Unbalanced active power distribution network energy storage system configuration method based on wind and light uncertainty

Abstract

The invention belongs to the technical field of active power distribution network energy storage, and particularly relates to a wind and light uncertainty-based unbalanced active power distribution network energy storage system configuration method, which comprises the following steps of: firstly, building a three-phase unbalanced active power distribution network energy storage system certainty model based on a three-phase unbalanced branch load flow model form; constructing a polyhedral uncertainty set containing budget parameters, establishing an unbalanced active power distribution network DES optimal configuration model considering wind and light uncertainty, and performing solving based on a column sum constraint generation algorithm and second-order cone programming; furthermore, according to a DES operation result obtained through calculation, a DES configuration strategy considering the comprehensive network loss sensitivity is formulated. Compared with a conventional energy storage configuration algorithm, the method is more suitable for considering uncertainty scenes under new energy access, meanwhile, two-way interaction between the energy storage system and the active power distribution network is considered, good economical efficiency is achieved, and the obtained configuration result better fits the actual operation condition.

Description

Technical Field

[0001] The present invention belongs to the technical field of active distribution network energy storage, and specifically relates to a method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty. Background Art

[0002] In response to sustainable development and environmental challenges, distributed generation (DG) technology has been rapidly promoted worldwide. However, due to its high intermittency and uncertainty, DG's high penetration in the distribution network has a negative impact on network planning and operation, and the spatiotemporal mismatch between load and DG output exacerbates the difference between load peaks and valleys, leading to rising energy costs. Distributed energy storage (DES) effectively alleviates these problems by storing electrical energy during load troughs and releasing it during load peaks, thereby decoupling the time of electricity production and consumption. However, due to the high cost of DES, achieving its expected benefits depends on optimal configuration. Therefore, how to effectively optimize the configuration of DES becomes a key issue.

[0003] The optimal configuration of battery energy storage systems covers the optimal selection of technology type, installation location, battery capacity and operation strategy. At present, the DES configuration problem has become a widely studied field. In modern power systems, the configuration and number of battery energy storage systems are mainly limited by their own costs. At the same time, policy requirements have promoted the widespread access of DG in distribution networks, increasing the uncertainty dimension of the system, thus bringing more challenges to the safe and stable operation of the power grid.

[0004] Therefore, when modeling the configuration of battery energy storage systems, economic efficiency should be the main goal, and the configuration of DES should be optimized to achieve the function of peak shaving and valley filling, thereby maintaining the safety and stability of the power grid. With the rapid development of distributed energy, some studies have also focused on the problem of optimal configuration of energy storage considering DG uncertainty. At present, the description methods for uncertain scenarios mainly use scenario generation method, chance constraint method and traditional robust algorithm. However, due to the high complexity, randomness and fuzziness of the data required for DES planning, these methods face significant challenges in characterizing uncertain scenarios and their characteristics, and it is difficult to achieve a balance between economy and efficiency. At the same time, the current models for DES site selection considering uncertainty factors mainly focus on analyzing the impact of various distribution network indicators on DES before installation, and have not yet achieved two-way coordination and interaction between DES and active distribution networks, thus failing to effectively integrate operation conditions into the planning process. This may lead to inefficient utilization of energy storage resources or increased operating costs. At the same time, improper site selection may also cause problems such as overload, overvoltage or undervoltage in local distribution networks. Therefore, it is necessary to study a DES configuration strategy based on uncertain scenarios that considers planning configuration and operation coordination issues.

[0005] In low-voltage distribution networks, due to the lack of line transposition and asymmetric layout, the actual three-phase parameters are often unbalanced, which leads to a decrease in the operating efficiency of the active distribution network and an increase in network losses. This may cause the DES to lack the necessary flexibility when dealing with sudden load fluctuations or unstable energy supply, and thus fail to give full play to its energy storage advantages. The original balanced network flow model is no longer applicable to this situation, so it is necessary to study the flow model applicable to the unbalanced low-voltage distribution network to effectively optimize the configuration of the DES. Summary of the invention

[0006] The purpose of the present invention is to provide a method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty, which overcomes the shortcomings of the prior art, comprehensively considers wind and solar uncertainty, the two-way coordinated interaction between DES and the associated distribution network, and takes into account the impact of three-phase imbalance. Compared with conventional energy storage configuration algorithms, it is more suitable for considering uncertainty scenarios under new energy access, has good economy, and the obtained configuration results are more in line with actual operating conditions.

[0007] To solve the above problems, the technical solutions adopted by the present invention are as follows:

[0008] The method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty includes the following steps:

[0009] S1. Build a deterministic model of energy storage system for unbalanced distribution network;

[0010] S2. Establish an optimal configuration model for the unbalanced active distribution network DES considering the uncertainty of wind and solar power, and use an optimization algorithm combining column and constraint generation algorithm with second-order cone programming to solve it;

[0011] S3. Formulate a DES sequence configuration strategy that takes into account the sensitivity of comprehensive network loss;

[0012] S4. The rationality and superiority of the DES sequence configuration strategy are verified through typical example system simulation.

[0013] Furthermore, the specific steps of building a deterministic model of an unbalanced distribution network energy storage system described in S1 include:

[0014] S11. Based on the three-phase unbalanced branch power flow model, build a network power flow model and input power flow model data;

[0015] S12. Considering minimizing the cost of DES while increasing the benefits, a deterministic model of the unbalanced active distribution network energy storage system is constructed, and its expression is:

[0016] minS LR +S LS -C I -C M -C depore

[0017] Among them, S LR In order to reduce network loss and profit, S LS is the load transfer benefit, C I is the DES investment cost, C M is the DES operation and maintenance cost, C depore is the DES life loss cost.

[0018] Furthermore, as described in S2, an optimal configuration model of DES for unbalanced active distribution network considering the uncertainty of wind and solar power is established, and an optimization algorithm combining column and constraint generation algorithm with second-order cone programming is used to solve the problem. The specific steps include:

[0019] S21. Construct a polyhedral uncertainty set containing budget parameters to characterize wind and solar uncertainty;

[0020] S22. Establish an optimal configuration model for DES in unbalanced active distribution network considering the uncertainty of wind and solar power;

[0021] S23. Based on the column and constraint generation algorithm and second-order cone programming, the DES optimization calculation of the unbalanced active distribution network considering the uncertainty of wind and solar power is carried out.

[0022] Further, the expression of the polyhedron uncertainty set in S21 is:

[0023]

[0024] Where P D ' G,t For DG, the output is uncertain, P DG,t is the predicted value of DG active output, It is an auxiliary variable for the photovoltaic output at the upper and lower boundaries, used to adjust the uncertainty interval of each time section. It is the fluctuation range of DG active output.

[0025] Furthermore, the unbalanced active distribution network DES optimization configuration model considering the uncertainty of wind and solar power is established in S22, and the compact form of the adaptive robust optimization model is adopted, and its expression is:

[0026]

[0027] The main problem is to minimize the objective function by adjusting the variable x, and the subproblem is to maximize the variable y under the worst case of uncertain parameter u.

[0028] Further, the specific steps of the unbalanced active distribution network DES optimization calculation in S22 are:

[0029] (1) Using the C&CG algorithm, the three-level min-max-min problem is decomposed into a single-level min main problem and a two-level max-min subproblem for solution. The main problem is:

[0030] minOF1=C I +C M +C depore

[0031] The sub-problems are:

[0032] maxminOF2=S LR +S LS

[0033] Among them, S LR In order to reduce network loss and profit, S LS is the load transfer benefit, C I is the DES investment cost, C M is the DES operation and maintenance cost, C depore is the DES life loss cost.

[0034] (4) Solve the main problem and sub-problems separately: the main problem is a conventional linear programming problem, which can be solved directly by calling the existing solver. The sub-problem is difficult to solve directly due to the existence of nonlinear power flow constraints and non-convex programming problems, and needs to be relaxed using SOCP;

[0035] The nonlinear power flow constraint in the original problem is:

[0036]

[0037]

[0038] By eliminating the quadratic terms in the constraints and the SOCP relaxation step, the above constraints can be transformed into the following form:

[0039]

[0040] (5) Since the subproblem is a max-min problem, the duality theory is used to transform the two-level model into a single-level linear programming model;

[0041] First, the relaxed model is written as follows:

[0042]

[0043] stCy≤f

[0044] Dy=g

[0045] Ey=u

[0046]

[0047] Where C, D, E, Q, q, c, d, f, g are all known coefficient matrices, u is an uncertain parameter, and y is a decision variable. Through the Lagrange duality theory, the dual "max" model is obtained as follows:

[0048]

[0049] ||ω i ||2≤λ i ,i=1,...,n

[0050] π1≤0

[0051] u min ≤u≤u max

[0052] In the formula, π1, π2, π3 are dual variables, λ i ω i is a quadratic dual variable, and n is the number of all uncertain scenarios. After the dual transformation, there is still a bilinear term u in the above formula T π3, which makes the model non-convex and unable to find the global optimal solution. Next, we introduce the dummy variable ζ t , use large M to transform:

[0053]

[0054] -M(1-ζt )+π 3,s ≤r s ≤M(1-ζ t )+π 3,s

[0055] -Mζ t ≤r s ≤Mζ t

[0056] In the formula, r s is an auxiliary variable, M is a large constant, ζ t is a 0-1 variable, s is a set of nodes with DG;

[0057] When t = 0, the bilinear term u T π3 gets u s =u min , corresponding to the worst scenario of uncertain variables; when ζ t = 1, the bilinear term u T π3 gets u s =u max , corresponding to the optimal scenario of uncertain variables;

[0058] (4) Introducing uncertainty margin τ DG , by adjusting its size, the model conservatism is improved:

[0059]

[0060] In the formula, τ DG is the uncertainty margin of DG output.

[0061] Furthermore, in the DES sequential configuration strategy considering comprehensive network loss sensitivity described in S3, network loss sensitivity analysis is used to determine the installation location and number of DES, and its expression is:

[0062]

[0063] In the formula, λ d , c are the network loss sensitivity weights corresponding to charging and discharging, which are set according to the actual charging and discharging duration and electricity price, so that λ d +λ c =1; are the network loss sensitivities corresponding to the r-th configuration charging and discharging respectively.

[0064] Further, the rationality and superiority of the DES sequence configuration strategy are verified by simulation of a typical example system in S4, and the specific steps include:

[0065] S41. Feasibility verification of algorithm results: Based on the proposed energy storage system configuration strategy, a feasibility analysis is performed on the obtained results.

[0066] S42. Verification of algorithm superiority: By comparing the calculation results with those of conventional methods, the superiority of the proposed algorithm is analyzed from the aspects of economy and result security.

[0067] Compared with the prior art, the present invention has the following beneficial effects:

[0068] The present invention comprehensively considers the uncertainty of wind and solar power, the two-way coordinated interaction between DES and the associated distribution network, and the impact of three-phase imbalance. Compared with conventional energy storage configuration algorithms, it is more suitable for considering uncertain scenarios under the access of new energy, has good economic efficiency, and the obtained configuration results are more in line with actual operating conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is the algorithm diagram of the present invention;

[0070] Figure 2 This is a wiring diagram of a real distribution network in Australia used for simulation in the present invention;

[0071] Figure 3 A network loss sensitivity explanation diagram for the DES configuration of the present invention;

[0072] Figure 4 This is a typical daily operation diagram of the DES of the present invention;

[0073] Figure 5 A comparison chart of the DES configuration results of the present invention and other experiments;

[0074] Figure 6 This is a post hoc analysis of the conservatism of the present invention. DETAILED DESCRIPTION

[0075] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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.

[0076] like Figure 1 As shown, the method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty of the present invention includes the following:

[0077] S1. Build a deterministic model of energy storage system for unbalanced distribution network;

[0078] S2. Establish an optimal configuration model for unbalanced active distribution network (distributed energy storage, DES) considering the uncertainty of wind and solar power, and use an optimization algorithm that combines column-and-constraint generation (C&CG) with second order cone programming (SOCP) to solve it;

[0079] S3. Formulate a DES sequence configuration strategy that takes into account the sensitivity of comprehensive network loss;

[0080] S4. Verify the rationality and superiority of the submitted DES optimization configuration method through simulation of typical example systems;

[0081] Further, as a preferred solution, the specific process of S1 is:

[0082] S11. Based on the three-phase unbalanced branch power flow model (distribution loadflow, DistFlow) form, build a network power flow model and input power flow model data;

[0083] S12. Considering minimizing DES costs while increasing revenue, a deterministic model of unbalanced active distribution network energy storage system is constructed:

[0084] minS LR +S LS -C I -C M -C depore

[0085] Among them, S LR In order to reduce network loss and profit, S LS is the load transfer benefit, C I is the DES investment cost, C M is the DES operation and maintenance cost, C depore is the DES life loss cost.

[0086] Further, as a preferred solution, the specific process of S2 is:

[0087] S21. Construct a polyhedral uncertainty set containing budget parameters to characterize wind and solar uncertainty:

[0088] S22. Establish an optimal configuration model for DES in unbalanced active distribution network considering the uncertainty of wind and solar power:

[0089] S23. Based on C&CG and SOCP, DES optimization calculation of unbalanced active distribution network considering the uncertainty of wind and solar power is performed.

[0090] Further, as a preferred solution, the specific process of S21 is:

[0091]

[0092] Where P D ' G,t For DG, the output is uncertain, P DG,t is the predicted value of DG active output, It is an auxiliary variable for the photovoltaic output at the upper and lower boundaries, used to adjust the uncertainty interval of each time section. is the fluctuation range of DG active output;

[0093] Further, as a preferred solution, the specific process of S21 is:

[0094] The compact form of the adaptive robust optimization (ARO) model used is as follows:

[0095]

[0096] The main problem is to minimize the objective function by adjusting the variable x, and the subproblem is to maximize the variable y under the worst case of uncertain parameter u.

[0097] Further, as a preferred solution, the specific process of S3 is:

[0098] S31. Using the C&CG algorithm, the three-level min-max-min problem is decomposed into a single-level min main problem and a two-level max-min sub-problem for solution. The main problem is:

[0099] minOF1=C I +C M +C depore

[0100] The sub-problems are:

[0101] maxminOF2=S LR +S LS

[0102] S32. Solve the main problem and sub-problems respectively: the main problem is a conventional linear programming problem, which can be solved directly by calling the existing solver; however, the sub-problem is difficult to solve directly due to the existence of nonlinear power flow constraints and non-convex programming problems, and needs to be relaxed using SOCP: the nonlinear power flow constraints in the original problem are:

[0103]

[0104] By eliminating the quadratic terms in the constraints and relaxing the SOCP, the above constraints can be transformed into the following form:

[0105]

[0106] S33. Since the subproblem is a max-min problem, it is difficult to solve it directly in actual operation, so the duality theory is used to transform the two-layer model into a single-layer linear programming model. First, the relaxed model is written in the following form:

[0107]

[0108] stCy≤f

[0109] Dy=g

[0110] Ey=u

[0111]

[0112] Where C, D, E, Q, q, c, d, f, g are all known coefficient matrices, u is an uncertain parameter, and y is a decision variable. Through the Lagrange duality theory, the dual "max" model is obtained as follows:

[0113]

[0114] ||ω i ||2≤λ i ,i=1,...,n

[0115] π1≤0

[0116] u min ≤u≤u max

[0117] In the formula, π1, π2, π3 are dual variables, λ i ω i is a quadratic dual variable, and n is the number of all uncertain scenarios. After the dual transformation, there is still a bilinear term u in the above formula T π3, which makes the model non-convex and unable to find the global optimal solution. Next, we introduce the dummy variable ζ t , using large M for transformation:

[0118]

[0119] -Mζ t ≤r s ≤Mζ t

[0120] In the formula, r s is an auxiliary variable, M is a large constant, ζ tis a 0-1 variable, and s is a set of nodes with DG. The result is: when ζ t = 0, the bilinear term u T π3 gets u s =u min , corresponding to the worst scenario of uncertain variables; when ζ t = 1, the bilinear term u T π3 gets u s =u max , corresponding to the optimal scenario of uncertain variables.

[0121] S34. To reduce the conservatism of the results, introduce uncertainty margin τ DG , by adjusting its size, the model conservatism is improved:

[0122]

[0123] In the formula, τ DG is the uncertainty margin of DG output. By adjusting its size, the model conservatism can be improved. As a preferred solution, the specific process of S3 is as follows:

[0124] S31. Use network loss sensitivity analysis to determine the installation location and quantity of DES:

[0125]

[0126] In the formula, λ d , c are the network loss sensitivity weights corresponding to charging and discharging, which are set according to the actual charging and discharging duration and electricity price, so that λ d +λ c =1; are the network loss sensitivities corresponding to the r-th configuration charging and discharging.

[0127] S32. Formulate a DES sequential configuration strategy that takes into account the sensitivity of comprehensive network loss. Considering that the selection of DES access location and charging and discharging strategy directly affect the distribution network flow, and then affect the network loss and voltage level, the DES on the distribution network side needs to take into account the comprehensive optimization of planning and operation in site selection. To this end, a configuration strategy is proposed to deal with the existing distribution network cluster layout with low benefits, and to support the application of the DES site selection proposed in the present invention in complex unbalanced distribution networks.

[0128] Further, as a preferred solution, the specific process of S4 is:

[0129] S41. Feasibility verification of algorithm results: Based on the proposed energy storage system configuration strategy, a feasibility analysis is performed on the obtained results.

[0130] S42. Verification of algorithm superiority: By comparing the calculation results with those of conventional methods, the superiority of the proposed algorithm is analyzed from the aspects of economy and result security.

[0131] The effect of the present invention can be verified and illustrated by the following performance analysis and simulation experiment test. Simulation experiment and performance analysis:

[0132] The present invention selects Figure 2 The real Australian distribution network shown in the figure is used as a simulation test environment. The 415V three-phase distribution network is powered by a distribution transformer with a capacity of 200kVA and a voltage of 22kV / 415V, including 101 busbars and 77 power-consuming devices. Among them, 51 users are single-phase users, 34 users install single-phase rooftop photovoltaic systems and 2 three-phase horizontal axis wind turbines, with an average installed capacity of 1.82kWp, a total installed capacity of 63.81kWp, and a penetration rate of 31.9%. The rated capacity of the wind turbines is 48kW, and the penetration rate is 24%.

[0133] Figure 3 A diagram showing the network loss sensitivity explanation of the DES configuration of the present invention. Figure 4 The following table shows the results of four optimization configurations of energy storage, which are shown in Table 1 and Table 2. Figure 3 , Figure 4 It can be seen that as the number of configurations increases, the objective function continues to increase, indicating that the benefits continue to increase. The DES capacity for the fifth configuration is 0 and the objective function remains unchanged, which shows that if DES continues to be configured on the basis of the continuous increase in cost, it will cause DES to lose benefits, which is obviously not feasible. In summary, the first four DES configurations are the optimal configuration solutions.

[0134]

[0135] The following table shows the results and benefits of the present invention compared with other DES configuration methods, showing the energy storage investment cost, operation and maintenance cost, battery life loss cost, network loss reduction benefit and load transfer benefit under each algorithm. Figure 5It can be seen that compared with random optimization and robust optimization (RO), ARO has increased the comprehensive cost of energy storage investment and operation. This is because the robust model takes into account the system safety constraints under extreme conditions when designing. ARO can not only significantly reduce system network losses, but also improve the load transfer benefits of the distribution network. This improvement stems from the investment of more DES capacity, thereby enhancing the load transfer capacity of the distribution network in time and space, so that peak loads can be transferred to low periods, avoiding overload operation of equipment and extending its service life. Therefore, although the two-stage adaptive robust algorithm adopted in the present invention is not as economical as random optimization and RO, its planning and operation results are superior in terms of safety, can reduce risk costs to a certain extent, and effectively support the practical application of the proposed DES optimization configuration model in unbalanced distribution networks.

[0136]

[0137] Figure 6 The conservative post-analysis heat map of the present invention is shown. Figure 6 It can be seen that when the 21 uncertain parameters deviate from their predicted values ​​by 10%, the system residual power drops to 0.53kWh, marking the lowest point of residual power, and the conservatism of the adaptive robust model is also at the lowest level. The size of residual power and its uncertainty margin directly affects the decision maker's assessment of the system's conservatism. Higher residual power usually prompts decision makers to adopt more cautious strategies to maintain the stability and security of the system; when the residual power is low, decision makers will focus more on improving energy efficiency, thereby reducing conservatism. In addition, the setting of uncertainty margin should be appropriately adjusted according to the needs and risk tolerance of DES. If the uncertainty margin is too small, it may cause the system to fail in abnormal situations, thereby affecting the reliability and stability of DES; while too large uncertainty margin will increase the cost and complexity of DES configuration. Therefore, the reasonable selection of residual power is crucial to reducing the conservatism of the ARO model.

[0138] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty, characterized by: The following steps are involved: S1. Build a deterministic model of energy storage system for unbalanced distribution network; S2. Establish an optimal configuration model for the unbalanced active distribution network DES considering the uncertainty of wind and solar power, and use an optimization algorithm that combines the column and constraint generation algorithm with the second-order cone programming to solve it; S3. Formulate a DES sequence configuration strategy that takes into account the sensitivity of comprehensive network loss; S4. The rationality and superiority of the DES sequence configuration strategy are verified through typical example system simulation.

2. The method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty according to claim 1 is characterized in that: The specific steps of building a deterministic model of an unbalanced distribution network energy storage system as described in S1 include: S11. Based on the three-phase unbalanced branch power flow model, build a network power flow model and input power flow model data; S12. Considering minimizing the cost of DES while increasing the benefits, a deterministic model of the unbalanced active distribution network energy storage system is constructed, and its expression is: my S LR +S LS -C I -C M -C depore Among them, S LR In order to reduce network loss and profit, S LS is the load transfer benefit, C I is the DES investment cost, C M is the DES operation and maintenance cost, C depore is the DES life loss cost.

3. The method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty according to claim 1 is characterized in that: As described in S2, an optimal configuration model of DES for unbalanced active distribution network considering the uncertainty of wind and solar power is established, and an optimization algorithm combining column and constraint generation algorithm with second-order cone programming is used to solve it. The specific steps include: S21. Construct a polyhedral uncertainty set containing budget parameters to characterize wind and solar uncertainty; S22. Establish an optimal configuration model for DES in unbalanced active distribution network considering the uncertainty of wind and solar power; S23. Based on the column and constraint generation algorithm and second-order cone programming, the DES optimization calculation of the unbalanced active distribution network considering the uncertainty of wind and solar power is carried out.

4. The method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty according to claim 3 is characterized in that: The expression of the polyhedral uncertainty set described in S21 is: In the formula, P′ DG,t For DG, the output is uncertain, P DG,t is the predicted value of DG active output, It is an auxiliary variable of the photovoltaic output at the upper and lower boundaries, used to adjust the uncertainty interval of each time section. It is the fluctuation range of DG active output.

5. The method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty according to claim 3 is characterized in that: The DES optimization configuration model for the unbalanced active distribution network considering the uncertainty of wind and solar power is established in S22, and the compact form of the adaptive robust optimization model is adopted, and its expression is: The main problem is to minimize the objective function by adjusting the variable x, and the subproblem is to maximize the variable y under the worst case of uncertain parameter u.

6. The method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty according to claim 3 is characterized in that: The specific steps of the unbalanced active distribution network DES optimization calculation in S22 are: (1) Using the C&CG algorithm, the three-level min-max-min problem is decomposed into a single-level min main problem and a two-level max-min subproblem for solution. The main problem is: my OF1=C I +C M +C depore The sub-problems are: max minOF2=S LR +S LS Among them, S LR In order to reduce network loss and profit, S LS is the load transfer benefit, C I is the DES investment cost, C M is the DES operation and maintenance cost, C depore is the DES life loss cost. (2) Solve the main problem and sub-problems separately: the main problem is a conventional linear programming problem, which can be solved directly by calling the existing solver. The sub-problem is difficult to solve directly due to the existence of nonlinear power flow constraints and non-convex programming problems, and needs to be relaxed using SOCP; The nonlinear power flow constraint in the original problem is: By eliminating the quadratic terms in the constraints and the SOCP relaxation step, the above constraints can be transformed into the following form: (3) Since the subproblem is a max-min problem, the duality theory is used to transform the two-level model into a single-level linear programming model; First, the relaxed model is written as follows: stCy≤f Dy=g Ey=u Where C, D, E, Q, q, c, d, f, g are all known coefficient matrices, u is an uncertain parameter, and y is a decision variable. Through the Lagrange duality theory, the dual "max" model is obtained as follows: ||ω i ||2≤λ i ,i=1,...,n π1≤0 in min Oh, oh. max In the formula, π1, π2, π3 are dual variables, λ i ω i is a quadratic dual variable, n is the number of all uncertain scenarios; After the dual transformation, there is still a bilinear term u in the above formula T π3, which makes the model non-convex and unable to find the global optimal solution. Next, we introduce the dummy variable ζ t , using large M for transformation: -M(1-ζ t )+π 3,s ≤r s ≤M(1-ζ t )+π 3,s -Mz t ≤r s ≤Mζ t In the formula, r s is an auxiliary variable, M is a large constant, ζ t is a 0-1 variable, s is a set of nodes with DG; When t = 0, the bilinear term u T π3 gets u s =u min , corresponding to the worst scenario of uncertain variables; when ζ t = 1, the bilinear term u T π3 gets u s =u max , corresponding to the optimal scenario of uncertain variables; (4) Introducing uncertainty margin τ DG , by adjusting its size, the model conservatism is improved: In the formula, τ DG is the DG output uncertainty margin.

7. The method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty according to claim 1 is characterized in that: In the DES sequential configuration strategy considering comprehensive network loss sensitivity described in S3, network loss sensitivity analysis is used to determine the installation location and number of DES, and its expression is: In the formula, λ d , c are the network loss sensitivity weights corresponding to charging and discharging, which are set according to the actual charging and discharging duration and electricity price, so that λ d +λ c =1; are the network loss sensitivities corresponding to the r-th configuration charging and discharging respectively.

8. The method for configuring an unbalanced active distribution network energy storage system based on wind and solar uncertainty according to claim 1 is characterized in that: The rationality and superiority of the DES sequence configuration strategy are verified by simulation of a typical example system in S4, and the specific steps include: S41. Feasibility verification of algorithm results: Based on the proposed energy storage system configuration strategy, conduct feasibility analysis on the obtained results; S42. Verification of algorithm superiority: By comparing the calculation results with those of conventional methods, the superiority of the proposed algorithm is analyzed from the aspects of economy and result security.

Images