Two-stage adaptive robust elasticity improvement method for high-proportion renewable energy power distribution network

Abstract

The invention discloses a high-proportion renewable energy power distribution network two-stage adaptive robust elastic improvement method, and the method comprises the specific steps: building a two-stage adaptive robust optimization model for a power system at a pre-disaster stage and a post-disaster stage, building an uncertainty scene set according to the output and load demands of renewable energy, determining an objective function of a min-max-min structure of the two-stage adaptive robust optimization model and constraints of a power system; and a min-max-min structure of the two-stage adaptive robust optimization model is decomposed, the model is decomposed into a main problem of the min structure and a sub-problem of the max-min structure, iteration is performed on the main problem and the sub-problem, and an optimal power supply recovery strategy is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency dispatch methods, and in particular to a two-stage adaptive robust elasticity enhancement method for a high-proportion renewable energy distribution network. Background Art

[0002] The power system is one of the most important infrastructures in modern society. However, in recent years, many power outages caused by natural disasters have occurred. Against the backdrop of frequent extreme disasters caused by global warming, the concept of power system resilience has emerged. Resilience is defined as the ability to reduce the duration and scope of high-impact, low-frequency extreme events in the power system.

[0003] Pre-disaster prevention strategies can be taken before a disaster occurs to resist extreme disasters and reduce power outage losses. Different pre-disaster prevention measures have different time scales. Existing studies have considered expansion planning methods for electric-natural gas coupling systems based on stochastic planning and robust optimization to enhance their resilience. System expansion planning usually aims at long-term profits and takes several months to several years. Other studies have strengthened fault-prone components to build a resilient power system that can resist extreme disasters. In addition, a pre-disaster active scheduling model for microgrids under typhoon disasters is proposed to improve system resilience. The time scale of the above method is days. Another study proposed a two-stage resilience improvement strategy for distribution networks that involves the coordination of multiple distributed resources. In the pre-disaster stage, an emergency mobile energy storage vehicle site selection model is established to deal with extreme disasters. The time scale of the above method is hours.

[0004] Post-disaster power supply restoration strategies mainly include real-time scheduling of emergency resources and network reconstruction. Emergency resources include mobile energy storage systems (MESS), mobile emergency generators (MEG), electric vehicle fleets, etc., and their optimal scheduling is of great significance for rapid load recovery. Existing studies guide electric vehicles to move to charging stations as energy storage elements to participate in V2G (vehicle to grid) reverse power supply to reduce the load shedding of urban power grids during typhoon disasters. There are also mobile power sources such as MESS and MEG to improve the resilience of distribution networks and discuss the feasibility of mobile power sources for distribution system restoration in terms of cost-effectiveness and scheduling. Network reconstruction can isolate faults and restore loads in non-faulty areas after disasters. Some studies identify radial network topologies based on the rank of the association matrix, achieve reconstruction by switching tie lines and segment lines, and perform distributed resource scheduling based on this to improve system resilience, and select the optimal topology that minimizes the cumulative cost from all radial topologies. The above methods verify that the coordination of real-time scheduling of emergency resources and network reconstruction after disasters can effectively restore system loads.

[0005] The access of intermittent distributed power sources such as photovoltaics and wind power brings uncertain risks to the power supply restoration of the distribution network. The output of distributed power sources is difficult to predict accurately, and the time-varying load demand in the distribution network further increases the uncertainty of power supply restoration. Therefore, it is necessary to propose a more robust power supply restoration strategy to improve system resilience. Existing studies have introduced quasi-Monte Carlo methods to model the uncertainty of distributed power output, or described the output fluctuations of distributed power sources as a predefined uncertainty set through a two-stage robust recovery optimization model. Other studies have considered the uncertainty of renewable energy output and load demand in the distribution network, and proposed a min-max two-layer robust power supply restoration model that considers the state correlation of multiple time periods of power supply restoration. However, the models in existing studies need to provide the distribution probability of random variables, and the model solution process is complex and the calculation speed is slow. Summary of the invention

[0006] The purpose of the present invention is to provide a two-stage adaptive robust elasticity improvement method for a high-proportion renewable energy distribution network. By decomposing the two-stage adaptive robust optimization model into a main-subproblem structure and performing dual and linear processing on the subproblems, the model can be solved directly without providing the distribution probability of random variables, thereby simplifying the complexity of the model solution and improving the calculation speed of the model.

[0007] In order to achieve the above object, the present invention provides a two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network, and the specific steps include:

[0008] Step 1: Establish a two-stage adaptive robust optimization model for the power system in the pre-disaster stage and the post-disaster stage, establish an uncertainty scenario set based on the output of renewable energy and load demand, and determine the objective function of the min-max-min structure of the two-stage adaptive robust optimization model and the constraints of the power system;

[0009] Step 2: Decompose the min-max-min structure of the two-stage adaptive robust optimization model into a main problem with a min structure and a sub-problem with a max-min structure, iterate the main problem-sub-problem, and find the optimal power supply restoration strategy.

[0010] Optionally, in step 2, the method for decomposing the objective function of the min-max-min structure includes: decomposing the objective function into the objective function of the min-structure main problem and the objective function of the max-min-structure sub-problem, and converting the constraints of the power system, the main problem and the objective functions of the sub-problems into matrix form.

[0011] Optionally, in step 2, the max-min structure subproblem is converted into a max structure subproblem, and the max structure subproblem is subjected to dual and linearization processing to convert the max structure subproblem into a mixed integer linear programming problem.

[0012] Optionally, the main-subproblem iterative algorithm can calculate the main problem of the min structure and the subproblem of the max structure; the main-subproblem iterative algorithm includes:

[0013] Step 2.1, setting an initial scenario set, setting initial upper and lower bounds of the objective function and a convergence gap, wherein the initial scenario set includes: output and load requirements of renewable energy units under various scenarios;

[0014] Step 2.2, search for the worst scenario set in the initial scenario set, and solve the optimal solution of the min structure main problem at this time;

[0015] Step 2.3, solve the optimal solution of the sub-problem in the worst scenario, and obtain a new worst scenario set and new upper and lower bounds of the objective function based on the optimal solutions of the main problem and sub-problems;

[0016] Step 2.4, determine whether the current objective function converges; when the difference between the upper and lower bounds of the objective function is less than or equal to the convergence gap, the optimal solution of the main problem and the sub-problem is the optimal solution of the two-stage adaptive robust optimization model; when the difference between the upper and lower bounds of the objective function is greater than the convergence gap, repeat steps 2.2 and 2.3 using the new worst scenario set to re-judge.

[0017] Optionally, in step 2.3, the upper bound of the new objective function is the value obtained by substituting the optimal solution of the objective function of the main problem into the objective function minus the real-time scheduling cost of emergency resources under the worst scenario, plus the value of the objective function of the sub-problem when the optimal solution is substituted; the lower bound of the new objective function is the value of the objective function of the main problem when the optimal solution is substituted.

[0018] Optionally, in step 1, the objective function of the min-max-min structure of the two-stage adaptive robust optimization model includes: prevention costs in the pre-disaster stage, and power supply restoration costs in the post-disaster stage; wherein, the prevention costs in the pre-disaster stage include: adjustment costs of the controllable unit combination, and the power supply restoration costs in the post-disaster stage include: load shedding costs, and start-up or shutdown operation costs of controllable units; wherein, the min structure represents the pre-disaster prevention costs and post-disaster power supply restoration costs under each scenario set, and the max-min structure represents the post-disaster power supply restoration costs under the worst scenario that can be searched in all scenario sets.

[0019] Optionally, in step 1, the objective function can be expressed as:

[0020]

[0021] In the formula, u s ,u d ,u o are binary variables representing the start-up, shutdown, and operation status of the controllable unit, respectively. When the variable is 1, the controllable unit is in this state; v is a binary variable representing the state variable of the power grid branch in the power system. When v is 1, the branch line is connected, and when v is 0, the branch line is disconnected; G is the set of controllable units in the power grid; are the startup and shutdown cost coefficients of the controllable units respectively; s is the worst fault scenario set; p g Output for controllable units; is the output cost coefficient of the controllable unit; ΔD 1 , ΔD 2 They are the load shedding capacity of important loads and the load shedding capacity of conventional loads; They are the important load shedding cost coefficient and the conventional load shedding cost coefficient respectively.

[0022] Optionally, in step 1, the power system constraints of the two-stage adaptive robust optimization model include: power grid constraints, scheduling constraints of mobile emergency generators and mobile energy storage systems, reconstruction radial constraints, operation constraints of controllable units, output constraints of renewable energy units, and load demand constraints.

[0023] Optionally, the power grid constraint adopts a regular dodecagon linearized power flow apparent power constraint, which can be expressed as:

[0024]

[0025] Where, E is the set of distribution network branches; N is the set of power grid nodes; i and j are both nodes in the power grid node set N; P jk , P ij are the active power flows of the power grid branches jk and ij respectively; Q jk , Q ij They are the reactive power flows of the power grid branches jk and ij respectively; are respectively the active power flow injection and reactive power flow injection of node j; v j 、v i are the voltages of node j and node i respectively; r ij 、x ij is the resistance and reactance of branch ij; U ac，0 is the rated voltage of the AC distribution network; are the maximum and minimum voltages at node j, respectively, v j，t is the voltage of node j at time t; α w , β w , δw are the coefficients of the regular dodecagon apparent power constraint, ranging from (0, 1), where w is the number of the 12 edge groups into which the power grid is divided; S ij is the maximum transmission capacity of branch ij.

[0026] Optionally, in step 1, robust parameters are used to represent the severity of the impact of uncertain scenarios on the power system during disasters, and an uncertain scenario set is established for the output and load demand of renewable energy, which can be expressed as:

[0027]

[0028]

[0029] In the formula, φ RES ,φ LOAD are the adjustable robust parameters of renewable energy units and load demands in the two-stage adaptive robust optimization model of the power system, respectively, which are used to characterize the output and load demands of renewable energy units when the power system is in the worst scenario. The severity range of uncertain scenarios can be adjusted by changing the parameters; m and n are the number of renewable energy units at the source end and the number of load nodes at the load end, respectively; are the predicted values ​​of renewable energy output at the source end and load demand at the load end; P RES (m), P LOAD (n) are the actual value of renewable energy output at the source end and the actual value of load demand at the load end; The upper and lower limits of renewable energy output at the source end; are the upper and lower limits of the load demand at the load end respectively; Both are binary variables, used to indicate whether the predicted value of renewable energy output can meet the actual needs of the power system. If the actual value of the renewable energy output at the source end or the actual value of the load demand at the load end is greater than the predicted value of the renewable energy output at the source end, then If the actual value of the renewable energy output at the source end and the actual value of the load demand at the load end are not greater than the predicted value of the renewable energy output at the source end

[0030] Compared with the prior art, the technical solution of the present invention has at least the following beneficial effects:

[0031] While meeting the operation constraints of the power system, the present invention takes into account the uncertainty of the output and load demand of renewable energy units, establishes a two-stage adaptive robust optimization model, and formulates the optimal strategy for optimal power supply restoration. The two-stage adaptive robust optimization model is decomposed into a main problem of a min structure and a sub-problem of a max-min structure, and the sub-problems are dualized and linearized so that the model can be solved directly without providing the distribution probability of random variables, simplifying the complexity of the model solution, and using the penalty factor method for iterative solution to improve the calculation speed of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is the iterative solution process of the main-subproblems of the two-stage adaptive robust resilience improvement method for a high-proportion renewable energy distribution network of the present invention. DETAILED DESCRIPTION

[0033] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] In the description of the present invention, it should be noted that the terms "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0035] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0036] The present invention proposes a two-stage adaptive robust resilience enhancement method for a high-proportion renewable energy distribution network, and uses a two-stage adaptive robust model to calculate the optimal strategy for power supply restoration of the power system.

[0037] The two-stage adaptive robust optimization model is a min-max-min structure, involving the four links of "source", "grid", "load" and "storage" of the power system. Among them, the min structure is used to formulate the preventive unit combination at the hour level in the pre-disaster stage, adjust the start and stop status of the controllable units and the output of renewable energy, and optimize the power network topology to ensure the reasonable distribution of the flow in the prevention stage; the max-min structure is used in the post-disaster stage to reconstruct the power network topology, real-time dispatch of emergency resources, adjust the generator output, differentiate the load shedding, and perform fault recovery of the power system.

[0038] Therefore, the two-stage adaptive robust optimization model includes: the start and stop status of controllable units, renewable energy output and post-disaster regulation in the pre-disaster stage at the source end (i.e., the power generation part of the power system), network topology reconstruction in the post-disaster stage at the network end (i.e., the power grid), differentiated load shedding in the post-disaster stage at the load end (i.e., the load side of the power system), and real-time scheduling of emergency resources in the post-disaster stage at the energy storage end; the emergency resources include: mobile energy storage system (MESS) and mobile emergency generator (MEG). An uncertainty scenario set can be established based on the output and load demand of renewable energy, and the objective function and constraints of the two-stage adaptive robust optimization model can be determined, and then the master-subproblem can be iteratively solved to obtain the optimal power supply restoration strategy. The specific steps include:

[0039] Step 1: Establish a two-stage adaptive robust optimization model for the power system in the pre-disaster stage and the post-disaster stage.

[0040] Step 1.1, establish an uncertainty scenario set.

[0041] Robust parameters are used to represent the severity of the impact of uncertain scenarios on the power system during disasters, and an uncertainty scenario set is established for the output and load demand of renewable energy. Since the output of renewable energy units is random, and the load demand at the load end is also random, the uncertainty scenario set can be established by the relationship between the predicted value and actual value of the output and load demand of renewable energy units, which can be expressed as:

[0042]

[0043] In the formula, φ RES ,φ LOADare the adjustable robust parameters of renewable energy units and load demands in the two-stage adaptive robust optimization model of the power system, respectively, which are used to characterize the output and load demands of renewable energy units when the power system is in the worst scenario. The severity range of uncertain scenarios can be adjusted by changing the parameters; m and n are the number of renewable energy units at the source end and the number of load nodes at the load end, respectively; are the predicted values ​​of renewable energy output at the source end and load demand at the load end; P RES (m), P LOAD (n) are the actual value of renewable energy output at the source end and the actual value of load demand at the load end; The upper and lower limits of renewable energy output at the source end; are the upper and lower limits of the load demand at the load end respectively; Both are binary variables, used to indicate whether the predicted value of renewable energy output can meet the actual needs of the power system. If the actual value of the renewable energy output at the source end or the actual value of the load demand at the load end is greater than the predicted value of the renewable energy output at the source end, then If the actual value of the renewable energy output at the source end and the actual value of the load demand at the load end are not greater than the predicted value of the renewable energy output at the source end

[0044] Step 1.2, establish the objective function of the two-stage adaptive robust optimization model.

[0045] In order to minimize the restoration cost of the power system, the objective function of the two-stage adaptive robust optimization model is established, which includes: the prevention cost in the pre-disaster stage and the power supply restoration cost in the post-disaster stage; the prevention cost in the pre-disaster stage includes: the cost of adjusting the combination of controllable units, and the power supply restoration cost in the post-disaster stage includes: the load shedding cost and the start-up or shutdown operation cost of the controllable units. The objective function can be represented by the min-max-min structure, as shown in Formula 3, where the min structure represents the prevention cost in the pre-disaster stage and the power supply restoration cost in the post-disaster stage under each scenario set, and the max-min structure represents the power supply restoration cost in the post-disaster stage under the worst scenario that can be searched in all scenario sets.

[0046]

[0047] In the formula, u s ,u d ,u o are binary variables representing the start-up, shutdown, and operation status of the controllable unit, respectively. When the variable is 1, the controllable unit is in this state; v is a binary variable representing the state variable of the power grid branch in the power system. When v is 1, the branch line is connected, and when v is 0, the branch line is disconnected; G is the set of controllable units in the power grid; are the startup and shutdown cost coefficients of the controllable units respectively; s is the worst fault scenario set; p g Output for controllable units; is the output cost coefficient of the controllable unit; ΔD 1 , ΔD 2 They are the load shedding capacity of important loads and the load shedding capacity of conventional loads; They are the important load shedding cost coefficient and the conventional load shedding cost coefficient respectively.

[0048] Step 1.3, set the power system constraints of the two-stage adaptive robust optimization model.

[0049] The linearized Distflow power flow model is used to set the constraints of the power grid. The regular dodecagon linearized power flow apparent power constraint is used to control the apparent power (i.e., effective power) of each node in the power grid to not exceed the maximum transmission capacity provided by the branch where the node is located, thereby completing the constraints on the power grid, which can be expressed as:

[0050]

[0051] Where, E is the set of distribution network branches; N is the set of power grid nodes; i and j are both nodes in the power grid node set N; P jk , P ij are the active power flows of the power grid branches jk and ij respectively; Q jk , Q ij They are the reactive power flows of the power grid branches jk and ij respectively; are respectively the active power flow injection and reactive power flow injection of node j; v j 、v i are the voltages of node j and node i respectively; r ij 、x ij is the resistance and reactance of branch ij; V ac，0 is the rated voltage of the AC distribution network; are the maximum and minimum voltages at node j, respectively, v j，t is the voltage of node j at time t; α w , β w , δ w are the coefficients of the regular dodecagon apparent power constraint, ranging from (0, 1), where w is the number of the 12 edge groups into which the power grid is divided; S ij is the maximum transmission capacity of branch ij.

[0052] Establish dispatch constraints for mobile emergency generators (MEGs) and mobile energy storage systems (MESSs). By using the conservation of the number of mobile emergency generators and mobile energy storage systems in the power grid and whether the output of mobile emergency generators and mobile energy storage systems meets the power supply demand of the node, dispatch constraints are established for mobile emergency generators and mobile energy storage systems, which can be expressed as:

[0053]

[0054] In the formula, Ω MEG ,Ω MESS are the set of the number of mobile emergency generators and the set of the number of mobile energy storage systems at each node of the power grid; are binary variables for the location scheduling of the wth group of mobile emergency generators and the location scheduling of the mobile energy storage system. If the variable is 1, it means that the group of mobile emergency generators or mobile energy storage systems needs to be moved to node k; are the output of the mobile emergency generator and the output of the mobile energy storage system of the wth group at the distribution network node k respectively; MEG，min 、p MEG，max are the lower and upper limits of the output of the mobile emergency generator respectively; q MESS，min ,q MESS，max The upper and lower limits of the output of the mobile emergency generator and the mobile energy storage system are set according to the operating parameters of the mobile emergency generator and the mobile energy storage system.

[0055] Establish a radial reconstruction constraint to ensure that the power grid maintains the radial topology. Using the fact that the power flow direction of the same branch can only be in one direction, the radial reconstruction constraint is expressed as:

[0056]

[0057] Where, d ij is a binary variable for the flow direction of branch ij. If the variable is 1, the flow direction is from node i to node j; if it is 0, the flow direction is from node j to node i; d ji It is a binary variable for the power flow direction of branch ij. If the variable is 1, the power flow direction is from node j to node i; if it is 0, the power flow direction is from node i to node j.

[0058] Establish the operation constraints of the controllable units, and constrain the controllable units by their ability to increase or decrease their output in a short time, which can be expressed as:

[0059]

[0060]

[0061] In the formula, S CG is a binary variable indicating the start-up state of the controllable unit. CG When it is 1, it means the controllable unit is turned on. CG When it is 0, it means the controllable unit is not turned on; P CG is the output of the controllable unit; is the maximum output of the controllable unit; The output of the controllable units in the post-disaster stage and the pre-disaster stage; They are the maximum ramp-up power and the maximum ramp-down power of the controllable unit, that is, the maximum output that the controllable unit can increase or decrease in unit time, where the unit time is usually 15 minutes; They are the upward and downward adjustment amounts of the controllable unit output respectively; It is the maximum active power increase and decrease of the controllable units; They are the upward and downward adjustment amounts of controllable units in the post-disaster stage respectively; They are respectively the upward and downward adjustments of controllable units in the pre-disaster stage.

[0062] Establish the output constraints of renewable energy units. The actual consumption of renewable energy cannot exceed the output requirements of renewable energy units. Establish the output constraints of renewable energy units, which can be expressed as:

[0063]

[0064] In the formula, is the actual consumption value of renewable energy, P RES The output of renewable energy units.

[0065] Establish load demand constraints. For the stability and security of the power system, the supply load of the power system must be less than the load demand. The load demand constraint can be expressed as:

[0066]

[0067] In the formula, is the actual supply load, P LOAD For load demand.

[0068] Step 2: Decompose the min-max-min structure of the two-stage adaptive robust optimization model and find the optimal power supply restoration strategy.

[0069] Formula 3 to Formula 9 are sorted and converted into a matrix form, as shown in Formula 10.

[0070]

[0071] In the formula, c, d, e are the coefficient matrices of the two-stage adaptive robust optimization model in Formula 3, including: cost coefficients of start-up and shutdown of controllable units, output cost coefficients of controllable units, cost coefficients of important load shedding, and cost of conventional load shedding; A, C, D are the coefficient matrices of the equality constraints in Formula 4; g and i are the constant column vectors of the equality constraints in Formula 4, including: active power flow and reactive power flow of power grid nodes; B, E, F, G are the coefficient matrices of the inequality constraints in Formula 4-9, which are generally set to 1; h and j are the constant column vectors of the inequality constraints in Formula 4-9, including: the maximum output of controllable units and the maximum transmission capacity of power grid branches; u is the worst fault scenario set.

[0072] Using the robust optimization (Column-and-Constraint Generation Algorithm, C&CG) algorithm, equation 10 is decomposed into a main problem of a min structure and a sub-problem of a max-min structure. The main problem can be expressed as:

[0073]

[0074] Where θ is the real-time scheduling cost of emergency resources under the worst known scenario set, is the objective function of the main problem, which represents the prevention cost in the pre-disaster stage and the power supply restoration cost in the post-disaster stage under the scenario set.

[0075] The sub-problem can be expressed as:

[0076]

[0077] At this time, Equation 12 cannot be solved directly. We can further transform Equation 12 into a sub-problem of the max structure, which can be expressed as:

[0078]

[0079] In the formula, α, β, γ are dual variables, and ξ is an auxiliary variable. Since u is an uncertain parameter interval, let u T ξ is a bilinear term and cannot be solved directly. To maximize the objective function, each positive value of ξ corresponds to the upper bound of u, and each negative value of ξ corresponds to the lower bound of u. Therefore, the penalty factor method is used to linearize Equation 13, so that the dual and linearized max structure subproblem is transformed into a mixed integer linear programming problem, which can be solved directly. The transformed max structure subproblem can be expressed as:

[0080]

[0081] In the formula, u up 、u down are the upper and lower bounds of renewable energy output and load demand range, respectively; ξ +, - is the positive or negative value of the auxiliary variable ξ; is an auxiliary binary variable; is the objective function of the sub-problem, which represents the power supply restoration cost in the post-disaster stage under the worst scenario searched in all scenario sets.

[0082] like Figure 1 As shown, the main-subproblem iterative algorithm is used to calculate the main problem-subproblem of the min-max-min structure.

[0083] Step 2.1, set initial parameters.

[0084] According to formula 1-2, the initial scene set u is obtained 1 The initial scenario set includes: the output and load requirements of renewable energy units in various scenarios; setting the initial upper and lower bounds of the objective function, and initializing the upper bound of the objective function U 0 Set to +∞, and initialize the lower bound of the objective function L 0 Set to -∞; and set the convergence gap ζ.

[0085] Step 2.2, search for the worst scenario set and solve the min structure main problem.

[0086] Search for the worst scenario set in the current initialization scenario set Substitute the output and load demand of the renewable energy units under this scenario into equation 11 to calculate and solve the optimal solution x for the current main problem. s , at this time, ε=1.

[0087] Step 2.3, find the optimal solution of the subproblem under this scenario set, and obtain a new worst scenario set and new upper and lower bounds of the objective function.

[0088] The optimal solution x of the main problem in step 2.2 s Substitute the objective function of the main problem and use the value of the objective function as the new lower bound value L of the objective function ε .

[0089] The worst scenario set among the initial scenario sets Under the condition of , the output and load demand of the renewable energy units in this scenario set are substituted into Equation 14 to calculate the optimal solution of the objective function of the sub-problem Obtain the power restoration strategy and convert the optimal solution into As the latest set of bad scenes

[0090] Replace c in the main problem T Substitute x into the optimal solution x ε The value of and the optimal solution of the sub-problem objective function The sum of the values ​​is taken as the upper bound of the new objective function U s At this time, the new upper and lower bounds of the objective function are U s , L ε .

[0091] Step 2.4, determine whether the current objective function converges.

[0092] When the difference between the upper and lower bounds of the objective function is less than or equal to the convergence gap, that is, U ε -L ε ≤ζ, the objective function (i.e., Equation 3) converges, and the optimal solution of the objective function of the main problem and the optimal solution of the objective function of the sub-problem are the optimal solutions of the two-stage adaptive robust optimization model.

[0093] When the difference between the upper and lower bounds of the objective function is greater than the convergence gap, substitute the new set of severe scenarios in step 2.3 into step 2.2, repeat steps 2.2 to 2.3, and re-obtain the optimal solution of the objective function of the main problem, the optimal solution of the objective function of the sub-problem, and the new upper and lower bounds of the objective function, and re-determine the difference between the upper and lower bounds.

[0094] The minimum restoration cost and optimal power supply restoration strategy of the power system can be obtained according to the optimal solution of the two-stage adaptive robust optimization model.

[0095] By simulating the PG&E-69 node test system using the method of the present invention, different power supply restoration strategies under various scenario sets can be obtained, and the performance of the power system under different power supply restoration strategies can be compared. It can also be verified that the obtained power supply restoration strategy can effectively cope with the uncertainty of renewable energy units and load demand. At the same time, through the reasonable selection of the adjustable robust parameters in the method of the present invention, the balance between the economy and robustness of the power supply restoration strategy can be obtained, and the effectiveness of the method of the present invention can be verified.

[0096] In summary, the present invention decomposes the two-stage adaptive robust optimization model into a main-subproblem structure and performs dual and linear processing on the subproblems, so that the model can be solved directly without providing the distribution probability of random variables; at the same time, the main-subproblems are iteratively solved to simplify the complexity of the model solution and improve the calculation speed of the model.

[0097] Although the content of the present invention has been described in detail through the above preferred embodiments, it should be appreciated that the above description should not be considered as a limitation of the present invention. After reading the above content, it will be apparent to those skilled in the art that various modifications and substitutions of the present invention will occur. Therefore, the protection scope of the present invention should be limited by the appended claims.

Claims

1. A two-stage adaptive robust resilience improvement method for a high proportion of renewable energy distribution network, characterized in that: The specific steps include: Step 1: Establish a two-stage adaptive robust optimization model for the power system in the pre-disaster stage and the post-disaster stage, establish an uncertainty scenario set based on the output of renewable energy and load demand, and determine the objective function of the min-max-min structure of the two-stage adaptive robust optimization model and the constraints of the power system; Step 2: Decompose the min-max-min structure of the two-stage adaptive robust optimization model into a main problem with a min structure and a sub-problem with a max-min structure, iterate the main problem-sub-problem, and find the optimal power supply restoration strategy.

2. The two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network according to claim 1 is characterized in that: In step 2, the method for decomposing the objective function of the min-max-min structure includes: decomposing the objective function into the objective function of the min-structure main problem and the objective function of the max-min-structure sub-problem, and converting the constraints of the power system, the main problem and the objective functions of the sub-problems into matrix form.

3. The two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network according to claim 2 is characterized in that: In step 2, the max-min structure subproblem is converted into a max structure subproblem, and the max structure subproblem is subjected to dual and linearization processing, so as to convert the max structure subproblem into a mixed integer linear programming problem.

4. The two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network according to claim 3 is characterized in that: The main-subproblem iterative algorithm can calculate the main problem of min structure and the subproblem of max structure; The iterative algorithm for the master-subproblem includes: Step 2.1, setting an initial scenario set, setting initial upper and lower bounds of the objective function and a convergence gap, wherein the initial scenario set includes: output and load requirements of renewable energy units under various scenarios; Step 2.2, search for the worst scenario set in the initial scenario set, and solve the optimal solution of the min structure main problem at this time; Step 2.3, solve the optimal solution of the sub-problem in the worst scenario, and obtain a new worst scenario set and new upper and lower bounds of the objective function based on the optimal solutions of the main problem and sub-problems; Step 2.4, determine whether the current objective function converges; when the difference between the upper and lower bounds of the objective function is less than or equal to the convergence gap, the optimal solution of the main problem and the sub-problem is the optimal solution of the two-stage adaptive robust optimization model; when the difference between the upper and lower bounds of the objective function is greater than the convergence gap, repeat steps 2.2 and 2.3 using the new worst scenario set to re-judge.

5. The two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network according to claim 4 is characterized in that: In step 2.3, the upper bound of the new objective function is the value obtained by substituting the optimal solution of the objective function of the main problem into the objective function minus the real-time scheduling cost of emergency resources in the worst scenario, plus the value of the objective function of the sub-problem when the optimal solution is substituted; the lower bound of the new objective function is the value of the objective function of the main problem when the optimal solution is substituted.

6. The two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network according to claim 1 is characterized in that: In step 1, the objective function of the min-max-min structure of the two-stage adaptive robust optimization model includes: prevention cost in the pre-disaster stage and power supply restoration cost in the post-disaster stage; wherein, the prevention cost in the pre-disaster stage includes: adjustment cost of controllable unit combination, and the power supply restoration cost in the post-disaster stage includes: load shedding cost, start-up or shutdown operation cost of controllable units; wherein, the min structure represents the pre-disaster prevention cost and the post-disaster power supply restoration cost under each scenario set, and the max-min structure represents the post-disaster power supply restoration cost under the worst scenario that can be searched in all scenario sets.

7. The two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network according to claim 6 is characterized in that: In step 1, the objective function can be expressed as: In the formula, u s ,u d ,u o are binary variables representing the start-up, shutdown, and operation status of the controllable unit, respectively. When the variable is 1, the controllable unit is in this state; v is a binary variable representing the state variable of the power grid branch in the power system. When v is 1, the branch line is connected, and when v is 0, the branch line is disconnected; G is the set of controllable units in the power grid; are the startup and shutdown cost coefficients of the controllable units respectively; s is the worst failure scenario set; p g Output for controllable units; is the output cost coefficient of the controllable unit; ΔD1 and ΔD2 are the load shedding amounts of important loads and conventional loads respectively; They are the important load shedding cost coefficient and the conventional load shedding cost coefficient respectively.

8. The two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network according to claim 1 is characterized in that: In step 1, the power system constraints of the two-stage adaptive robust optimization model include: grid constraints, scheduling constraints of mobile emergency generators and mobile energy storage systems, reconstructed radial constraints, operation constraints of controllable units, output constraints of renewable energy units, and load demand constraints.

9. The two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network according to claim 8 is characterized in that: The power grid constraint adopts regular dodecagon linearized power flow apparent power constraint, which can be expressed as: Where, E is the set of distribution network branches; N is the set of power grid nodes; i and j are both nodes in the power grid node set N; P jk , P ij are the active power flows of the power grid branches jk and ij respectively; Q jk , Q ij They are the reactive power flows of the power grid branch jk and branch ij respectively; are respectively the active power flow injection and reactive power flow injection of node j; v j 、v i are the voltages of node j and node i respectively; r ij 、x ij is the resistance and reactance of branch ij; U ac，0 is the rated voltage of the AC distribution network; are the maximum and minimum voltages of node j, respectively, v j，t is the voltage of node j at time t; α w , β w ,δ w are the coefficients of the regular dodecagon apparent power constraint, ranging from (0, 1), where w is the number of the 12 edge groups into which the power grid is divided; S ij is the maximum transmission capacity of branch ij.

10. The two-stage adaptive robust elasticity improvement method for a high proportion of renewable energy distribution network according to claim 1, characterized in that: In step 1, robust parameters are used to represent the severity of the impact of uncertain scenarios on the power system during disasters, and an uncertainty scenario set is established for the output and load demand of renewable energy, which can be expressed as: In the formula, φ RES ,φ LOAD are the adjustable robust parameters of renewable energy units and load demands in the two-stage adaptive robust optimization model of the power system, respectively, which are used to characterize the output and load demands of renewable energy units when the power system is in the worst scenario. The severity range of uncertain scenarios can be adjusted by changing the parameters; m and n are the number of renewable energy units at the source end and the number of load nodes at the load end, respectively; are the predicted values ​​of renewable energy output at the source end and load demand at the load end; P RES (m), P LOAD (n) are the actual value of renewable energy output at the source end and the actual value of load demand at the load end; The upper and lower limits of renewable energy output at the source end; are the upper and lower limits of the load demand at the load end respectively; Both are binary variables, used to indicate whether the predicted value of renewable energy output can meet the actual needs of the power system. If the actual value of the renewable energy output at the source end or the actual value of the load demand at the load end is greater than the predicted value of the renewable energy output at the source end, then If the actual value of the renewable energy output at the source end and the actual value of the load demand at the load end are not greater than the predicted value of the renewable energy output at the source end

Images