexible resources are used to ensure the existence of the real-time dispatching scheme under the worst scenario, if the check fails, the parameters of the corresponding worst scenario and the corresponding system operation constraint set are fed back to the main problem for updating, and the next iteration optimization is carried out;

[0127] S4, the upper and lower double-layer model optimization solution:

[0128] The lower-layer robust optimization model establishes a double-layer iterative solution framework based on the target cascade analysis method, then introduces column and constraint generation algorithm to transform and solve the two-stage robust optimization model of each interconnected subsystem, and feeds back the active power plan of the virtual unit to the upper-layer coordinated optimization model;

[0129] The upper-layer coordinated optimization model optimizes and updates the DC tie-line power transmission scheme according to the active power plan of the virtual unit obtained by the lower-layer robust optimization model, and transmits the optimization result back to the lower-layer robust optimization model as the reference value for the next optimization;

[0130] The upper and lower double-layer models are iteratively solved and interacted until the convergence criteria are met and the iteration stops.

[0131] In step S1, the DC tie-line can continuously and flexibly adjust the power transmission, and can be used as a flexible resource to promote the effective coordination of the interconnected power system. The operation constraints of the DC tie-line are considered in the dispatching model to avoid frequent fluctuations of the power and ensure the operation safety of the line and the interconnected system. The operation characteristics of the DC tie-line are modeled as shown in formulas (1) to (6), which specifically include:

[0132] 1) Tie-line power interval constraint: the tie-line power of the interconnected region must be operated within the set power interval, and cannot exceed the upper and lower power limits, which is expressed as follows:

[0133] (1)

[0134] In the formula, 、 and are the maximum transmission power, the minimum transmission power and the transmission power of the time period t of the DC tie-line, respectively.

[0135] 2) Tie-line power interval constraint, tie-line can only adjust to a single direction in each time interval, and the adjacent time interval cannot be adjusted in the opposite direction, which is expressed as follows:

[0136] (2)

[0137] wherein, , and represent the state of the DC tie-line power adjustment, power increase and power decrease in time interval t, and represent the state of power increase and power decrease in time interval t+1, , , , and are 0-1 integer variables;

[0138] 3) Scheduling intra-day adjustment times constraint, which is expressed as follows:

[0139] (3)

[0140] wherein, φ is the adjustment times limit value in the scheduling day, to avoid frequent adjustment of DC tie-line power transmission in a single scheduling day, the adjustment times limit in the day is set, which can be adjusted according to actual needs;

[0141] 4) Transmission power adjustment rate constraint, the power change of adjacent time intervals of the DC tie-line is jointly constrained by the adjustment state and the adjustment rate, there is a coupling relationship between each time interval, which is expressed as follows:

[0142] (4)

[0143] wherein, represents the transmission power in time interval t+1, and are the upward and downward power adjustment rates of the DC tie-line; M is a constant term, which takes the maximum value;

[0144] 5) Transmission power step operation constraint, the DC tie-line needs to meet the step constraint when it is running, and it needs to be stable for a certain period of time after power adjustment, which is expressed as follows:

[0145] (5)

[0146] wherein, is a state variable representing whether the tie-line starts to adjust the power, which is a 0-1 integer variable; is a state variable representing whether the tie-line stops power adjustment, which is a 0-1 integer variable; to adjust the interval time of the action; T is the total time period of scheduling, and τ is an auxiliary variable;

[0147] 6) Transmission power deviation constraint: the total power delivered in each period of the day should meet the predetermined power range. The power deviation coefficient is set to adjust the power, which is expressed as follows:

[0148] (6)

[0149] In the formula, is the power delivered in advance; and ρ is the power deviation coefficient.

[0150] In the upper-layer coordination optimization model, the upper-layer control center does not need to know the detailed operation parameters of the lower-layer regional power system, but only needs to know the operation information of the DC tie line, thereby avoiding a large amount of information transmission and processing. The upper-layer optimization main problem is responsible for optimizing the power transmission plan of the DC tie line and issuing it to the lower-layer regional optimization sub-problem as a reference value, that is, the upper-layer coordination problem optimizes the virtual unit active power plan obtained by the lower-layer problem, optimizes and updates the DC tie line power transmission scheme, and returns the optimization result to the lower-layer control center as a reference value. In step S2, the upper-layer coordination optimization model is as follows:

[0151] (7)

[0152] In the formula, and are the active power outputs of the virtual units m and n optimized by the lower-layer robust optimization model, respectively; , are the penalty coefficients of the linear term and the quadratic term of the virtual unit m, respectively, for optimizing the tie line power; and are the penalty coefficients of the linear term and the quadratic term of the virtual unit n, respectively, for optimizing the tie line power; is the set of DC tie lines.

[0153] In step S3, each interconnected subsystem establishes a corresponding two-stage robust optimization model considering multiple types of flexible resources based on robust optimization theory. In the first stage, considering the system comprehensive operation constraints of wind power basic prediction scenarios and the worst-case scenarios of wind power optimization feedback in the second stage, the joint optimization of thermal power unit start-stop plan, active plan of various flexible resources and flexible reserve capacity scheme is performed, and the active plan of virtual units is also optimized. In the second stage, based on the wind power uncertainty set and the optimization result of the first stage, the feasibility of the dispatching solution under the worst-case scenario is checked and analyzed, that is, the correction measures of flexible resources are used to ensure the existence of the real-time scheduling scheme under the worst-case scenario. If the check fails, the parameters of the corresponding worst-case scenario and the corresponding system operation constraint set are fed back to the main problem for updating, and the next iteration optimization is performed. The lower two-stage robust optimization model is as follows, which specifically includes:

[0154] S301, establishing an objective function: the objective function is shown as formula (8), which includes four parts, wherein the first part, is the operation cost of the thermal power unit, including generation cost, start-up cost, shutdown cost and flexible reserve capacity cost; the second part, is the operation cost of the energy storage system, including energy storage charging and discharging cost and flexible reserve capacity cost; the third part, is the flexible reserve capacity cost of demand response; the fourth part, is the augmented Lagrange penalty function composed of shared variables between interconnected systems, that is, tie line power and power variables of virtual units in the region;

[0155] (8)

[0156] In the formula, , and are the planned output, start-up cost and shutdown cost of the thermal power unit i at time period t; and are the upward and downward flexible reserve capacities reserved by the conventional unit i at time period t, and are the corresponding capacity costs; and are the charging and discharging power of the energy storage system s, and are the corresponding operation costs; and are the upward and downward flexible reserve capacities reserved by the energy storage system s at time period t; and are the capacity cost and interruptible reserve capacity of demand response k at time period t; T is the total number of scheduling periods; a set of thermal power units, energy storage, interruptible load, respectively;

[0157] S302, constructing operation constraints under the basic scenario, including:

[0158] 1) thermal power unit operation constraints, as shown in formulas (9)-(20), wherein formulas (9)-(11) are minimum operation duration constraints of the thermal power unit, formulas (12)-(14) are minimum shutdown duration constraints of the thermal power unit, formulas (15) and (16) represent startup cost and shutdown cost of the thermal power unit respectively, formulas (17) and (18) represent upper bound constraint and lower bound constraint of active power output of the thermal power unit respectively, while considering the coupling relationship between flexible reserve capacity and active power output, formulas (19) and (20) represent upward ramp rate constraint and downward ramp rate constraint of the thermal power unit respectively, while considering the adjacent period restraint coupling relationship when the flexible reserve capacity is called;

[0159] (9)

[0160] (10)

[0161] (11)

[0162] (12)

[0163] (13)

[0164] (14)

[0165] (15)

[0166] (16)

[0167] (17)

[0168] (18)

[0169] (19)

[0170] (20)

[0171] In formulas (9)-(20), is the operation state of the conventional unit i at the time period t; and are the minimum startup operation time and the minimum shutdown duration of the conventional unit i, respectively; and respectively the start-up cost and shut-down cost of conventional unit i; and respectively the minimum and maximum active power output of conventional unit i; and respectively the upward ramping capability and downward ramping capability of conventional unit i;

[0172] 2) the operation constraints of energy storage system, as shown in equations (21)-(35), wherein equation (21) represents the state of charge of energy storage system in each time interval, i.e. the storage capacity after charging and discharging in the time interval, equation (22) represents the upper and lower bound constraints of storage capacity; equation (23) represents that the initial capacity of energy storage system is consistent with the final capacity, equation (24) represents the power output of energy storage, equations (25) and (26) represent the charging power and discharging power limit constraints of energy storage system respectively, equation (27) represents the operation state constraint of energy storage system, i.e. charging and discharging cannot be performed simultaneously in the same time interval, equations (28) and (29) represent the upward flexible reserve capacity supply capability constraints of energy storage system in charging state and discharging state respectively, equations (30) and (31) represent the downward flexible reserve capacity supply capability constraints of energy storage system in charging state and discharging state respectively, equations (32) and (33) represent the total upward and downward flexible reserve capacity of energy storage system respectively, and equations (34) and (35) represent the lower and upper bound constraints of energy storage system considering the calling situation of flexible reserve capacity;

[0173] (21)

[0174] (22)

[0175] (23)

[0176] (24)

[0177] (25)

[0178] (26)

[0179] (27)

[0180] (28)

[0181] (29)

[0182] (30)

[0183] (31)

[0184] (32)

[0185] (33)

[0186] (34)

[0187] (35)

[0188] In formula (21) to formula (35), is the storage capacity value of the energy storage s at time period t; , and are the initial capacity value, the upper limit of capacity storage, the lower limit of capacity storage, the charging efficiency and the discharging efficiency of the energy storage s respectively; and are the charging and discharging power limits of the energy storage s respectively; and are the upward flexible reserve capacity of the energy storage s in the charging and discharging state at time period t respectively; and are the downward flexible reserve capacity of the energy storage s in the charging and discharging state at time period t respectively;

[0189] 3) Demand response operation constraints, as shown in formula (36) to formula (40), considering incentive-based demand response, power users participating in incentive-based demand response can sign a contract in advance, respond to power dispatching instructions, and provide certain flexible reserve capacity, wherein formula (36) is the load reduction capacity constraint of incentive-based demand response, formula (37) and formula (38) are the minimum duration and maximum duration constraints of continuous reduction of incentive-based demand response respectively, formula (39) is the interval time constraint of adjacent two load reduction actions, formula (40) is the calling frequency limit constraint of demand response in the dispatching day, to avoid frequent calling affecting normal load of the load;

[0190] (36)

[0191] (37)

[0192] (38)

[0193] (39)

[0194] (40)

[0195] In formula (36) to formula (40), is the state of interruptible load k participating in reserve supply in time period t, taking value 1 to represent participation and value 0 to represent non-participation; is the cumulative interruption time of interruptible load k in time period t-1; , are the minimum interruption time and the maximum interruption time of interruptible load k, respectively; is the cumulative non-interruption time of interruptible load k in time period t-1; is the minimum interruption interval time of interruptible load k; is the maximum interruption number of interruptible load k;

[0196] 4) System operation constraints, including power balance constraints and power flow constraints, wherein equation (41) is the system power balance constraint and equation (42) is the system line power flow constraint;

[0197] (41)

[0198] (42)

[0199] In the formula, and are the power flow distribution transfer factors of conventional unit i, wind farm w, energy storage s, node load j and virtual unit m to network line l, respectively; is the power transmission limit of network line l;

[0200] S303, wind power uncertainty set: for each regional optimization sub-problem, the optimization objective is to determine the scheduling scheme under the wind power basic prediction scenario, while ensuring the scheduling feasibility under the worst wind power scenario. Considering the spatial distance of wind farms in each regional power system, the invention does not consider the correlation of wind farm output of interconnected regional systems, and the wind power uncertainty in the region is represented by the uncertainty set constructed by each region. The regional wind power uncertainty set is shown in equation (43).

[0201] (43)

[0202] In the formula, and are the actual power and predicted power of wind farm w in time period t, respectively; and are the upper and lower bounds of the active power output prediction error of wind farm w in time period t, respectively; and are 0-1 variables representing positive and negative prediction power deviation of wind farm, and are optimization variables; is the uncertainty budget, which is used to adjust the robustness of the scheduling strategy, and can be determined according to equation (44) by confidence level a calculation;

[0203] (44)

[0204] S304, real-time operation constraints under uncertain scenarios, the influence of the random characteristics of wind power on the real-time operation stage needs to be considered when determining the day-ahead scheduling plan, therefore, the real-time operation constraints are considered in the second stage. The real-time stage rescheduling plan needs to ensure the feasibility under any wind power scenario by calling flexible resources, the rescheduling model of the real-time operation stage is shown in equations (45)-(50), wherein equation (45) is the second stage objective function, including the wind curtailment cost and the load shedding cost, equation (46) is the system power balance constraint considering the wind curtailment variable and the load shedding variable, equations (47) and (48) respectively represent the active power output constraints of the thermal power unit and the energy storage system considering the calling of flexible reserve capacity, equation (49) is the regulation capacity constraint of the incentive demand response, and equation (50) is the line power flow constraint considering the uncertainty of wind power;

[0205] (45)

[0206] (46)

[0207] (47)

[0208] (48)

[0209] (49)

[0210] (50)

[0211] In equations (45)-(50), and are the wind curtailment and load shedding power in period t under random error scenarios, respectively; and are the wind curtailment and load shedding penalty costs, respectively.

[0212] Step S4 proposes a decentralized solution algorithm based on the target cascade method, column and constraint generation method for the constructed hierarchical optimization model. First, a double-layer iterative solution framework based on the target cascade analysis method is proposed, and then the column and constraint generation algorithm is introduced to transform and solve the two-stage robust optimization problem of each interconnected regional system. The proposed decentralized solution algorithm is based on the modeling idea of the target cascade analysis method, in which the upper regulatory department plays a coordinating role. Based on the scheduling plan of the virtual unit obtained by sub-problem optimization, the transmission plan of the direct current tie line is coordinated and optimized, and sent to the lower regional regulatory department as the reference value for the next optimization. The lower regional optimization problem uses the column and constraint generation method to solve the two-stage robust optimization model of each region, and feeds back the output plan of the virtual unit to the upper regulatory department. The upper and lower double-layer models are iteratively solved until the convergence criteria are met, and the iteration stops. For ease of analysis, it is considered that the direct current tie line transmission plan remains unchanged within the dispatching day and is not adjusted.

[0213] In step S4, the upper and lower double-layer models are iteratively solved until the convergence criteria are met, and the iteration stops, specifically including:

[0214] S401, convergence criteria and penalty coefficient updating principle:

[0215] The iterative convergence criteria of the decentralized solution algorithm is shown in equation (51). When the optimization result meets the iterative convergence requirement, the scheduling plan of the interconnected region and the transmission plan of the direct current tie line are obtained;

[0216] (51)

[0217] In the formula, ε is the convergence threshold;

[0218] If the iterative optimization result does not meet the convergence criteria, the penalty coefficient is updated using equation (52) for the next iteration optimization,

[0219] (52)

[0220] In the formula, γ is a constant term;

[0221] S402, the iterative solution process is as shown in Figure 2 :

[0222] The proposed decentralized optimization algorithm combines the target cascade analysis method, column and constraint generation optimization solution framework for iterative solution, and uses the update mechanism of the direct current tie line transmission plan to realize the iterative interaction of the upper and lower double-layer optimization problems. The specific solution process of the proposed decentralized optimization algorithm is as follows:

[0223] 1) Initialize the iteration index , the penalty coefficient and the DC tie-line transmission power plan;

[0224] 2) solving the lower-level optimization sub-problems of each regional power grid constituted by formula (1) to formula (6) and formula (8) to formula (50) by using the column and constraint generation algorithm, determining the regional power grid scheduling plan and the output plan of the virtual unit, and feeding back the output information of the virtual unit to the upper-level main problem for coordinated optimization.

[0225] 3) solving the upper-level main problem constituted by formula (1) to formula (7), optimizing the DC tie-line transmission power plan, and issuing the DC tie-line transmission power plan as a reference value to each lower-level optimization sub-problem;

[0226] 4) if the optimization results of the upper and lower double-layer problems meet the convergence condition formula (51), the iteration is terminated and the scheduling plan of each region and the DC tie-line transmission plan are obtained, otherwise the penalty coefficient of each region is updated and the next iteration is returned to step 2).

[0227] In addition, the method further comprises: step S5, based on the lower-level robust optimization model, the upper-level coordinated optimization model and the optimization solving algorithm proposed in step S4, solving by using the Yalmip toolbox combined with the optimization software CPLEX to obtain the running state of the i th thermal power unit of the interconnected region in the t th time period , the planned active power output , the flexible reserve capacity , the charging and discharging state of the s th energy storage device in the t th time period and the charging and discharging power , the flexible reserve capacity the calling state of the k th interruptible load in the t th time period and the capacity and the tie-line power plan , realizing the decentralized robust optimization scheduling of the AC-DC interconnected power system.

[0228] To sum up, the present application establishes a decentralized robust optimization scheduling model of the AC-DC interconnected power system, the upper-level main problem of which is used for coordinating the power transmission plan of the DC tie-line, and the lower-level sub-problem of the regional system is used for optimizing the robust scheduling scheme of each region and obtaining the active power plan of the virtual unit of the tie-line, realizing the coordinated optimization of multiple regions and ensuring the information privacy. The decentralized robust optimization algorithm proposed in the present application constructs the upper and lower double-layer problems in the framework of the target cascade analysis method, performs iterative optimization, and solves the two-stage robust optimization model of the lower layer by using the column and constraint generation algorithm, which can provide an economic and reliable scheduling scheme for the AC-DC interconnected system and has good scalability, and the algorithm has good convergence and calculation efficiency.

[0229] The principles and implementation manners of the present application are described by using specific examples in the present application, and the above examples are only used to help understand the method of the present application and its core idea; meanwhile, for the general technical personnel in the art, the specific implementation manners and application ranges will be changed according to the idea of the present application. In conclusion, the content of the present specification should not be understood as the limitation of the present application.

Claims

1. A distributed robust optimization scheduling method for AC / DC interconnected power systems, characterized in that, Includes the following steps: S1, construct operating constraints based on the operating characteristics of the DC tie line; S2, Constructing a higher-level coordination optimization model as the main problem: Based on the operating characteristics and constraints of the DC tie line, an upper-level coordination and optimization model is constructed. S3, constructing a lower-level robust optimization model as a sub-problem, specifically includes: S301, Establish the objective function: The objective function is shown in equation (8), which consists of four parts, of which the first part, The first part covers the operating costs of thermal power units, including power generation costs, start-up costs, shutdown costs, and flexible reserve capacity costs; the second part... The operating costs of the energy storage system include the costs of charging and discharging energy and the costs of flexible backup capacity; Part Three, Cost of flexible standby capacity for demand response; Part Four, The augmented Lagrange penalty function is constructed for the shared variables between interconnected systems, namely tie-line power, which includes the tie-line power obtained from the upper-level optimization and the power variables of the virtual machine group in this region; （8） In the formula, , and thermal power units i During the period t The planned output, start-up costs, and downtime costs; and They are conventional units i During the period t Reserved flexible backup capacity both upwards and downwards. and These are the corresponding capacity costs; and energy storage system s The charging and discharging power, and These are the corresponding operating costs; and energy storage system s During the period t Reserved flexible backup capacity for both upward and downward travel; and Demand response k Capacity cost and time period t Interruptible backup capacity; T This represents the total number of scheduling periods; These are collections of thermal power units, energy storage, and interruptible loads. S302, establish the operational constraints for the basic scenario, including: 1) Operating constraints of thermal power units; 2) Energy storage system operation constraints; 3) Demand response operational constraints; 4) System operational constraints; S303, Wind power uncertainty set: The uncertainty of wind power in the region is characterized by the uncertainty set constructed by each region, as shown in equation (43); （43） In the formula, and Time periods t wind farm w Actual power and predicted power; and Time periods t wind farm w The upper bound and lower bound of the active power output prediction error; and These are 0-1 variables representing the positive and negative predicted power deviations of wind farms, respectively, and are optimization variables; It is an uncertainty budget used to adjust the robustness of the scheduling strategy, which can be determined by the confidence level according to equation (44). α Calculation determined; （44） S304, Real-time operation constraints under uncertain scenarios. The real-time stage rescheduling plan needs to call flexible resources to ensure the feasibility of any wind power scenario. The rescheduling model of the real-time operation stage is shown in Equations (45) to (50). Equation (45) is the objective function of the second stage, including wind curtailment cost and load shedding cost. Equation (46) is the system power balance constraint considering wind curtailment variables and load shedding variables. Equations (47) and (48) respectively represent the active power output constraints of thermal power units and energy storage systems after considering the call of flexible reserve capacity. Equation (49) is the adjustment capacity constraint of incentive-type demand response. Equation (50) is the line power flow constraint considering the uncertainty of wind power. （45） （46） （47） （48） （49） （50） In equations (45) to (50), and These are time periods under random error scenarios. t The power curtailment and load shedding; and These are the penalties for wind curtailment and load shedding, respectively. S4, optimization solution for the upper and lower two-layer model: The lower-level robust optimization model establishes a two-level iterative solution framework based on the objective cascade analysis method. Then, a column and constraint generation algorithm is introduced to transform and solve the two-stage robust optimization model of each interconnected subsystem, and the active power plan of the virtual machine group is fed back to the upper-level coordinated optimization model. The upper-level coordinated optimization model optimizes and updates the DC tie-line transmission scheme based on the virtual machine group active power plan optimized by the lower-level robust optimization model, and transmits the optimization results back to the lower-level robust optimization model as reference values ​​for the next optimization. The upper and lower two-layer models are used for iterative interactive solution until the convergence criterion is met, at which point the iteration stops.

2. The distributed robust optimization scheduling method for AC / DC interconnected power systems according to claim 1, characterized in that, In step S1, the operational constraints specifically include: 1) Tie-line power range constraints: The transmission power of tie-lines in interconnected areas must operate within a set power range and cannot exceed the upper and lower power limits. The formula is as follows: （1） In the formula, , and These represent the maximum transmission power, minimum transmission power, and time period of the DC tie line. t The transmission power; 2) Tie line power range constraint: The tie line can only be adjusted in one direction during each time period, and cannot be adjusted in opposite directions between adjacent time periods. The formula is as follows: （2） In the formula, , and These represent the DC tie lines during the time period. t The states of adjusting power, increasing power, and decreasing power. and These represent the DC tie lines during the time period. t +1 indicates states of increasing and decreasing power. , , , as well as All are integer variables between 0 and 1; 3) Constraints on the number of intraday adjustments during scheduling, expressed by the following formula: （3） In the formula, The limit for the number of adjustments within a scheduling day is set and adjusted according to actual needs. 4) Transmission power regulation rate constraint: The power change of adjacent time periods of the DC tie line is constrained by both the regulation state and the regulation rate. There is a coupling relationship between the time periods, which is expressed by the following formula: （4） In the formula, Indicates time period t +1 transmission power, and For the upward and downward power regulation rates of the DC tie line; M For the constant term, take the maximum value; 5) Stepped power transmission constraints: DC tie lines must meet stepped constraints during operation, requiring stable operation for a certain period after power adjustment. The formula is as follows: （5） In the formula, The state variable representing whether the tie line has started power adjustment is an integer variable between 0 and 1; The state variable representing whether the tie line stops power adjustment is an integer variable between 0 and 1; To adjust the interval between actions; T The total number of scheduling periods. τ As an auxiliary variable; 6) Power transmission deviation constraint: The total power transmitted during each time period of the scheduling day must meet the predetermined power range. Power adjustment is achieved by setting a power deviation coefficient, expressed by the following formula: （6） In the formula, For the electricity transmission volume agreed upon in advance; ρ This is the power deviation coefficient.

3. The distributed robust optimization scheduling method for AC / DC interconnected power systems according to claim 2, characterized in that, In step S2, the upper-level coordination optimization model is as follows: （7） In the formula, and The virtual machine groups obtained by optimizing the lower-level robust optimization model are respectively m , n Those who have made meritorious contributions; , Virtual machine groups m The penalty coefficients for the first and second terms are used to optimize tie-line power; and Virtual machine groups n The penalty coefficients for the first and second terms are used to optimize tie-line power; It is a collection of DC tie lines.

4. The distributed robust optimization scheduling method for AC / DC interconnected power systems according to claim 3, characterized in that, S302, establish the operational constraints for the basic scenario, including: 1) Operating constraints of thermal power units, as shown in equations (9) to (20), where equations (9) to (11) are the minimum operating duration constraints of thermal power units, equations (12) to (14) are the minimum shutdown duration constraints of thermal power units, equations (15) and (16) represent the start-up cost and shutdown cost of thermal power units, equations (17) and (18) represent the upper and lower bound constraints of active power output of thermal power units, while considering the coupling relationship between flexible reserve capacity and active power output, equations (19) and (20) represent the upward ramp rate constraints and downward ramp rate constraints of thermal power units, while considering the adjacent time period constraint coupling relationship when flexible reserve capacity is called. （9） （10） （11） （12） （13） （14） （15） （16） （17） （18） （19） （20） In equations (9) to (20), For conventional units i During the period t The running status; and They are conventional units i Minimum start-up time and minimum downtime; and They are conventional units i Start-up and downtime costs; and They are conventional units i The minimum and maximum active power output; and They are conventional units i The ability to climb uphill and downhill; 2) Energy storage system operation constraints, as shown in equations (21) to (35), where equation (21) represents the state of charge of the energy storage system at each time period, i.e., the stored energy after charging and discharging at that time period; equation (22) represents the upper and lower bound constraints of the stored energy; equation (23) represents that the initial energy of the energy storage system is consistent with the final energy; equation (24) represents the power output of the energy storage; equations (25) and (26) represent the charging power and discharging power limit constraints of the energy storage system, respectively; and equation (27) represents the operating state constraints of the energy storage system, i.e., at the same time... The segment cannot be charged and discharged at the same time. Equations (28) and (29) are the upward flexible reserve capacity supply capacity constraints of the energy storage system under charging and discharging states, respectively. Equations (30) and (31) are the downward flexible reserve capacity supply capacity constraints of the energy storage system under charging and discharging states, respectively. Equations (32) and (33) represent the overall upward and downward flexible reserve capacity of the energy storage system, respectively. Equations (34) and (35) are the lower and upper bound constraints of the energy storage system under the condition of flexible reserve capacity call-up, respectively. （21） （22） （23） （24） （25） （26） （27） （28） （29） （30） （31） （32） （33） （34） （35） In equations (21) to (35), For energy storage s During the period t The stored power value; , and Energy storage s Initial battery level, upper limit of battery storage, lower limit of battery storage, charging efficiency, and discharging efficiency; and Energy storage s Charging and discharging power limits; and Energy storage s During the period t Upward flexible standby capacity in charging and discharging states; and Energy storage s During the period t Downward flexible standby capacity in charging and discharging states; 3) Demand response operation constraints, as shown in equations (36) to (40), where equation (36) is the load reduction capacity constraint for incentive-type demand response, equations (37) and (38) are the minimum and maximum duration constraints for continuous reduction of incentive-type demand response, respectively, equation (39) is the time interval constraint between two adjacent load reduction actions, and equation (40) is the limit constraint for the number of times demand response can be called within the dispatching day, so as to avoid frequent calls affecting the normal power consumption of the load; （36） （37） （38） （39） （40） In equations (36) to (40), For time period t Interruptible load k The status of participation in the standby supply is indicated by a value of 1, which represents participation and a value of 0, which represents non-participation. for t -1 Period of interruptible load k The cumulative interruption time; , Interruptible loads k Minimum interrupt time and maximum interrupt time; for t -1 Period of interruptible load k The cumulative uninterrupted time; interruptible load k The minimum interruption interval time; interruptible load k The maximum number of interruptible cycles; 4) System operation constraints, including power balance constraints and power flow constraints, where equation (41) is the system power balance constraint and equation (42) is the system line power flow constraint; （41） （42） In the formula, and They are conventional units i Wind farm w Energy storage s Node load j and virtual machine groups m For grid lines l The power flow distribution transfer factor; For grid lines l The power delivery limit.

5. The distributed robust optimization scheduling method for AC / DC interconnected power systems according to claim 4, characterized in that, In step S4, the upper and lower two-layer models are iteratively solved until the convergence criterion is met, at which point the iteration stops. Specifically, this includes: S401, Convergence Criteria and Penalty Coefficient Update Principles: The iterative convergence criterion of the distributed solution algorithm is shown in Equation (51). When the optimization result meets the iterative convergence requirement, the scheduling plan of the interconnected area and the DC tie line transmission plan are obtained. （51） In the formula, ε This is the convergence threshold; If the optimization result of the iteration does not meet the convergence criterion, the penalty coefficient is updated using equation (52) for the next iteration optimization. （52） In the formula, γ For constant terms; S402, Iterative solution process: An iterative solution is adopted using an optimization framework that combines objective cascade analysis and column and constraint generation. Furthermore, the update mechanism of the DC tie-line transmission plan is utilized to achieve iterative interaction between the upper and lower layers of the optimization problem. The specific solution process of the proposed distributed optimization algorithm is as follows: 1) Initialize the iteration index Penalty coefficient and DC-DC power transmission plan; 2) Use column and constraint generation algorithms to solve the lower-level optimization subproblems of each regional power grid consisting of equations (1) to (6) and equations (8) to (50), determine the regional power grid scheduling plan and the output plan of the virtual machine group, and feed back the output information of the virtual machine group to the upper-level main problem for coordination and optimization; 3) Solve the upper-level master problem consisting of equations (1) to (7), optimize the DC tie line power transmission plan, and use it as a reference value to distribute to each lower-level optimization sub-problem; 4) If the optimization results of the upper and lower double-layer problems satisfy the convergence condition (51), then terminate the iteration and obtain the scheduling plan and DC tie line transmission plan for each region; otherwise, update the penalty coefficient of each region and return to step 2) for the next iteration.

6. The distributed robust optimization scheduling method for AC / DC interconnected power systems according to claim 5, characterized in that, The method further includes: step S5, based on the lower-level robust optimization model, the upper-level coordinated optimization model, and the optimization solution algorithm proposed in step S4, using the Yalmip toolbox combined with the optimization software CPLEX to solve the problem, and obtaining the first interconnection region. i The thermal power unit was in the first t Operating status for each time period Plans and efforts Flexible standby capacity , No. s The energy storage device in the first t State of charge and discharge during a period of time and charging / discharging power Flexible standby capacity No. k The interruptible load is in the first t Time period call status and capacity and tie-line power plan This enables decentralized, robust, and optimized scheduling of AC / DC interconnected power systems.

Images