A method and device for affine adjustable robust optimization of reactive voltage in active power grid

By establishing a distributed photovoltaic controllable model, equivalent ZIP load model and battery energy storage model, combined with ellipsoid uncertainty, the problems of insufficient load model and low optimization efficiency are solved, and the voltage reactive power grid is realized. Robust optimization reduces conservatism and optimization costs.

CN115036931BActive Publication Date: 2024-11-22WUHAN INST OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210550229.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-05-20
Publication Date
2024-11-22
Estimated Expiration
2042-05-20

AI Technical Summary

Technical Problem

In the prior art, the load model is insufficiently considered, the optimization efficiency of the stochastic optimization method is low, and the conservatism of the robust optimization method cannot be overcome, resulting in the reactive control of the active distribution network voltage facing stochastic disturbance and conservatism.

Method used

Using distributed photovoltaic controllable model, equivalent ZIP load model based on linear regression method, and battery energy storage model considering cycle life loss, combined with ellipsoid uncertainty set, the voltage of the active power grid is established through a linear affine mechanism. The work affine adjustable robust optimization model is converted into a deterministic hybrid second-order cone planning model for solving.

Benefits of technology

It reduces the conservatism of the optimization results, improves the economy and efficiency of voltage reactive power control, reduces the blindness of finding optimization, and achieves more accurate voltage reactive power optimization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115036931B_ABST
    Figure CN115036931B_ABST
Patent Text Reader

Abstract

The present application provides a method and device for robust optimization of active power grid reactive power and voltage affine adjustable. The method includes collecting relevant parameters, the relevant parameters including line parameters of the active power grid, distributed photovoltaic output and load forecast reference values; based on the relevant parameters, establishing a source-load-storage operation model in the active power grid, and establishing voltage-reactive operation constraints of the active power grid; based on the source-load-storage operation model and the voltage-reactive operation constraints, establishing a voltage-reactive optimization objective function of the active power grid; based on the voltage-reactive optimization objective function and the uncertainty of distributed photovoltaic output, establishing a voltage-reactive affine adjustable robust optimization model of the active power grid; solving the voltage-reactive affine adjustable robust optimization model, and outputting the optimization results. The voltage-reactive affine adjustable robust optimization method for active distribution network establishes a linear affine relationship between decision variables and uncertainty variables, reduces the blindness of optimization, and thus reduces the conservatism of the optimization results.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of reactive power optimization of active power distribution networks, and in particular to an active power grid reactive power voltage affine adjustable robust optimization method and device. Background Art

[0002] With the large-scale layered distribution of distributed photovoltaics, battery energy storage, etc. connected to the distribution network, the distribution network has changed from passive to active, and the flow has changed from unidirectional flow to bidirectional flow, making the operation mode of the distribution network more diversified; on the other hand, the randomness and volatility of distributed photovoltaic output make the voltage and reactive power control of the active distribution network face the risk of random disturbance. Therefore, the coordinated operation of distributed photovoltaics, energy storage, and loads in spatial distribution is conducive to solving the problem of frequent voltage limit violations at high-proportion distributed photovoltaic access points, which plays an important supporting role in promoting the development goals of the distribution network to achieve clean, low-carbon, safe, and efficient.

[0003] How to achieve source-load-storage synergy to improve the economy of voltage-reactive power control of active distribution network? Existing voltage-reactive power optimization technology mainly works from two aspects: source-load-storage operation model and uncertainty optimization method. The source-load-storage operation model mainly establishes detailed models of distributed power sources, energy storage, and load models, but the existing source-load-storage operation model does not consider the impact of battery charging and discharging process on its cycle life; secondly, the load model adopts constant power model, and rarely considers ZIP load model. The existing active distribution network uncertainty optimization methods are mainly random optimization methods and robust optimization methods. The random optimization method needs to generate sample data from the probability distribution of uncertainty variables, and cluster and reduce the scene. The higher the dimension of its typical scene, the more accurate the optimization result, but the lower its optimization efficiency. To this end, the robust optimization method uses intervals, polyhedrons, ellipsoids and other sets to describe uncertainty and seek optimization results in the worst scenario. How to overcome the conservatism of the robust optimization method has always been a difficulty restricting its application. Summary of the invention

[0004] The purpose of the present invention is to overcome the above-mentioned technical deficiencies and provide an active power grid reactive voltage affine adjustable robust optimization method and device to solve the technical problems in the prior art of insufficient consideration of load models, low optimization efficiency of random optimization methods and inability to overcome the conservatism of robust optimization methods.

[0005] In order to achieve the above technical objectives, the technical solution of the present invention provides an active power grid reactive voltage affine adjustable robust optimization method, comprising:

[0006] Collect relevant parameters, including line parameters of the active power grid, distributed photovoltaic output and load forecast reference values;

[0007] Based on the relevant parameters, a source-load-storage operation model in the active power grid is established;

[0008] Based on the relevant parameters, establishing voltage and reactive power operation constraints of the active power grid;

[0009] Based on the source-load-storage operation model and the voltage-reactive operation constraints, a voltage-reactive optimization objective function of the active power grid is established;

[0010] Based on the voltage-reactive power optimization objective function and the uncertainty of the distributed photovoltaic output, a voltage-reactive power affine adjustable robust optimization model of the active power grid is established;

[0011] Solve the voltage-reactive affine adjustable robust optimization model and output the optimization result.

[0012] According to the above technical solution of the present invention, an active power grid reactive voltage affine adjustable robust optimization method is provided, and the source-load-storage operation model specifically includes:

[0013] Distributed photovoltaic controllable model, equivalent ZIP load model based on linear regression method and battery energy storage model considering cycle life degradation.

[0014] According to the above technical solution of the present invention, an active power grid reactive voltage affine adjustable robust optimization method is provided, wherein the voltage reactive operation constraint includes:

[0015] The single-unit constraints of the distributed photovoltaic controllable model, the single-unit constraints of the ZIP load model and the constraints of the battery energy storage model.

[0016] According to the above technical solution of the present invention, an active power grid reactive voltage affine adjustable robust optimization method is provided, wherein the voltage reactive operation constraint further includes:

[0017] Power system flow constraints, node voltage constraints, branch current constraints, on-load tap-changing transformer operation constraints, group switching capacitor operation constraints, SVC operation constraints and distribution transformer gateway power constraints.

[0018] According to the above technical solution of the present invention, an active power grid reactive voltage affine adjustable robust optimization method is provided, and the voltage reactive optimization objective function specifically includes:

[0019] Network loss costs, on-load tap-changing transformer operation costs, capacitor switching costs, and battery energy storage cycle life reduction costs.

[0020] According to the active grid reactive voltage affine adjustable robust optimization method provided by the technical solution of the present invention, the determination of the uncertainty of the distributed photovoltaic output specifically includes:

[0021] The uncertainty of distributed photovoltaic output is represented by an ellipsoid uncertainty set.

[0022] According to the above technical solution of the present invention, a method for robust optimization of active power grid reactive power affine adjustable voltage is provided. Based on the voltage reactive power optimization objective function and the uncertainty of the distributed photovoltaic output, a robust optimization model of active power grid reactive power affine adjustable voltage is established, which specifically includes:

[0023] The relationship between the adjustable decision variables and the uncertainty in the voltage-reactive power optimization objective function is established through a linear affine mechanism, and a voltage-reactive power affine adjustable robust optimization model for the active power grid is established.

[0024] According to the above technical solution of the present invention, an active power grid reactive power voltage affine adjustable robust optimization method is provided, the voltage reactive power affine adjustable robust optimization model is solved, and the optimization result is output, specifically including:

[0025] Based on robust equivalence, the voltage-reactive affine adjustable robust optimization model is transformed into a deterministic hybrid second-order cone programming model for solution, and the optimization result is output.

[0026] The technical solution of the present invention also provides an active power grid reactive voltage affine adjustable robust optimization device, comprising:

[0027] A collection unit, used to collect relevant parameters, including line parameters of the active power grid, distributed photovoltaic output and load prediction reference values;

[0028] A first model unit, used to establish a source-load-storage operation model in the active power grid based on the relevant parameters;

[0029] A constraint unit, used for establishing voltage and reactive power operation constraints of the active power grid based on the relevant parameters;

[0030] A second model unit is used to establish a voltage-reactive power optimization objective function of the active power grid based on the source-load-storage operation model and the voltage-reactive power operation constraint;

[0031] A third model unit is used to establish a voltage-reactive affine adjustable robust optimization model of the active power grid based on the voltage-reactive optimization objective function and the uncertainty of the distributed photovoltaic output;

[0032] The solving unit is used to solve the voltage-reactive affine adjustable robust optimization model and output the optimization result.

[0033] The technical solution of the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that when the processor executes the program, the steps of the active grid reactive voltage affine adjustable robust optimization method as described in any one of the above technical solutions provided by the present invention are implemented.

[0034] Compared with the prior art, the beneficial effects of the present invention include: the active distribution network voltage reactive affine adjustable robust optimization method establishes a linear affine relationship between decision variables and uncertainty variables, reduces the blindness of optimization, and thus reduces the conservatism of the optimization result. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flow chart of an active power grid reactive voltage affine adjustable robust optimization method provided by the present invention;

[0036] Figure 2 It is a structural schematic diagram of a radially operated active power distribution network provided by the present invention;

[0037] Figure 3 It is a schematic diagram of a branch model including an on-load tap-changing transformer provided by the present invention;

[0038] Figure 4 It is a structural schematic diagram of an active power grid reactive voltage affine adjustable robust optimization device provided by the present invention;

[0039] Figure 5 It is a flow chart of a robust optimization method for voltage and reactive power affine adjustable active distribution network considering source-load-storage coordination provided by the present invention;

[0040] Figure 6 It is a structural schematic diagram of an improved IEEE-33 node system provided by the present invention;

[0041] Figure 7 A schematic diagram of the physical structure of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the following is a diagram of the present invention. Figure 1-Figure 7 It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0043] The present invention provides an active power grid reactive voltage affine adjustable robust optimization method. Figure 1 FIG. 1 is a flow chart of an active grid reactive voltage affine adjustable robust optimization method provided by the present invention, such as Figure 1 As shown, the method includes:

[0044] Step 110, collecting relevant parameters, wherein the relevant parameters include line parameters of the active power grid, distributed photovoltaic output and load prediction reference values;

[0045] Step 120, establishing a source-load-storage operation model in the active power grid based on the relevant parameters;

[0046] Step 130, establishing voltage and reactive power operation constraints of the active power grid based on the relevant parameters;

[0047] Step 140, establishing a voltage-reactive power optimization objective function of the active power grid based on the source-load-storage operation model and the voltage-reactive power operation constraint;

[0048] Step 150, based on the voltage-reactive power optimization objective function and the uncertainty of the distributed photovoltaic output, establishing a voltage-reactive power affine adjustable robust optimization model of the active power grid;

[0049] Step 160, solving the voltage-reactive affine adjustable robust optimization model and outputting the optimization result.

[0050] Specifically, the source-load-storage operation model and voltage-reactive power operation constraints are first established based on the collected relevant parameters, and then a deterministic voltage-reactive power optimization objective function is established. The uncertainty of distributed photovoltaic output is then introduced to construct an uncertain voltage-reactive power affine adjustable robust optimization model. Finally, the voltage-reactive power affine adjustable robust optimization model is converted into a deterministic model for solution to obtain the optimization result.

[0051] The active grid reactive voltage affine adjustable robust optimization method provided by the embodiment of the present invention establishes a linear affine relationship between decision variables and uncertainty variables, reduces the blindness of optimization, and thus reduces the conservatism of the optimization result.

[0052] Based on the above embodiment, in this method, the source-load-storage operation model specifically includes:

[0053] Distributed photovoltaic controllable model, equivalent ZIP load model based on linear regression method and battery energy storage model considering cycle life degradation.

[0054] Specifically, the source-load-storage operation model mainly establishes detailed models of distributed power sources, energy storage, and load models, but the existing source-load-storage operation models do not consider the impact of the battery charging and discharging process on its cycle life; secondly, the load models all adopt constant power models, and the ZIP load model is rarely considered. This embodiment, on the basis of establishing a distributed photovoltaic controllable model, a battery daily cycle life model, and a ZIP load model, constructs an active distribution network voltage and reactive power optimization model under the coordination of source-load-storage.

[0055] (1) Establishing a controllable model for distributed photovoltaics

[0056] The control strategy of auxiliary services adopted by distributed photovoltaic inverters can simultaneously realize the decoupled optimal control of the active output and reactive output of distributed photovoltaics. It can be concluded that the operating area of ​​distributed photovoltaic inverter j is:

[0057]

[0058] Where: are the active power and reactive power of the distributed photovoltaic connected to node j in period t respectively; is the predicted power value of the distributed photovoltaic connected to node j during period t; is the apparent power of the distributed photovoltaic inverter connected to node j; θ j is the power factor angle of the distributed photovoltaic inverter connected to node j.

[0059] Based on the linearization method of analytic geometry, the second circular constraint in the above equation is transformed into an equal n polygon for approximation. When n→∞, this equal n polygon approaches the circle infinitely. Therefore, the linearization expression of the equal n polygon in the above equation is:

[0060]

[0061] Where: k∈[0,n-1].

[0062] (2) Establishing an equivalent ZIP load model based on linear regression method

[0063] The ZIP load model is represented by a second-order polynomial:

[0064]

[0065] Where: are the proportional coefficients of constant power load, constant load and constant current load at node j respectively; are respectively the active power demand and reactive power demand of the load at node j under rated voltage; U j,t is the voltage amplitude of node j in period t. are the active power and reactive power of the load at node j respectively.

[0066] Since the ZIP load model is difficult to embed into the second-order cone power flow model of the distribution network, the present invention adopts the linear regression method to approximately transform the ZIP load model into the equivalent ZP load model, which is:

[0067]

[0068] Where: and is the equivalent coefficient in the ZP load model; For error.

[0069] and The approximate calculation method of can be approximated by Taylor series or estimated by linear regression method (LRM). The linear regression method is based on minimizing the sum of square errors and sampling within the specified voltage amplitude range to obtain n voltage amplitude data. Let U j,t is a real vector of sampled voltage amplitudes, is the vector U j,t The new vector is obtained by raising each element in to the mth power. According to the linear regression equation:

[0070]

[0071] When the error vector When it is approximately equal to 0, the optimal equivalent ZP load model parameters can be obtained. Therefore, the approximate expression of LRM can be expressed as:

[0072]

[0073] Where: is defined as a matrix The analytical expression of the transformation matrix C is:

[0074]

[0075] Where: The overline represents the vector average operator, then is the vector U j,t Therefore, the transformation matrix C is optimized and adjusted according to the voltage amplitude sampling interval to ensure sufficient approximate accuracy.

[0076] (3) Establish a battery energy storage model that takes into account cycle life degradation

[0077] 1) Daily depreciation cost of battery energy storage cycle life

[0078] The daily depreciation cost of battery energy storage cycle life is mainly due to the loss cost caused by the charge and discharge process within a day. Here, the full life cycle cost of battery energy storage is converted to daily according to 365 days. The full life cycle cost of battery energy storage is mainly composed of the investment cost of battery energy storage and its auxiliary equipment, which is determined by the rated capacity and rated power of the battery energy storage, that is:

[0079]

[0080] Where: Respectively represent the rated capacity and rated power of the battery energy storage at node j; ΩESS A collection of nodes that are installed to store battery energy. They represent the annualized investment cost per unit capacity and per unit power of battery energy storage at node j respectively, and the calculation formula is:

[0081]

[0082] Where: They represent the unit capacity investment cost and unit power investment cost of battery energy storage at node j respectively; r is the discount rate; is the cycle life of the battery energy storage at node j.

[0083] From the daily depreciation cost formula of battery energy storage cycle life, it can be seen that the daily depreciation cost of battery energy storage cycle life is mainly affected by the cycle life. The impact of It mainly depends on the number of cycles of battery charge and discharge, and its expression is as follows:

[0084]

[0085] Where: The maximum number of cycles that a battery can be discharged at the same depth of discharge until the battery life is exhausted; It is the number of cycles in which the battery is discharged at the same depth of discharge in one day, that is, the number of daily cycles.

[0086] The above formula is used to obtain the cycle life by discharging at the same discharge depth. However, in the actual operation of the distribution network, the charging and discharging behavior of battery energy storage does not follow a repetitive pattern, and even a complete charging and discharging cycle may not be completed during the entire life of the battery. The present invention calculates the cycle life at different discharge depths based on the power function of the battery cycle life, which can be expressed as:

[0087]

[0088] Where: DOD j is the discharge depth of the battery; N j,DOD It is the number of cycles that the battery is discharged at the same depth of discharge within a period of time; It is the equivalent number of cycles of the battery being discharged at 100% depth of discharge in the same time. It is expressed as the battery life loss; kp is the constant obtained by fitting, and this parameter is provided by the battery manufacturer.

[0089] The cycle life of the battery mainly depends on the discharge process of the battery. Therefore, the battery life loss can be expressed as:

[0090]

[0091] Where: s represents the time of a battery discharge process, which includes the continuous time period t during which the battery is discharged at different depths of discharge; is the discharge depth of the battery at the end of discharge within time s; ΔDOD is the depth of discharge of the battery at the beginning of discharge within time s; j,s is the discharge depth increment of the battery in time s; when the battery is discharged at a depth of ΔDOD j,s When discharging, is the equivalent number of cycles of the battery being discharged at 100% depth of discharge within time s.

[0092] Combined with the above formula, the equivalent number of cycles of a battery based on different discharge depths in a day can be expressed as:

[0093]

[0094] Where: is the equivalent number of cycles of the battery based on different depths of discharge in a day, that is, the daily equivalent number of cycles. T is the number of scheduled operation hours.

[0095] Therefore, the cycle life It can be expressed as

[0096]

[0097] Where: The number of cycles a battery can be discharged at 100% depth of discharge until the battery life is exhausted.

[0098] The cycle life of the battery depends mainly on the discharge depth increment ΔDOD j,s , but the endpoint that determines the depth of discharge and It has strong randomness and it is difficult to obtain its data and apply it to the battery energy storage optimization model. Therefore, the present invention adopts an efficient linearization method to calculate ΔDOD j,s , as follows

[0099] First, the increment of the depth of discharge during the period t when the battery is charged and discharged can be expressed as:

[0100] ΔDoD j,t =DoD j,t -DoD j,t-1 =SoC j,t-1 -SoC j,t

[0101] Define the discharge depth increment during the discharge process for:

[0102]

[0103] Where: Represented as a 0-1 variable, when When the battery is in the discharge state, further It can be linearized as:

[0104]

[0105] Where: M is a sufficiently large positive number.

[0106] Because the battery discharge process may be a continuous period of time, define a 0-1 variable To represent the continuous discharge period:

[0107]

[0108] Where: In a discharge cycle s, t = 1 represents the start of discharge. If the battery is still in a continuous discharge state in the next period t, then otherwise further It can be linearized as:

[0109]

[0110] Finally, the cumulative increment of the discharge depth at time period t is It can be expressed as:

[0111]

[0112] In the formula: In a continuous discharge period, is a cumulative value of the discharge depth with overlapping time intervals, but what is actually needed is the incremental value of the discharge depth in the last time interval of the discharge cycle, which is also the incremental value of the discharge depth of the entire discharge cycle. Therefore, another 0-1 variable is defined To determine whether the battery discharge period is in the last period of the discharge cycle:

[0113]

[0114]

[0115] Where: If t is the last time interval of a discharge cycle, then otherwise So the cumulative discharge depth increment at the last time interval of each discharge cycle is It can be expressed as:

[0116]

[0117] The equivalent number of battery cycles per day is for:

[0118]

[0119] The power function shown in the above formula is piecewise linearized as follows:

[0120]

[0121] Where: D is the total number of piecewise linearization, g j,t,d is a 0-1 variable, indicating the increment of the discharge depth of the BESS at time t In the dth segment; K j,d and B j,d is the piecewise linear fitting coefficient; The upper and lower limits of the discharge depth increment for each segment.

[0122] Therefore, the daily depreciation cost of battery energy storage cycle life is C ESS The linearized cost model is:

[0123]

[0124] Where: and are the linear fitting coefficients of each order.

[0125] 2) Battery energy storage operation constraints

[0126] Battery energy storage has three states: charging, discharging, and static (neither charging nor discharging). 0-1 variables are introduced Indicates that the battery energy storage at node j is in the charging state during period t, a 0-1 variable Indicates that the battery energy storage at node j is in a discharging state during period t. Since the battery energy storage can only be in one of the three states during the same period, we have:

[0127]

[0128] During operation, battery energy storage is also subject to constraints on charge and discharge rates, energy balance, and charge and discharge processes, which can be expressed as:

[0129] a) Charge and discharge rate constraints

[0130]

[0131] Where: are the upper and lower limits of the charging power of the battery energy storage at node j respectively; They are the upper and lower limits of the battery energy storage discharge power at node j respectively; are the charging and discharging powers of the battery energy storage at node j in time period t, respectively.

[0132] b) Charge and discharge energy balance constraints

[0133]

[0134] Where: SoC j,t are the state of charge of the battery energy storage at node j during period t; η ch,j , η dis,j are the charging and discharging efficiencies of the battery energy storage at node j, respectively; Δt is the scheduling time interval; SoC j are the upper and lower limits of the state of charge of the battery energy storage at node j respectively.

[0135] Based on the above embodiment, in the method, the voltage reactive operation constraint includes:

[0136] The single-unit constraints of the distributed photovoltaic controllable model, the single-unit constraints of the ZIP load model and the constraints of the battery energy storage model.

[0137] Specifically, the voltage and reactive power operation constraints include three individual constraints: a distributed photovoltaic controllable model, a ZIP load model, and a battery energy storage operation constraint.

[0138] Based on the above embodiment, in the method, the voltage reactive operation constraint further includes:

[0139] Power system flow constraints, node voltage constraints, branch current constraints, on-load tap-changing transformer operation constraints, group switching capacitor operation constraints, SVC operation constraints and distribution transformer gateway power constraints.

[0140] Specifically, in addition to the three individual constraints of distributed photovoltaic controllable model, ZIP load model and battery energy storage operation constraints introduced in the above embodiments, the operation constraints of the active distribution network also include system flow constraints, node voltage constraints, branch current constraints, on-load tap-changing transformer operation constraints, group switching capacitor operation constraints, SVC operation constraints, and distribution transformer gateway power constraints.

[0141] 1) Power flow constraints of active distribution networks

[0142] Figure 2 is a schematic diagram of the structure of the radially operated active distribution network provided by the present invention, such as Figure 2 As shown in Figure 2, the power flow equation constraints using second-order cone relaxation are as follows:

[0143]

[0144]

[0145] Where: set u(j) is the set of head nodes of the branch with j as the terminal node in the distribution network; set v(j) is the set of terminal nodes of the branch with j as the head node in the distribution network; U i,t is the voltage amplitude of node i in period t; I ij,t is the current amplitude of branch (i, j) in period t, l ij,t =(I ij,t ) 2 , V i,t =(U i,t ) 2 .P ij,t , Q ij,t are the active power and reactive power of branch (i, j) in period t respectively; P j,t , Q j,t are the active power and reactive power injected into node j respectively; are respectively the active power and reactive power injected into the main grid by the gateway at node j during period t; are the continuous compensation power of the SVC connected to node j and the discrete compensation power of the group switching capacitor bank CB respectively; ij is the reactance of branch (i, j), r ij is the resistance of branch (i, j); T is the number of scheduling operation hours.

[0146] 2) Node voltage and branch current constraints

[0147]

[0148] Where: are the upper and lower limits of the voltage amplitude at node j respectively; is the upper limit of the current amplitude of branch (i, j).

[0149] 3) On-load tap-changing transformer (OLTC) operation constraints

[0150] Figure 3 Schematic diagram of a branch model including an on-load tap-changing transformer provided by the present invention, such as Figure 3 As shown, Figure 3 An on-load tap-changing transformer is added between the intermediate nodes i and j, and an auxiliary node o is added. The OLTC operation has the following constraints:

[0151]

[0152] Where: V o,t N is the square of the voltage amplitude of the auxiliary node o during the period t; ij,t is the optimal gear of the transformer of branch (i, j) in period t, which is an integer variable; Δk ijk is the increment of each level of the transformer tap of branch (i, j); ij,t K is the transformation ratio of the transformer of branch (i, j) in period t; ij is the number of gear positions of the transformer tap of branch (i, j); They are the upper and lower limits of the transformer ratio of branch (i, j) respectively.

[0153] Further, the positive integer N ij,t Convert to binary number λ ij,n,t ∈{0,1}, n=0,1,…K ij , and linearize it by combining Big-M, we can get

[0154]

[0155] Where: m ij,t 、x ij,n,t =λ ij,n,t V j,t ,y ij,n,t =λ ij,n,t m ij,t is an intermediate auxiliary variable.

[0156] 4) Operation constraints of group switching capacitors (CB)

[0157] The group switching capacitor is a discrete reactive power compensation mechanical device, and its frequent operation needs to be limited to prevent it from affecting its service life. The specific constraints are as follows:

[0158]

[0159] Where: is the compensation power of the CB connected to the jth node in period t; is the number of CBs connected to the jth node put into operation in period t; Compensate the power of each group of CBs connected to the jth node in time period t; The number of CB actions within T cycles per day is limited;

[0160] Introducing auxiliary variables and Represents the positive and negative changes in the number of CB groups at node i during period t, and the change values ​​are all expressed as non-negative numbers, which can be linearized as follows:

[0161]

[0162] 5) SVC operation constraints

[0163] The operating constraints of the continuously regulated reactive power compensation device SVC are as follows:

[0164]

[0165] Where: are the lower limit and upper limit of the adjustable reactive power of the SVC connected to node j respectively.

[0166] 6) Power Constraints on Distribution Gateway Ports

[0167]

[0168] Where: are the upper and lower limits of active power at the node j of the distribution network respectively; are the upper and lower limits of reactive power at node j in the distribution network respectively.

[0169] Based on the above embodiment, in this method, the voltage-reactive power optimization objective function specifically includes:

[0170] Network loss costs, on-load tap-changing transformer operation costs, capacitor switching costs, and battery energy storage cycle life reduction costs.

[0171] Specifically, the voltage and reactive power optimization objective function of the active distribution network is established, including network loss cost, on-load tap-changing transformer operation cost, capacitor switching cost, and battery energy storage cycle life daily loss cost, as shown below:

[0172]

[0173] Where: c loss 、c OLTC and c CB They are network loss cost coefficient, on-load tap-changing transformer operation cost coefficient and capacitor switching cost coefficient respectively; Ω L is the branch set; Ω CB It is the set of CB nodes.

[0174] Based on the above embodiment, in the method, the determination of the uncertainty of the distributed photovoltaic output specifically includes:

[0175] The uncertainty of distributed photovoltaic output is represented by an ellipsoid uncertainty set.

[0176] Specifically, an ellipsoid uncertainty set is used to describe the uncertainty of distributed photovoltaic output.

[0177] Based on the above embodiment, in the method, based on the voltage-reactive power optimization objective function and the uncertainty of the distributed photovoltaic output, a voltage-reactive power affine adjustable robust optimization model of the active power grid is established, which specifically includes:

[0178] The relationship between the adjustable decision variables and the uncertainty in the voltage-reactive power optimization objective function is established through a linear affine mechanism, and a voltage-reactive power affine adjustable robust optimization model for the active power grid is established.

[0179] Specifically, an ellipsoid uncertainty set is used to represent the uncertainty of distributed photovoltaic output, and the relationship between adjustable decision variables and uncertainty variables is established through a linear affine mechanism to establish a voltage-reactive AARO (Affinely Adjustable Robust Optimization, AARO) model of active distribution network.

[0180] 1) Establish an ellipsoid set of distributed photovoltaic output

[0181] The present invention uses an ellipsoid set to describe the uncertainty of distributed photovoltaic output, as shown below:

[0182]

[0183] Where: is the distributed photovoltaic active power vector; is the expected value vector of distributed photovoltaic output, is the fluctuation vector of distributed photovoltaic output, N PV is the number of distributed photovoltaics; ∑ is The covariance matrix of is obtained from historical data; Λ is the radius of the ellipsoid, also called the uncertainty budget.

[0184] 2) Establish the voltage-reactive AARO model of active distribution network

[0185] The traditional adjustable robust optimization model of distribution network is divided into adjustable decision variables and non-adjustable decision variables. The non-adjustable decision variables cannot change regardless of how the uncertainty variables change. Therefore, the non-adjustable decision variables are determined before the uncertainty variables are realized; while the adjustable decision variables are based on the set of uncertainty variables and are adjusted along with the uncertainty variables. Therefore, in essence, the traditional adjustable robust optimization model of distribution network is a two-stage optimization model that requires iterative solution.

[0186] To this end, this embodiment proposes a voltage-reactive AARO model for an active distribution network. The voltage-reactive AARO model uses a linear affine mechanism to establish the relationship between adjustable decision variables and uncertainty variables, so that the adjustable decision variables are adjusted within the adjustable range according to the linear affine rule, rather than performing disordered optimization within the entire adjustable range. Therefore, the voltage-reactive AARO model is established by establishing a linear affine relationship between the adjustable decision variables (gateway power, distributed photovoltaic reactive output, SVC reactive output, CB reactive output) in the active distribution network and the uncertainty of distributed photovoltaic active output, as shown below:

[0187]

[0188] Where: y 0 is the expected operating point of the adjustable decision variable; ζ is the affine proportional coefficient.

[0189] The voltage-reactive AARO model of active distribution network is equivalent to adding an affine linear decision mechanism to the traditional adjustable robust optimization model of distribution network, that is, the above constraints are valid for any uncertain quantity in the uncertainty set, which simplifies the model structure.

[0190] Define x as an unadjustable decision variable and y as an adjustable decision variable. Based on the active distribution network voltage-reactive power optimization model established in the second and third steps and the above-mentioned linear affine strategy, the vector form of the active distribution network voltage-reactive power AARO model is obtained:

[0191]

[0192] Where: Represents an unadjustable decision variable, which is a decision variable before the uncertainty variable is obtained. The adjustable decision variables are decision variables that change in real time after the uncertainty variables are obtained. 1 、c 2 It is represented as the coefficient matrix corresponding to the objective function; R, A, B, D, G, C, g are the coefficient matrices of the variables under the corresponding constraints; r, h, w are constant column vectors. The first row of constraints in the formula represents the constraints related only to x, including battery energy storage operation constraints, branch power balance constraints and voltage drop constraints in the second-order cone power flow model, node voltage and branch current constraints, and OLTC operation constraints; the second row of constraints in the formula represents the second-order cone relaxation constraints in the second-order cone power flow model; the third row of constraints in the formula represents the coupling constraints related to x and y, and the node injection power balance constraints in the second-order cone power flow model; the fourth row of constraints in the formula represents the constraints related only to y, including SVC operation constraints, gateway power constraints, and battery energy storage operation constraints.

[0193] Based on the above embodiment, in the method, solving the voltage-reactive affine adjustable robust optimization model and outputting the optimization result specifically includes:

[0194] Based on robust equivalence, the voltage-reactive affine adjustable robust optimization model is transformed into a deterministic hybrid second-order cone programming model for solution, and the optimization result is output.

[0195] Specifically, the third constraint in the active distribution network voltage reactive power AARO model contains uncertainty ΔP PV , for ΔP tPV ∈Z Ellipse , the inequality constraint is robust if and only if its robust counterpart (RC) holds, namely:

[0196]

[0197] In the formula, a i , b i , D i 、h i are the coefficient matrices A, B, D and the i-th row of the constant vector h respectively. Let L = Σ 1 / 2 , ΔP PV Variable substitution is the standard norm δ, Ω t It can be expressed as The formula can be further equivalent to:

[0198]

[0199] That is

[0200]

[0201] By taking the uncertain variable ΔP PV The constraints are converted into corresponding mixed integer second-order cone robust equivalent constraints. Therefore, the active distribution network voltage-reactive AARO model established by the present invention is converted into a deterministic mixed integer second-order cone programming model for solution, and the commercial solver Cplex can be directly called for solution.

[0202] The present invention also provides an active power grid reactive voltage affine adjustable robust optimization device, Figure 4 is a structural schematic diagram of an active grid reactive voltage affine adjustable robust optimization device provided by the present invention, such as Figure 4 As shown, the device includes a collection unit 410, a first model unit 420, a constraint unit 430, a second model unit 440, a third model unit 450 and a solution unit 460, wherein:

[0203] The acquisition unit 410 is used to acquire relevant parameters, including line parameters of the active power grid, distributed photovoltaic output and load prediction reference values;

[0204] A first model unit 420, configured to establish a source-load-storage operation model in the active power grid based on the relevant parameters;

[0205] A constraint unit 430, configured to establish voltage and reactive power operation constraints of the active power grid based on the relevant parameters;

[0206] A second model unit 440 is used to establish a voltage-reactive power optimization objective function of the active power grid based on the source-load-storage operation model and the voltage-reactive power operation constraint;

[0207] A third model unit 450 is used to establish a voltage-reactive affine adjustable robust optimization model of the active power grid based on the voltage-reactive optimization objective function and the uncertainty of the distributed photovoltaic output;

[0208] The solving unit 460 is used to solve the voltage-reactive power affine adjustable robust optimization model and output the optimization result.

[0209] The device provided by the embodiment of the present invention establishes a linear affine relationship between decision variables and uncertainty variables, reduces the blindness of optimization, and thus reduces the conservatism of optimization results.

[0210] Based on the above embodiment, in the device, the source-load-storage operation model specifically includes:

[0211] Distributed photovoltaic controllable model, equivalent ZIP load model based on linear regression method and battery energy storage model considering cycle life degradation.

[0212] Based on the above embodiment, in the device, the voltage reactive operation constraint includes:

[0213] The single-unit constraints of the distributed photovoltaic controllable model, the single-unit constraints of the ZIP load model and the constraints of the battery energy storage model.

[0214] Based on the above embodiment, in the device, the voltage reactive operation constraint further includes:

[0215] Power system flow constraints, node voltage constraints, branch current constraints, on-load tap-changing transformer operation constraints, group switching capacitor operation constraints, SVC operation constraints and distribution transformer gateway power constraints.

[0216] Based on the above embodiment, in the device, the voltage-reactive power optimization objective function specifically includes:

[0217] Network loss costs, on-load tap-changing transformer operation costs, capacitor switching costs, and battery energy storage cycle life reduction costs.

[0218] Based on the above embodiment, in the device, the determination of the uncertainty of the distributed photovoltaic output specifically includes:

[0219] The uncertainty of distributed photovoltaic output is represented by an ellipsoid uncertainty set.

[0220] Based on the above embodiment, in the device, based on the voltage-reactive power optimization objective function and the uncertainty of the distributed photovoltaic output, a voltage-reactive power affine adjustable robust optimization model of the active power grid is established, which specifically includes:

[0221] The relationship between the adjustable decision variables and the uncertainty in the voltage-reactive power optimization objective function is established through a linear affine mechanism, and a voltage-reactive power affine adjustable robust optimization model for the active power grid is established.

[0222] Based on the above embodiment, in the device, solving the voltage-reactive affine adjustable robust optimization model and outputting the optimization result specifically includes:

[0223] Based on robust equivalence, the voltage-reactive affine adjustable robust optimization model is transformed into a deterministic hybrid second-order cone programming model for solution, and the optimization result is output.

[0224] Based on the above embodiments, compared with the existing active distribution network voltage and reactive power optimization method, the present invention can achieve the following beneficial effects:

[0225] 1) Compared with the existing robust optimization methods, the voltage-reactive power AARO method based on ellipsoid sets of the present invention converts the robust constraints into deterministic robust equivalent constraints one by one, without increasing the constraint dimension during the conversion process, and does not require iterative solution.

[0226] 2) Compared with the interval set AARO method and the polyhedron set AARO method, the voltage-reactive power optimization cost solved by the voltage-reactive power AARO method based on the ellipsoid set of the present invention is smaller, and the node voltage distribution fluctuates within a smaller range, indicating that the voltage-reactive power AARO method based on the ellipsoid set of the present invention has the best robust performance, that is, the least conservatism.

[0227] In order to more clearly illustrate the technical solution and advantages of the present invention, Figure 5 It is a flow chart of a robust optimization method for voltage and reactive power affine adjustable active distribution network considering source-load-storage coordination provided by the present invention. Figure 6 is a schematic diagram of the structure of the improved IEEE-33 node system provided by the present invention, such as Figure 5 and Figure 6 The present invention is further described in detail.

[0228] The active distribution network voltage reactive affine adjustable robust optimization method considering source-load-storage coordination of the present invention comprises the following steps:

[0229] Step 1: Initialization, input the improved IEEE-33 node distribution network line parameters, distributed photovoltaic output and load forecast benchmark values; Table 1 is the parameter settings of the access system. Specifically, Table 1 gives the parameter settings of PV, BESS, CB, and OLTC of the access system. The system reference voltage is 12.66kV, and the reference capacity is 10MVA. The gateway power active power range is: 0~5000kW, the reactive power range is: 0~2500kvar; the branch current amplitude upper limit is 0.4pu; the node voltage amplitude range is 0.95pu~1.05pu. Based on the three-year historical data of photovoltaics in a certain area, a random model of distributed photovoltaic output was constructed, and the expected and covariance matrices of distributed photovoltaic output were obtained, and the uncertainty budget Λ=1. Assume that all loads are ZIP loads, and the three proportional systems are respectively The network loss cost coefficient is 0.2 yuan / (kW·h), the on-load tap-changing transformer operation cost coefficient is 0.5 yuan / (kW·h), and the capacitor bank operation cost coefficient is 0.2 yuan / (kW·h).

[0230] Table 1 Parameter settings of access system

[0231]

[0232] Step 2: Establish a source-load-storage operation model in the active distribution network, which includes a distributed photovoltaic controllable model, an equivalent ZP load model based on the linear regression method, and a battery energy storage model considering cycle life degradation;

[0233] (1) Establishing a controllable model for distributed photovoltaics

[0234]

[0235] Where: are the active power and reactive power of the distributed photovoltaic connected to node j in period t respectively; is the predicted power value of the distributed photovoltaic connected to node j during period t; is the apparent power of the distributed photovoltaic inverter connected to node j; θ j is the power factor angle of the distributed photovoltaic inverter connected to node j.

[0236] (2) Establishing an equivalent ZP load model based on linear regression method

[0237] Since the ZIP load model is difficult to embed into the second-order cone power flow model of the distribution network, the present invention adopts the linear regression method to approximately transform the ZIP load model into the equivalent ZP load model, which is:

[0238]

[0239] Where: Hekou is the equivalent coefficient in the ZP load model; For error.

[0240] and The approximate calculation method of can be approximated by Taylor series or estimated by linear regression method (LRM). The linear regression method is based on minimizing the sum of square errors and sampling within the specified voltage amplitude range to obtain n voltage amplitude data. Let U j,t is a real vector of sampled voltage amplitudes, is the vector U j,t The new vector is obtained by raising each element in to the power of m. When it is approximately equal to 0, the optimal equivalent ZP load model parameters can be obtained. Therefore, the approximate expression of LRM can be expressed as:

[0241]

[0242] Where: is defined as a matrix The analytical expression of the transformation matrix C is:

[0243]

[0244] Where: The overline represents the vector average operator, then is the vector U j,t Therefore, the transformation matrix C is optimized and adjusted according to the voltage amplitude sampling interval to ensure sufficient approximate accuracy.

[0245] (3) Establish a battery energy storage model that takes into account cycle life degradation

[0246] 1) Daily depreciation cost of battery energy storage cycle life

[0247] The daily depreciation cost of battery energy storage cycle life is mainly due to the loss cost caused by the charge and discharge process within a day. Here, the full life cycle cost of battery energy storage is converted to daily according to 365 days. The full life cycle cost of battery energy storage is mainly composed of the investment cost of battery energy storage and its auxiliary equipment, which is determined by the rated capacity and rated power of the battery energy storage, that is:

[0248]

[0249] Where: Respectively represent the rated capacity and rated power of the battery energy storage at node j; Ω ESS A collection of nodes installed for battery energy storage. They represent the annualized investment cost per unit capacity and per unit power of battery energy storage at node j respectively, and the calculation formula is:

[0250]

[0251] Where: They represent the unit capacity investment cost and unit power investment cost of battery energy storage at node j respectively; r is the discount rate; is the cycle life of the battery energy storage at node j.

[0252] From the daily depreciation cost formula of battery energy storage cycle life, it can be seen that the daily depreciation cost of battery energy storage cycle life is mainly affected by the cycle life. The impact of It mainly depends on the number of cycles of battery charge and discharge, and its expression is as follows:

[0253]

[0254] Where: The maximum number of cycles that a battery can be discharged at the same depth of discharge until the battery life is exhausted; It is the number of cycles in which the battery is discharged at the same depth of discharge in one day, that is, the number of daily cycles.

[0255] The above formula is used to obtain the cycle life by discharging at the same discharge depth. However, in the actual operation of the distribution network, the charging and discharging behavior of battery energy storage does not follow a repetitive pattern, and even a complete charging and discharging cycle may not be completed during the entire life of the battery. The present invention calculates the cycle life at different discharge depths based on the power function of the battery cycle life, which can be expressed as:

[0256]

[0257] Where: DOD j is the discharge depth of the battery; N j,DOD It is the number of cycles that the battery is discharged at the same depth of discharge within a period of time; It is the equivalent number of cycles of the battery being discharged at 100% depth of discharge in the same time. It is expressed as the battery life loss; kp is the constant obtained by fitting, and this parameter is provided by the battery manufacturer.

[0258] The cycle life of the battery mainly depends on the discharge process of the battery. Therefore, the battery life loss can be expressed as:

[0259]

[0260] Where: s represents the time of a battery discharge process, which includes the continuous time period t during which the battery is discharged at different depths of discharge; is the discharge depth of the battery at the end of discharge within time s; ΔDOD is the depth of discharge of the battery at the beginning of discharge within time s; j,s is the discharge depth increment of the battery in time s; when the battery is discharged at a depth of ΔDOD j,s When discharging, is the equivalent number of cycles of the battery being discharged at 100% depth of discharge within time s.

[0261] Combined with the above formula, the equivalent number of cycles of a battery based on different discharge depths in a day can be expressed as:

[0262]

[0263] Where: is the equivalent number of cycles of the battery based on different depths of discharge in a day, that is, the daily equivalent number of cycles. T is the number of scheduled operation hours.

[0264] Therefore, the cycle life It can be expressed as:

[0265]

[0266] Where: The number of cycles a battery can be discharged at 100% depth of discharge until the battery life is exhausted.

[0267] The cycle life of the battery depends mainly on the discharge depth increment ΔDOD j,s , but the endpoint that determines the depth of discharge and It has strong randomness and it is difficult to obtain its data and apply it to the battery energy storage optimization model. Therefore, the present invention adopts an efficient linearization method to calculate ΔDOD j,s , as follows

[0268] First, the increment of the depth of discharge during the period t when the battery is charged and discharged can be expressed as:

[0269] ΔDoD j,t =DoD j,t -DoD j,t-1 =SoC j,t-1 -SoC j,t

[0270] Define the discharge depth increment during the discharge process for:

[0271]

[0272] Where: Represented as a 0-1 variable, when When the battery is in the discharge state, further It can be linearized as:

[0273]

[0274] Where: M is a sufficiently large positive number.

[0275] Because the battery discharge process may be a continuous period of time, define a 0-1 variable To represent the continuous discharge period:

[0276]

[0277] Where: In a discharge cycle s, t = 1 represents the start of discharge. If the battery is still in a continuous discharge state in the next period t, then otherwise further It can be linearized as:

[0278]

[0279] Finally, the cumulative increment of the discharge depth at time period t is It can be expressed as:

[0280]

[0281] In the formula: In a continuous discharge period, is a cumulative value of the discharge depth that overlaps with the previous time interval, but what is actually needed is the incremental value of the discharge depth of the last time interval of the discharge cycle, which is also the incremental value of the discharge depth of the entire discharge cycle. Therefore, another 0-1 variable is defined To determine whether the battery discharge period is in the last period of the discharge cycle:

[0282]

[0283]

[0284] Where: If t is the last time interval of a discharge cycle, then otherwise So the cumulative discharge depth increment at the last time interval of each discharge cycle is It can be expressed as:

[0285]

[0286] The equivalent number of battery cycles per day is for:

[0287]

[0288] The power function shown in the above formula is piecewise linearized as follows:

[0289]

[0290] Where: D is the total number of piecewise linearization, g j,t,d is a 0-1 variable, indicating the increment of the discharge depth of the BESS at time t In the dth segment; K j,d and B j,d is the piecewise linear fitting coefficient; The upper and lower limits of the discharge depth increment for each segment.

[0291] Therefore, the daily depreciation cost of battery energy storage cycle life is C ESS The linearized cost model is:

[0292]

[0293] Where: and are the linear fitting coefficients of each order.

[0294] 2) Battery energy storage operation constraints

[0295] Battery energy storage has three states: charging, discharging, and static (neither charging nor discharging). 0-1 variables are introduced Indicates that the battery energy storage at node j is in the charging state during period t, a 0-1 variable Indicates that the battery energy storage at node j is in a discharging state during period t. Since the battery energy storage can only be in one of the three states during the same period, we have:

[0296]

[0297] During operation, battery energy storage is also subject to constraints on charge and discharge rates, energy balance, and charge and discharge processes, which can be expressed as:

[0298] a) Charge and discharge rate constraints

[0299]

[0300] Where: are the upper and lower limits of the charging power of the battery energy storage at node j respectively; They are the upper and lower limits of the battery energy storage discharge power at node j respectively; are the charging and discharging powers of the battery energy storage at node j in time period t, respectively.

[0301] b) Charge and discharge energy balance constraints

[0302]

[0303] Where: SoC j,t are the state of charge of the battery energy storage at node j during period t; η ch,j , η dis,j are the charging and discharging efficiencies of the battery energy storage at node j, respectively; Δt is the scheduling time interval; SoC j are the upper and lower limits of the state of charge of the battery energy storage at node j respectively.

[0304] Step 3: Establish voltage and reactive power operation constraints of the active distribution network. In addition to the three individual constraints of distributed photovoltaic controllable model, ZIP load model and battery energy storage operation constraints introduced in the second step, the operation constraints of the active distribution network also include system flow constraints, node voltage constraints, branch current constraints, on-load tap-changing transformer operation constraints, group switching capacitor operation constraints, SVC operation constraints, and distribution transformer gateway power constraints.

[0305] 1) Power flow constraints of active distribution networks

[0306] like Figure 2 The radial active distribution network shown in the figure uses the second-order cone relaxation power flow equation constraint as follows:

[0307]

[0308]

[0309] Where: set u(j) is the set of head nodes of the branch with j as the terminal node in the distribution network; set v(j) is the set of terminal nodes of the branch with j as the head node in the distribution network; U i,t is the voltage amplitude of node i in period t; I ij,t is the current amplitude of branch (i, j) in period t, l ij,t =(I ij,t ) 2 , V i,t =(U i,t ) 2 .P ij,t , Q ij,t are the active power and reactive power of branch (i, j) in period t respectively; P j,t , Q j,t are the active power and reactive power injected into node j respectively; are respectively the active power and reactive power injected into the main grid by the gateway at node j during period t; are the continuous compensation power of the SVC connected to node j and the discrete compensation power of the group switching capacitor bank CB respectively; ij is the reactance of branch (i, j), r ij is the resistance of branch (i, j); T is the number of scheduling operation hours.

[0310] 2) Node voltage and branch current constraints

[0311]

[0312] Where: are the upper and lower limits of the voltage amplitude at node j respectively; is the upper limit of the current amplitude of branch (i, j).

[0313] 3) On-load tap-changing transformer (OLTC) operation constraints

[0314] The branch model including the on-load tap-changing transformer is as follows: Figure 3 As shown, Figure 3 An on-load tap-changing transformer is added between the intermediate nodes i and j, and an auxiliary node o is added. The OLTC operation has the following constraints:

[0315]

[0316] Where: V o,t N is the square of the voltage amplitude of the auxiliary node o during the period t; ij,t is the optimal gear of the transformer of branch (i, j) in period t, which is an integer variable; Δk ij k is the increment of each level of the transformer tap of branch (i, j); ij,t is the transformation ratio of the transformer of branch (i, j) in period t; K i j is the number of gear positions of the transformer tap of branch (i, j); They are the upper and lower limits of the transformer ratio of branch (i, j) respectively.

[0317] Further, the positive integer N ij,t Convert to binary number λ ij,n,t ∈{0,1}, n=0,1,…K ij , and linearize it by combining Big-M, we can get

[0318]

[0319] Where: m ij,t 、x ij,n,t =λ ij,n,t Vj,t ,y ij,n,t =λ ij,n,t m ij,t is an intermediate auxiliary variable.

[0320] 4) Operation constraints of group switching capacitors (CB)

[0321] The group switching capacitor is a discrete reactive power compensation mechanical device, and its frequent operation needs to be limited to prevent it from affecting its service life. The specific constraints are as follows:

[0322]

[0323] Where: is the compensation power of the CB connected to the jth node in period t; is the number of CBs connected to the jth node put into operation in period t; Compensate the power of each group of CBs connected to the jth node in time period t; The number of CB actions within T cycles per day is limited;

[0324] Introducing auxiliary variables and Represents the positive and negative changes in the number of CB groups at node i during period t, and the change values ​​are all expressed as non-negative numbers, which can be linearized as follows:

[0325]

[0326] 5) SVC operation constraints

[0327] The operating constraints of the continuously regulated reactive power compensation device SVC are as follows:

[0328]

[0329] Where: are the lower limit and upper limit of the adjustable reactive power of the SVC connected to node j respectively.

[0330] 6) Power Constraints on Distribution Gateway Ports

[0331]

[0332] Where: are the upper and lower limits of active power at the node j of the distribution network respectively; are the upper and lower limits of reactive power at node j in the distribution network respectively.

[0333] Step 4: Establish the voltage and reactive power optimization objective function of the active distribution network, including network loss cost, on-load tap-changing transformer operation cost, capacitor switching cost, and battery energy storage cycle life daily depreciation cost, as shown below:

[0334]

[0335] Where: c loss 、c OLTC and c CB They are network loss cost coefficient, on-load tap-changing transformer operation cost coefficient and capacitor switching cost coefficient respectively; Ω L is the branch set; Ω CB It is the set of CB nodes.

[0336] Step 5: Use the ellipsoid uncertainty set to represent the uncertainty of distributed photovoltaic output, and establish the relationship between the adjustable decision variables and the uncertainty variables through the linear affine mechanism to establish the voltage-reactive AARO model of the active distribution network.

[0337] 1) Establish an ellipsoid set of distributed photovoltaic output

[0338] The present invention uses an ellipsoid set to describe the uncertainty of distributed photovoltaic output, as shown below:

[0339]

[0340] Where: is the distributed photovoltaic active power vector; is the expected value vector of distributed photovoltaic output, is the fluctuation vector of distributed photovoltaic output, N PV is the number of distributed photovoltaics; ∑ is The covariance matrix of is obtained from historical data; Λ is the radius of the ellipsoid, also called the uncertainty budget.

[0341] 2) Establish the voltage-reactive AARO model of active distribution network

[0342] The present invention proposes a voltage-reactive AARO model for an active distribution network. The voltage-reactive AARO model uses a linear affine mechanism to establish the relationship between adjustable decision variables and uncertainty variables, so that the adjustable decision variables are adjusted within the adjustable range according to the linear affine rule, rather than performing disordered optimization within the entire adjustable range. Therefore, the voltage-reactive AARO model is established by establishing a linear affine relationship between the adjustable decision variables (gateway power, distributed photovoltaic reactive output, SVC reactive output, CB reactive output) in the active distribution network and the uncertainty of distributed photovoltaic active output, as shown below:

[0343]

[0344] Where: y 0 is the expected operating point of the adjustable decision variable; ζ is the affine proportional coefficient.

[0345] Define x as an unadjustable decision variable and y as an adjustable decision variable. Based on the active distribution network voltage-reactive power optimization model established in the second and third steps and the above-mentioned linear affine strategy, the vector form of the active distribution network voltage-reactive power AARO model is obtained:

[0346]

[0347] Where: Represents an unadjustable decision variable, which is a decision variable before the uncertainty variable is obtained. The adjustable decision variables are decision variables that change in real time after the uncertainty variables are obtained. 1 、c 2 It is represented as the coefficient matrix corresponding to the objective function; R, A, B, D, G, C, g are the coefficient matrices of the variables under the corresponding constraints; r, h, w are constant column vectors. The first row of constraints in the formula represents the constraints related only to x, including battery energy storage operation constraints, branch power balance constraints and voltage drop constraints in the second-order cone power flow model, node voltage and branch current constraints, and OLTC operation constraints; the second row of constraints in the formula represents the second-order cone relaxation constraints in the second-order cone power flow model; the third row of constraints in the formula represents the coupling constraints related to x and y, and the node injection power balance constraints in the second-order cone power flow model; the fourth row of constraints in the formula represents the constraints related only to y, including SVC operation constraints, gateway power constraints, and battery energy storage operation constraints.

[0348] Step 6: Based on robust equivalence, the voltage-reactive AARO model of the active distribution network is transformed into a deterministic mixed integer second-order cone programming model for solution, and the optimization results are output.

[0349] The third constraint in the voltage-reactive AARO model of active distribution network contains uncertainty ΔP PV , for ΔP t PV ∈Z Ellipse , the inequality constraint is robust if and only if its robust counterpart (RC) holds, namely:

[0350]

[0351] In the formula, a i , b i , D i 、h i are the coefficient matrices A, B, D and the i-th row of the constant vector h respectively. Let L = ∑ 1 / 2 , ΔP PVVariable substitution is the standard norm δ, Ω t It can be expressed as The formula can be further equivalent to:

[0352]

[0353] That is

[0354]

[0355] By taking the uncertain variable ΔP PV The constraints are converted into corresponding mixed integer second-order cone robust equivalent constraints. Therefore, the active distribution network voltage-reactive AARO model established by the present invention is converted into a deterministic mixed integer second-order cone programming model for solution, and the commercial solver Cplex can be directly called for solution.

[0356] Figure 7 A schematic diagram of the physical structure of an electronic device provided by the present invention, such as Figure 7 As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communication interface 720 and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute the active power grid reactive voltage affine adjustable robust optimization method, the method comprising: collecting relevant parameters, the relevant parameters including the line parameters of the active power grid, the distributed photovoltaic output and the load prediction reference value; based on the relevant parameters, establishing the source-load-storage operation model in the active power grid; based on the relevant parameters, establishing the voltage-reactive operation constraints of the active power grid; based on the source-load-storage operation model and the voltage-reactive operation constraints, establishing the voltage-reactive optimization objective function of the active power grid; based on the voltage-reactive optimization objective function and the uncertainty of the distributed photovoltaic output, establishing the voltage-reactive affine adjustable robust optimization model of the active power grid; solving the voltage-reactive affine adjustable robust optimization model and outputting the optimization result.

[0357] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0358] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the active power grid reactive voltage affine adjustable robust optimization method provided by the above methods, and the method includes: collecting relevant parameters, the relevant parameters include line parameters of the active power grid, distributed photovoltaic output and load forecast reference values; based on the relevant parameters, establishing a source-load-storage operation model in the active power grid; based on the relevant parameters, establishing the voltage-reactive operation constraints of the active power grid; based on the source-load-storage operation model and the voltage-reactive operation constraints, establishing the voltage-reactive optimization objective function of the active power grid; based on the voltage-reactive optimization objective function and the uncertainty of the distributed photovoltaic output, establishing the voltage-reactive affine adjustable robust optimization model of the active power grid; solving the voltage-reactive affine adjustable robust optimization model and outputting the optimization result.

[0359] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the active power grid reactive voltage affine adjustable robust optimization method provided by the above-mentioned methods, the method comprising: collecting relevant parameters, the relevant parameters including line parameters of the active power grid, distributed photovoltaic output and load forecast reference value; based on the relevant parameters, establishing a source-load-storage operation model in the active power grid; based on the relevant parameters, establishing the voltage-reactive operation constraints of the active power grid; based on the source-load-storage operation model and the voltage-reactive operation constraints, establishing the voltage-reactive optimization objective function of the active power grid; based on the voltage-reactive optimization objective function and the uncertainty of the distributed photovoltaic output, establishing the voltage-reactive affine adjustable robust optimization model of the active power grid; solving the voltage-reactive affine adjustable robust optimization model and outputting the optimization result.

[0360] The server embodiment described above is merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0361] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0362] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0363] The specific implementation of the present invention described above does not constitute a limitation on the protection scope of the present invention. Any other corresponding changes and modifications made based on the technical concept of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A robust optimization method for active power grid reactive power and voltage affine adjustment, It is characterized in that include: Collect relevant parameters, including line parameters of the active power grid, distributed photovoltaic output and load forecast reference values; Based on the relevant parameters, a source-load-storage operation model in an active power grid is established, wherein the source-load-storage operation model specifically includes: a distributed photovoltaic controllable model, an equivalent ZIP load model based on a linear regression method, and a battery energy storage model considering cycle life loss; Based on the relevant parameters, establishing voltage and reactive power operation constraints of the active power grid; Based on the source-load-storage operation model and the voltage-reactive operation constraints, the voltage-reactive optimization objective function of the active power grid is established, wherein the voltage-reactive optimization objective function of the active distribution network is established, including network loss cost, on-load tap-changing transformer operation cost, capacitor switching cost and battery energy storage cycle life daily depreciation cost, as shown below: Where: , and They are network loss cost coefficient, on-load tap-changing transformer operation cost coefficient and capacitor switching cost coefficient respectively; For branch collection; is the set of CB nodes; Based on the voltage-reactive power optimization objective function and the uncertainty of the distributed photovoltaic output, a voltage-reactive power affine adjustable robust optimization model of the active power grid is established; Solving the voltage-reactive affine adjustable robust optimization model and outputting the optimization result; Among them, the distributed photovoltaic controllable model is established, including: The distributed photovoltaic inverter adopts the control strategy of auxiliary services to simultaneously realize the decoupled optimal control of the active output and reactive output of distributed photovoltaics, and the operating area of ​​the distributed photovoltaic inverter j is obtained as follows: Where: , are the active power and reactive power of the distributed photovoltaic connected to node j in period t respectively; is the predicted power value of the distributed photovoltaic connected to node j during period t; is the apparent power of the distributed photovoltaic inverter connected to node j; is the power factor angle of the distributed photovoltaic inverter connected to the j node; Based on the linearization method of analytic geometry, the second circular constraint in the above equation is transformed into an equal The polygon is approximated when When, this The polygon approaches the circle infinitely, and the above equation is equal The linearization expression of the polygon is: Where: , , ; An equivalent ZIP load model based on linear regression method was established, including: The ZIP load model is represented by a second-order polynomial: Where: , , are the proportional coefficients of constant power load, constant load and constant current load at node j respectively; , are respectively the active power demand and reactive power demand of the load at node j under rated voltage; is the voltage amplitude of node j in period t, , are the active power and reactive power of the load at node j respectively; Establish a battery energy storage model that takes into account cycle life degradation, including: The daily depreciation cost of battery energy storage cycle life is determined by the rated capacity and rated power of the battery energy storage, namely: Where: , They represent the rated capacity and rated power of the battery energy storage at node j respectively; A collection of installed nodes for battery energy storage; , They represent the annualized investment cost per unit capacity and per unit power of battery energy storage at node j respectively, and the calculation formula is: , Where: , They represent the unit capacity investment cost and unit power investment cost of battery energy storage at node j respectively; is the discount rate; is the cycle life of the battery energy storage at node j; The daily depreciation cost of battery energy storage cycle life is affected by the cycle life The impact of It depends on the number of cycles of battery charge and discharge, and its expression is as follows: Where: The maximum number of cycles that a battery can be discharged at the same depth of discharge until the battery life is exhausted; It is the number of cycles in which the battery is discharged at the same depth of discharge in one day, i.e. the number of daily cycles; The cycle life at different discharge depths is calculated based on the power function of the battery cycle life, expressed as: Where: is the discharge depth of the battery; It is the number of cycles that the battery is discharged at the same depth of discharge within a period of time; It is the equivalent number of cycles of the battery being discharged at 100% depth of discharge in the same time. Expressed as battery life loss; is the constant obtained by fitting; The battery life loss can be expressed as: Where: s represents the time of a battery discharge process, which includes the continuous time period t during which the battery is discharged at different depths of discharge; is the discharge depth of the battery at the end of discharge within time s; is the discharge depth of the battery at the beginning of discharge within time s; is the discharge depth increment of the battery in time s; when the battery is discharged at a depth of When discharging, is the equivalent number of cycles of the battery being discharged at 100% depth of discharge within time s; The equivalent number of cycles of a battery based on different depths of discharge in a day can be expressed as: Where: It is the equivalent number of cycles of the battery based on different discharge depths in a day, that is, the daily equivalent cycle number. The number of hours for scheduling operation; Therefore, the cycle life It can be expressed as: Where: The number of cycles a battery can be discharged at 100% depth of discharge until the battery life is exhausted.

2. The active grid reactive voltage affine adjustable robust optimization method according to claim 1, It is characterized in that The voltage and reactive power operation constraints include: The single-unit constraints of the distributed photovoltaic controllable model, the single-unit constraints of the ZIP load model and the constraints of the battery energy storage model.

3. The active grid reactive voltage affine adjustable robust optimization method according to claim 1, It is characterized in that The voltage reactive power operation constraint further includes: Power system flow constraints, node voltage constraints, branch current constraints, on-load tap-changing transformer operation constraints, group switching capacitor operation constraints, SVC operation constraints and distribution transformer gateway power constraints.

4. The active grid reactive voltage affine adjustable robust optimization method according to claim 1, It is characterized in that The determination of the uncertainty of the distributed photovoltaic output specifically includes: The uncertainty of distributed photovoltaic output is represented by an ellipsoid uncertainty set.

5. The active grid reactive voltage affine adjustable robust optimization method according to claim 4, It is characterized in that The voltage-reactive affine adjustable robust optimization model of the active power grid is established based on the voltage-reactive optimization objective function and the uncertainty of the distributed photovoltaic output, specifically including: The relationship between the adjustable decision variables and the uncertainty in the voltage-reactive power optimization objective function is established through a linear affine mechanism, and a voltage-reactive power affine adjustable robust optimization model for the active power grid is established.

6. The active grid reactive voltage affine adjustable robust optimization method according to claim 1, It is characterized in that The step of solving the voltage-reactive affine adjustable robust optimization model and outputting the optimization result specifically includes: Based on robust equivalence, the voltage-reactive affine adjustable robust optimization model is transformed into a deterministic hybrid second-order cone programming model for solution, and the optimization result is output.

7. An active power grid reactive voltage affine adjustable robust optimization device, used to execute the active power grid reactive voltage affine adjustable robust optimization method according to any one of claims 1 to 6, It is characterized in that include: A collection unit, used to collect relevant parameters, including line parameters of the active power grid, distributed photovoltaic output and load prediction reference values; A first model unit, used to establish a source-load-storage operation model in the active power grid based on the relevant parameters; A constraint unit, used for establishing voltage and reactive power operation constraints of the active power grid based on the relevant parameters; A second model unit is used to establish a voltage-reactive power optimization objective function of the active power grid based on the source-load-storage operation model and the voltage-reactive power operation constraint; A third model unit is used to establish a voltage-reactive affine adjustable robust optimization model of the active power grid based on the voltage-reactive optimization objective function and the uncertainty of the distributed photovoltaic output; The solving unit is used to solve the voltage-reactive affine adjustable robust optimization model and output the optimization result.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the program, the steps of the active grid reactive power voltage affine adjustable robust optimization method according to any one of claims 1 to 6 are implemented.