SoC interval managed battery energy storage and photovoltaic inverter droop curve collaborative optimization method and system

Abstract

The application discloses a kind of SoC section management battery energy storage and photovoltaic inverter droop curve collaborative optimization method and system.The method optimizes the droop curve of battery energy storage and photovoltaic inverter from power distribution network dispatching level, including: establishing day-ahead SoC management model, solving optimal SoC section and calculating battery energy storage maximum charge-discharge power based on this during daily scheduling period;Considering the uncertainty of daily source and load scene, a five-section droop curve model based on binary counting and large M method is constructed, the model is simplified by subsection reduction method, and the constraints are reconstructed by penalty linear approximation method, the droop curve is optimized and issued to battery energy storage and photovoltaic inverter, and the device controller adjusts the output power according to voltage measurement and droop curve, to realize efficient response to power distribution network source and load and node voltage change.The application optimizes SoC section management and droop curve in layers, reduces the power loss, voltage deviation and operating cost of power distribution network.

Description

Technical Field

[0001] This invention belongs to the field of power distribution networks, specifically relating to a method and system for co-optimizing the droop curves of battery energy storage and photovoltaic inverters under SoC range management. Background Technology

[0002] The high penetration rate of renewable distributed energy and the emergence of new loads may lead to problems such as voltage deviation and fluctuations, supply-demand mismatch, and power backflow in distribution networks. Thanks to the energy time-shifting function and rapid response capability of battery energy storage, installing battery energy storage in distribution networks and coordinating its operation with photovoltaic systems has become one of the most effective solutions to these problems. At the same time, with the development of battery technology and battery management, its operating costs have been significantly reduced, promoting the widespread application of battery energy storage in distribution networks.

[0003] However, current research on the operation and configuration of battery energy storage and photovoltaic systems mainly focuses on optimizing the power setpoint to ensure that the battery energy storage and photovoltaic inverters maintain a constant power output over a certain period of time. This method is insufficient to quickly respond to power fluctuations caused by renewable energy output and load variations, especially under conditions of grid instability or significant load changes, where the system struggles to achieve rapid adjustment and efficient response. Considering these issues and in accordance with the IEEE 1547 standard, battery energy storage and photovoltaic inverters can operate in a droop control mode. Specifically, distribution networks are typically connected to the upstream grid and have relatively stable frequencies. Therefore, droop control can be implemented separately for the active-voltage (PV) aspect of battery energy storage and the reactive-voltage (QV) aspect of the photovoltaic inverter, allowing for real-time and flexible adjustment of active and reactive power, thereby further improving the operational performance of the distribution network.

[0004] Generally, sag curves are typically modeled in two forms: with or without dead zones. Figure 2As shown. The former is represented as a three-segment droop curve, including the droop slope and upper and lower power limits, aiming to fully utilize the inverter's capacity to keep the node voltage within acceptable range. However, frequently adjusting the power output based on this dead-zone-free droop curve may affect the inverter's stability and accelerate equipment aging. A droop curve with a dead zone can be divided into five intervals, and to some extent, the modeling of a five-segment droop curve incorporates the characteristics of a three-segment curve. Specifically, if the dead zone width is zero and the upper and lower slopes are the same, it can be considered a three-segment linear droop curve. However, regardless of whether the droop curve modeling has a dead zone or not, it inevitably introduces a bilinear constraint, namely the product of the droop slope and the voltage offset. This significantly increases computational complexity and reduces solution speed. To address this issue, several simplification techniques have been adopted in droop curve modeling and optimization, including pre-determining certain droop parameters (such as the droop slope and inflection point voltage), discretizing the node voltage using binary auxiliary variables, and reducing source-load uncertainty scenarios. However, these methods may potentially compromise certain droop characteristics, limiting the adaptability of droop curve modeling. Therefore, developing a general and efficient model and providing a solution for sag curve optimization remains a key unsolved challenge in the current research field.

[0005] Furthermore, due to the limitations of battery energy storage capacity and its energy-time coupling characteristics, the available charge and discharge capabilities are affected by the System-on-Chips (SoC). Therefore, traditional droop optimization techniques for photovoltaic inverters or generators are not suitable for direct application to battery energy storage. For example, in droop curve design, improper configuration of power setpoints or limits may lead to excessively high or insufficient SoCs during subsequent charging or discharging. Therefore, the SoC of battery energy storage must be kept within an appropriate range. To this end, it is necessary to study a hierarchical scheduling optimization framework for battery energy storage and photovoltaic inverter droop curves that considers SoC range management, achieving synergy between SoC range management and droop curve optimization. Summary of the Invention

[0006] Objective: This invention proposes a method and system for co-optimizing the droop curves of battery energy storage and photovoltaic inverters under SoC (System-on-Chips) interval management. It provides a multi-stage hierarchical coordination framework for distribution network operation to coordinate SoC interval management and droop curve optimization of battery energy storage. Simultaneously, it fully models the five-segment droop curves of battery energy storage (PV) and photovoltaic inverter (QV), achieving efficient optimization of droop parameters.

[0007] Technical solution: A method for co-optimizing the droop curves of battery energy storage and photovoltaic inverters under SoC range management, comprising the following steps:

[0008] In the day-ahead phase, a battery energy storage state of charge (SoC) interval management model is established. Based on the generated day-ahead source-load uncertainty scenario of the distribution network, a stochastic optimization method is used to solve the battery energy storage SoC interval management model to obtain the optimal SoC interval for battery energy storage during the day. Based on the obtained optimal SoC interval, the maximum charging and discharging power of battery energy storage during the day-ahead scheduling period is calculated.

[0009] During the intraday phase, considering the uncertainty of the distribution network's source and load during the day, an intraday scheduling model based on the five-segment droop curves of battery energy storage active power-voltage (PV) and photovoltaic inverter reactive power-voltage (QV) is established.

[0010] The segmented reduction method is adopted to identify and discard unnecessary segments in the five-segment droop curve intraday scheduling model for each source load uncertainty scenario, thereby reducing the number of segments in the droop curve intraday scheduling model corresponding to each uncertainty scenario.

[0011] A penalized linear approximation solution method is adopted to reconstruct the bilinear constraints in the intraday scheduling model of the droop curve. The droop curves of battery energy storage (PV) and photovoltaic inverters (QV) during the intraday scheduling period are obtained iteratively and sent to the battery energy storage and photovoltaic inverter devices. The device controllers dynamically adjust the active and reactive power outputs based on the real-time measurement information of the local node voltage and the optimized droop curves, so as to achieve efficient response to changes in the power grid source load and node voltage.

[0012] Furthermore, the battery energy storage state of charge (SoC) interval management model aims to minimize network power loss, node voltage deviation, and distribution network operating costs. The constraints include battery energy storage operation constraints, battery energy storage SoC interval constraints, distribution network power flow constraints, node voltage constraints, and line transmission capacity constraints.

[0013] The objective function of the battery energy storage state of charge (SoC) interval management model is as follows:

[0014]

[0015] In the formula, N s Let S be the total number of scenarios, ω1, ω2, and ω3 be the set of scenarios s, and ω1, ω2, and ω3 be the weight coefficients for different objectives, respectively. For the power loss of the distribution network in scenario s, The average voltage offset of the distribution network in scenario s. To determine the transaction costs between the distribution network and the upstream power grid in scenario s, The calculation formulas for the battery energy storage operating cost in scenario s are as follows:

[0016]

[0017] Where T is the set of time period indices t, J is the set of distribution network nodes j, L is the set of branches jk, and r jk Let P be the resistance of branch jk. jk,t,s and Q jk,t,s Let N be the active power and reactive power flowing through branch jk in scenario s during time period t, respectively, and let V0 be the reference voltage of the root node. j Let V be the total number of distribution network nodes j. j,t,s This represents the voltage at node j in scenario s during time period t. and These represent the unit purchase price and selling price of electricity in the distribution network during time period t, respectively. and These represent the power purchased and power sold by the distribution network in scenario s during time period t, respectively, where τ is the charging and discharging time. The unit operating cost of battery energy storage, This represents the battery energy storage discharge power or charging power in scenario s during time period t at node j. and Let represent the battery energy storage charging efficiency and discharging efficiency at node j, respectively.

[0018] The operating constraints of battery energy storage are as follows:

[0019]

[0020]

[0021] The constraints for the battery energy storage SoC range are as follows:

[0022]

[0023] in, The rated charge and discharge power of the battery energy storage at node j, For the remaining battery energy stored in scenario s during time period t at node j, The initial electrical energy stored in the battery at node j. The remaining energy of the battery storage in scenario s at node j after 24 hours of charge and discharge. and These represent the lower and upper bounds of the battery energy storage SoC range at node j and time period t, respectively. For the rated battery energy storage capacity at node j, SOC min Minimum SoC value preset for battery energy storage, SOC max The maximum SoC value preset for battery energy storage. β t and These represent the minimum and maximum widths of the battery energy storage SoC range during time period t, respectively.

[0024] Furthermore, the step of using a stochastic optimization method to solve the battery storage SoC interval management model to obtain the optimal SoC interval for battery storage during the day, and calculating the maximum charging and discharging power of battery storage during the day's scheduling period based on the obtained optimal SoC interval, includes:

[0025] By introducing slack variables and polygon approximation to linearize the model, the battery energy storage state of charge (SoC) interval management model is transformed into a mixed-integer linear model. Then, a stochastic optimization method is used to solve the mixed-integer linear model to obtain the optimal SoC upper limit for intraday battery energy storage operation. and lower limit And the remaining electrical energy stored in the battery during the time period t-1 at node j.

[0026] Based on the obtained and The value is used to calculate the maximum charging and discharging power of the battery energy storage during the daily scheduling period, as shown below:

[0027]

[0028] in, This represents the maximum charging power of battery energy storage during node j at time t within the day. This represents the maximum discharge power of the battery energy storage during the intraday phase at node j and time period t.

[0029] Furthermore, an intraday scheduling model for the five-segment droop curve of battery energy storage PV and photovoltaic inverter QV based on binary counting and the Big M method is established, including:

[0030] Establish droop curve models for active charging and discharging of battery energy storage and reactive power of photovoltaic inverters, in order to... Indicates the active power of battery energy storage discharge. Negative value of active power for battery energy storage charging Reactive power of photovoltaic inverters For any one of the following, the intraday scheduling model with five downward-sloping curves is shown below:

[0031]

[0032] in, Let be the voltage value at the inflection point of the droop curve of the corresponding device during time period t at node j. For the upper / lower bound of the power of the device droop curve model at node j and time period t, when express hour, and when express hour, and when express hour, and This represents the minimum reactive power / maximum reactive power of the photovoltaic inverter. The dead zone power of the corresponding device droop curve at node j during time period t. and These are the adjustment powers for the second and fourth segments of the droop curve corresponding to the equipment. The voltage of node j in time period t and the voltage at the inflection point of the droop curve. The difference, and These are the slopes of the second and fourth segments of the drooping curve at node j and time period t, respectively. and These represent the minimum and maximum slopes of the droop curve for the corresponding device during time period t at node j.

[0033] The five-segment droop curve model is reconstructed based on binary counting and the Big M method. First, a vector consisting of three binary decision variables is defined. Five possible binary combinations in this vector are selected to represent the five segments of the droop curve. The binary variable constraints of the droop curve are as follows:

[0034]

[0035] in, Let be the binary decision variables for the drooping curve. A vector consisting of three binary decision variables, obtained by... Different binary combinations are used to represent the five segments of the drooping curve;

[0036] Then, define five coefficient matrices as follows:

[0037] γ1=[1,1,1], γ2=[-1,1,1], γ3=[1,-1,1], γ4=[-1,-1,1], γ5=[1,1,-1]

[0038] Finally, the power and voltage corresponding to each segment of the droop curve are represented by multiple inequality constraints. The droop curve reconstruction constraints are shown below:

[0039]

[0040] Where M is a specified constant, and the superscript T denotes the matrix transpose symbol;

[0041] Ultimately, a five-segment droop curve intraday scheduling model A for battery energy storage (PV) and photovoltaic inverter (QV) based on binary counting and the Big M method was formed. The objective of model A is to minimize the power loss of the distribution network, the average voltage deviation of the distribution network, the transaction cost between the distribution network and the upper-level grid, and the operating cost of battery energy storage. The main constraints include battery energy storage operation constraints, distribution network power flow constraints, node voltage constraints, line transmission capacity constraints, binary variable constraints of the droop curve, and droop curve reconfiguration constraints.

[0042] Furthermore, a segmented reduction method is employed to identify and discard unnecessary segments in the five-segment droop curve intraday scheduling model for each source load uncertainty scenario, including:

[0043] A model B for optimal scheduling of active power of battery storage and reactive power of photovoltaic inverter under droop-free control is established. Its objective is to minimize the power loss of distribution network, the average voltage deviation of distribution network, the transaction cost between distribution network and the upper-level grid, and the operating cost of battery storage. The constraints include battery storage operation constraints, distribution network power flow constraints, node voltage constraints, and line transmission capacity constraints. All uncertain scenarios correspond to only one set of optimal decision variables. Solving model B yields a set of optimal active power of battery storage and reactive power output of photovoltaic inverter, which is used as the dead zone power of the droop curve. Based on the dead zone power and the upper and lower limits of equipment power as boundaries, the droop curve is divided into three regions: above the dead zone power, dead zone power, and below the dead zone power.

[0044] An optimal scheduling model C for the active power of battery storage and the reactive power of photovoltaic inverters without droop control is established. The objective is to minimize distribution network power loss, average voltage deviation, transaction costs between the distribution network and the upstream grid, and operating costs of battery storage. Constraints include battery storage operation constraints, distribution network power flow constraints, node voltage constraints, and line transmission capacity constraints. Each set of uncertainties has a corresponding set of optimal decision variables. Based on the power output values ​​of the corresponding devices obtained from model C, their positions within the defined regions are determined, and unnecessary segments of the droop curve are identified. The judgment criteria are as follows:

[0045] When the power value is above the dead zone power, the fourth and fifth segments of the droop curve are unnecessary segments; when the power value is equal to the dead zone power, the first and fifth segments of the droop curve are unnecessary segments; when the power value is below the dead zone power, the first and second segments of the droop curve are unnecessary segments.

[0046] Furthermore, a penalized linear approximation solution method is adopted to reconstruct the bilinear constraints in the intraday scheduling model of the droop curve, and iterative solutions are used to obtain the droop curves of battery storage PV and photovoltaic inverter QV during the intraday scheduling period, including:

[0047] (a) Ignoring variables The superscripts and subscripts in the text will be used to represent the bilinear constraints in model A. In short, ΔG = η·ΔV, and the McCormick envelope is used to relax this bilinear constraint as shown below:

[0048] ΔG≥η min ·ΔV+η·ΔV min -η min ·ΔV min

[0049] ΔG≥η max ·ΔV+η·ΔV max -η max ·ΔV max

[0050] ΔG≤η max ·ΔV+η·ΔV min -η max ·ΔV min

[0051] ΔG≤η min ·ΔV+η·ΔV max -η min ·ΔV max

[0052] Where, η min and η max ΔV represents the minimum and maximum values ​​of the slope of the sag curve, respectively. min and ΔV max Let represent the minimum and maximum values ​​of the voltage offset, respectively. After relaxation, model D is obtained, and the slope of the droop curve and the initial values ​​of the voltage offset are solved to obtain the initial values. At this point, n = 0;

[0053] (b) Based on the first-order Taylor expansion formula at point A linear approximation of the bilinear terms in the vicinity is expressed as:

[0054]

[0055] in, It is a linearized function;

[0056] (c) Introduce an auxiliary variable μ to represent With ΔG n The absolute difference between them is expressed as:

[0057]

[0058] (d) Add the auxiliary variable as a penalty term to the objective function to establish a penalized linear approximation model E, whose objective function is expressed as:

[0059]

[0060] Its constraints include battery energy storage operation constraints, distribution network power flow constraints, node voltage constraints, line transmission capacity constraints, droop curve binary variable constraints, droop curve reconstruction constraints, relaxation constraints based on McCormick envelopes, linear approximation constraints based on first-order Taylor expansion in step (b), and related inequality constraints of auxiliary variable μ in step (c), where π n It is a penalty factor that gradually increases with each iteration;

[0061] (e) Solve for model E to obtain the optimal droop slope result. Replace the constraints in step (c) of model E with:

[0062]

[0063] A model F with a fixed slope is obtained, and the optimization result of its objective equation is set as obj2. Solving this model yields the optimal voltage offset result. Check if the termination criterion is met; if it is, stop the iteration and output the result.

[0064] The termination criteria are as follows:

[0065]

[0066] Where δ is a preset threshold;

[0067] The sag slope boundary for the next iteration is updated based on the current optimal sag slope value, expressed as:

[0068]

[0069] in, and These are the maximum and minimum values ​​of the droop slope in the (n+1)th iteration, respectively. Let ψ be the droop slope obtained in the nth iteration. n It is a parameter greater than 0, used to gradually tighten the boundary of the sag slope;

[0070] Update penalty factor π n The method is as follows:

[0071]

[0072] in, π is a coefficient greater than 1. max The maximum value of the pre-set penalty factor.

[0073] Furthermore, the solution process of the penalized linear approximation method includes:

[0074] (a) Initialize δ, 1>ψ>0, n=0π max ≥π n >0,

[0075] (b) Establish models A, B, and C, and perform piecewise reduction to identify unnecessary line segments of the drooping curve and reduce model complexity;

[0076] (c) Based on the McCormick envelope, relax the bilinear constraints in model A to obtain model D, and solve for the initial value of the slope of the sag curve.

[0077] (d) Based on Establish a penalized linear approximation model E, and solve it to obtain...

[0078] (e) Based on droop slope Establish a slope-fixed model F and solve for it.

[0079] (f) Determine if the termination criterion is met. If it is met, stop the iteration and output the result. Otherwise, update the droop slope boundary and penalty factor, let n = n + 1, and go to step (d).

[0080] A SoC-range managed battery energy storage and photovoltaic inverter droop curve co-optimization system includes:

[0081] The SoC interval management module is used to establish a battery energy storage state of charge (SoC) interval management model during the day-ahead phase. Based on the generated day-ahead source-load uncertainty scenario of the distribution network, the module uses a stochastic optimization method to solve the battery energy storage SoC interval management model to obtain the optimal SoC interval for battery energy storage during the day. Based on the obtained optimal SoC interval, the module calculates the maximum charging and discharging power of battery energy storage during the day-ahead scheduling period.

[0082] The droop curve model construction module is used to establish an intraday scheduling model of five droop curves for battery energy storage active power-voltage (PV) and photovoltaic inverter reactive power-voltage (QV) based on binary counting and the big M method, considering the uncertainty of the source and load of the distribution network during the intraday stage.

[0083] The model simplification module is used to identify and discard unnecessary segments in the five-segment droop curve intraday scheduling model for each source load uncertainty scenario using a segmented reduction method, thereby reducing the number of segments in the droop curve intraday scheduling model corresponding to each uncertainty scenario.

[0084] The droop curve optimization solution and distribution module is used to reconstruct the bilinear constraints in the intraday scheduling model of the droop curve using a penalized linear approximation method. Iteratively, it obtains the droop curves of the battery storage PV and the photovoltaic inverter QV during the intraday scheduling period, and distributes them to the battery storage and photovoltaic inverter devices. The device controllers then dynamically adjust the active and reactive power outputs based on the real-time measurement information of the local node voltage and the optimized droop curves, achieving efficient response to changes in the power distribution network source load and node voltage.

[0085] The present invention also provides a computer device, comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, wherein when the programs are executed by the processors, they implement the steps of the SoC range-managed battery energy storage and photovoltaic inverter droop curve co-optimization method as described above.

[0086] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the SoC range management method for co-optimizing battery energy storage and photovoltaic inverter droop curves as described above are implemented.

[0087] Beneficial effects:

[0088] (1) This invention coordinates the SoC range management of battery energy storage with the PV droop control of battery energy storage and the QV droop control of photovoltaic inverters in a multi-time scale and multi-stage manner, optimizes the SoC range and intraday droop curve parameters throughout the day, and minimizes the power loss, voltage deviation and operating cost of the distribution network.

[0089] (2) This invention proposes a five-segment droop curve modeling method and a corresponding piecewise simplification technique to simplify the droop curve model. At the same time, it adopts a penalized linear approximation solution method for linearization processing, which has a fast solution speed, meets the intraday scheduling requirements, and efficiently obtains droop curve parameters. Attached Figure Description

[0090] Figure 1 This is a diagram illustrating the overall framework of the battery energy storage and photovoltaic inverter droop curve co-optimization method considering SoC range management in this invention.

[0091] Figure 2 This is a schematic diagram of the three-segment drooping curve and the five-segment drooping curve in this invention;

[0092] Figure 3 This is the current SoC range management optimization result in this invention;

[0093] Figure 4 This is a droop control curve diagram of the photovoltaic inverter in this invention;

[0094] Figure 5 This is a graph showing the battery energy storage droop control in this invention. Detailed Implementation

[0095] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0096] This embodiment discloses an adaptive optimization method for the droop curves of battery energy storage and photovoltaic inverters considering SoC range management in distribution network scheduling. The specific framework is as follows: Figure 1 As shown, it includes the following steps:

[0097] Step (1): In the day-ahead phase, establish a battery storage load SoC interval management model. Based on the generated day-ahead source-load uncertainty scenario of the distribution network, use a stochastic optimization method to obtain the optimal SoC interval for battery storage during the day. Based on this SoC interval, calculate the maximum charging and discharging power of battery storage during the day-ahead scheduling period. Specifically, this includes:

[0098] (a) The following constraints are established for the operation of battery energy storage:

[0099]

[0100] in, This represents the battery energy storage discharge power or charging power in scenario s during time period t at node j. The rated charge and discharge power of the battery energy storage at node j, For the remaining battery energy stored in scenario s during time period t at node j, and These represent the battery energy storage charging efficiency and discharging efficiency at node j, respectively, where τ is the charging / discharging time. For the rated battery energy storage capacity at node j, The initial electrical energy stored in the battery at node j. The remaining energy of the battery storage in scenario s at node j after 24 hours of charging and discharging. and These represent the lower and upper bounds of the battery energy storage SoC range at node j and time period t, respectively. T is the set of time period indices t, and J is the set of distribution network nodes j.

[0101] (b) The following constraints are established for the battery energy storage SoC range:

[0102]

[0103] Among them, SOC min Minimum SoC value preset for battery energy storage, SOC max The maximum SoC value preset for battery energy storage. β t and These represent the minimum and maximum widths of the battery energy storage SoC range during time period t, respectively.

[0104] (c) To minimize network power loss, node voltage deviation, and distribution network operating costs (including transaction costs between the distribution network and the upstream grid and battery storage operating costs), a day-ahead battery storage SoC range management model based on stochastic optimization is established:

[0105]

[0106] st(1)-(6)

[0107]

[0108] Where, N s Let S be the total number of scenes, S be the set of scenes s, L be the set of branches jk, and ω1, ω2, and ω3 be the weight coefficients for different objectives. For the power loss of the distribution network in scenario s, The average voltage offset of the distribution network in scenario s. To determine the transaction costs between the distribution network and the upstream power grid in scenario s, For the battery energy storage operating cost in scenario s, r represents the unit operating cost of battery energy storage jk and x jk The resistance and reactance of branch jk are respectively, P jk,t,s and Q jk,t,s P represents the active and reactive power flowing through branch jk in scenario s during time period t. ij,t,s and Q ij,t,s Let Vij be the active power and reactive power flowing through branch ij in scenario s during time period t, respectively, and let V0 be the reference voltage of the root node. j Let j be the total number of nodes in the distribution network. and These represent the unit purchase price and selling price of electricity in the distribution network during time period t, respectively. and These represent the power purchased and power sold by the distribution network in scenario s during time period t, respectively. This represents the photovoltaic power generation in scenario s at node j during time period t. This represents the wind turbine power generation capacity in scenario s during time period t at node j. This represents the active power of the load in scenario s during time period t at node j. This represents the load reactive power in scenario s during time period t at node j. V represents the reactive power of the photovoltaic inverter in scenario s at node j during time period t. j,t,s V represents the voltage at node j in scenario s during time period t. min and Vmax These represent the minimum and maximum allowable node voltages, respectively. This indicates the maximum apparent power allowed to flow through branch jk. and These represent the minimum and maximum reactive power of the photovoltaic inverter, P. 01,t,s This refers to the active power flowing on the branch connecting the distribution network and the upstream power grid during time period t in scenario s.

[0109] (d) Introducing slack variable ζ j,t,s , such that the absolute value |V exists in the above constraint (9) j,t,s -V0| is represented as ζ j,t,s ≥V j,t,s -V0 and ζ j,t,s ≥V0-V j,t,s And the constraint (16) is linearized by the polygon approximation method, thereby converting the daytime battery storage SoC interval management model into a mixed integer linear model;

[0110] (e) The mixed-integer linear model is solved using a stochastic optimization method to obtain the following results: Figure 3 The upper and lower limits of the optimal SoC range for battery energy storage during the day are shown. And the remaining electrical energy stored in the battery during the time period t-1 at node j. And based on the obtained and The value is used to calculate the maximum charging and discharging power of the battery energy storage during the daily scheduling period, as shown below:

[0111]

[0112] in, This represents the maximum charging power of battery energy storage during node j at time t within the day. This represents the maximum discharge power of the battery energy storage during the intraday phase at node j and time period t.

[0113] Step (2), during the intraday phase, considering the intraday source-load uncertainty scenario of the distribution network, establish an intraday scheduling model for battery energy storage PV and photovoltaic inverter QV based on binary counting and the Big M method, specifically including:

[0114] (a) Establish a droop curve model for the active charging and discharging of battery energy storage and the reactive power of photovoltaic inverter, using the active power of battery energy storage discharge as the basis. For example (the negative value of the active power of battery energy storage charging) PV droop curve, reactive power of photovoltaic inverter The modeling method for the QV sag curve is the same as this, as shown below:

[0115]

[0116] in, Let be the voltage value at the inflection point of the battery energy storage discharge droop curve during time period t at node j. The maximum discharge power of the battery energy storage during time period t at node j. The minimum discharge power of the battery energy storage during time period t at node j. Let represent the dead zone power of the battery energy storage discharge droop curve at node j and time period t. and These represent the adjustment power for the second and fourth segments of the battery energy storage discharge droop curve, respectively. The voltage of node j in time period t and the inflection point voltage of the battery energy storage discharge droop curve. The difference, and These represent the slopes of the second and fourth segments of the battery energy storage discharge droop curve at node j and time period t, respectively. and These represent the minimum and maximum slopes of the battery energy storage discharge droop curve at node j and time period t, respectively.

[0117] (b) Define a vector consisting of three binary decision variables, and select five possible binary combinations from this vector to represent the five segments of the drooping curve, as shown below:

[0118]

[0119] in, Let be the binary decision variables for the discharge droop curve of the battery system. A vector consisting of three binary decision variables, obtained by... Different binary combinations are used to represent the five segments of the drooping curve. For example, when the value is [0,0,0], it represents the first segment of the drooping curve; when the value is [1,0,0], it represents the second segment; when the value is [0,1,0], it represents the third segment; when the value is [1,1,0], it represents the fourth segment; and when the value is [0,0,1], it represents the fifth segment.

[0120] (c) Define five coefficient matrices as follows:

[0121] γ1=[1,1,1], γ2=[-1,1,1], γ3=[1,-1,1], (30)

[0122] γ4=[-1,-1,1], γ5=[1,1,-1]

[0123] (d) Based on the Big M method, (21) is relaxed, and the power and voltage corresponding to each segment of the droop curve are represented by multiple inequality constraints. The droop curve reconstruction constraints are shown below:

[0124]

[0125] Here, γ is the coefficient matrix, M is a constant that is much larger than the values ​​of rated power and node voltage, and the superscript T indicates the matrix transpose symbol.

[0126] (e) Considering the uncertainty of the source and load in the distribution network during the day, an intraday scheduling model A based on binary counting and the Big M method for battery energy storage PV and photovoltaic inverter QV with five droop curves is established, as shown below:

[0127]

[0128] st(1)-(4),(8)-(18),(22)-(29),(31),(32)

[0129] Step (3) employs a segmented reduction method to identify and discard unnecessary segments in the five-segment droop curve intraday scheduling model A for each source load uncertainty scenario. This reduces the number of segments in the droop curve model corresponding to each uncertainty scenario, thereby lowering the complexity of the droop model. Specifically, this includes:

[0130] (a) Establish an optimal scheduling model B for the active power of battery storage and the reactive power of photovoltaic inverters without droop control, where all uncertain scenarios correspond to only one set of optimal decision variables. This model can be expressed as:

[0131]

[0132] st(1)-(4),(8)-(18)

[0133] Among them, the variables in the constraints C BS Without subscript 's', model B is solved to obtain a set of optimal battery energy storage active power and photovoltaic inverter reactive power outputs, which are used as the dead zone power of the droop curve. Based on the dead zone power and the upper and lower limits of the equipment power as boundaries, the droop curve is divided into three regions: above the dead zone power, dead zone power, and below the dead zone power.

[0134] (b) Establish an optimal scheduling model C for the active power of battery storage and the reactive power of photovoltaic inverters without droop control, where each set of uncertainty scenarios has a corresponding set of optimal decision variables. This model can be expressed as:

[0135]

[0136] st(1)-(4),(8)-(18)

[0137] Among them, variables in constraints Each has a subscript s. Solving model C yields multiple sets of active power output from battery storage and reactive power output from photovoltaic inverters.

[0138] (c) Determine the position of the power output value obtained from model C within the region defined in step (a) above, and identify potential unnecessary segments of the drooping curve. The judgment criteria are as follows:

[0139] When the power value is above the dead zone power, the fourth and fifth segments of the drooping curve are unnecessary segments; when the power value is equal to the dead zone power, the first and fifth segments of the drooping curve are unnecessary segments; when the power value is below the dead zone power, the first and second segments of the drooping curve are unnecessary segments. Therefore, after the segmented reduction method, constraints (31) and (32) can be expressed as inequality constraints corresponding to the three necessary segments of the drooping curve and represented by two binary variables. (28)-(29) are replaced by the following constraints:

[0140]

[0141] Step (4) employs a penalized linear approximation solution method to reconstruct the bilinear constraints in the intraday scheduling model of the droop curve. Iterative solutions are used to obtain the droop curves of the battery storage PV and the photovoltaic inverter QV during the intraday scheduling period. These curves are then distributed to the battery storage and photovoltaic inverter devices respectively, enabling adaptive adjustment of the device droop curves under changes in the power grid's source and load. Specifically, this includes:

[0142] (a) Let's ignore the variables for now. Using superscripts and subscripts, the bilinear constraint in formulas (22) and (23) can be simplified as ΔP=η·ΔV. The bilinear constraint is relaxed based on the McCormick envelope as follows:

[0143] ΔP≥η min ·ΔV+η·ΔV min -η min ·ΔV min (38a)

[0144] ΔP≥η max ·ΔV+η·ΔV max -η max ·ΔV max (38b)

[0145] ΔP≤η max ·ΔV+η·ΔV min -η max ·ΔVmin (38c)

[0146] ΔP≤η min ·ΔV+η·ΔV max -η min ·ΔV max (38d)

[0147] Where, η min and η max ΔV represents the minimum and maximum values ​​of the slope of the sag curve, respectively. min and ΔV max These represent the minimum and maximum values ​​of the voltage offset, respectively. After relaxation, model D is obtained and solved to obtain the initial values ​​of the droop curve slope and the voltage offset. At this point, n = 0.

[0148] (b) Based on the first-order Taylor expansion formula at point A linear approximation of the bilinear terms in the vicinity is expressed as:

[0149]

[0150] in, It is a linearized function.

[0151] (c) Introduce an auxiliary variable μ to represent With ΔP n The absolute difference between them is expressed as:

[0152]

[0153] (d) Add the auxiliary variable as a penalty term to the objective function to establish a penalized linear approximation model E, expressed as:

[0154]

[0155] st(1)-(4),(8)-(18),(24)-(27),(31),(32),(36)-(40)

[0156] Where, π n It is a penalty factor that gradually increases with each iteration.

[0157] (e) Solve model (41) to obtain the optimal droop slope result. Establish a slope-fixed model F, expressed as:

[0158]

[0159] st(1)-(4),(8)-(18),(24)-(27),(31),(32),(36)-(38)

[0160]

[0161] Solving this model yields the optimal voltage offset result. Check the termination criterion (44). If it is satisfied, stop the iteration and output the result; otherwise, update the droop slope boundary and the penalty factor π. n Then return to step (d) to rebuild the penalized linear approximation model E and continue iteratively solving.

[0162] The termination criteria are as follows:

[0163]

[0164] Where δ is a preset threshold.

[0165] The droop slope boundary for the next iteration is updated based on the current optimal droop slope, expressed as:

[0166]

[0167] in, and These are the maximum and minimum values ​​of the droop slope in the (n+1)th iteration, respectively. Let ψ be the droop slope obtained in the nth iteration. n It is a parameter greater than 0, used to gradually tighten the boundary of the sag slope.

[0168] Update penalty factor π n The method is as follows:

[0169]

[0170] in, π is a coefficient greater than 1. max The maximum value of the pre-set penalty factor.

[0171] According to an embodiment of the present invention, the solution process of the penalized linear approximation method is as follows:

[0172] (a) Initialize δ, 1>ψ>0, n=0π max ≥π n >0,

[0173] (b) Establish models A, B, and C, and perform piecewise reduction to identify unnecessary line segments of the drooping curve and reduce model complexity;

[0174] (c) Based on the McCormick envelope, relax the bilinear constraints in model A to obtain model D, and solve for the initial value of the slope of the sag curve.

[0175] (d) Based on Establish a penalized linear approximation model E, and solve it to obtain...

[0176] (e) Based on droop slope Establish a slope-fixed model F and solve for it.

[0177] (f) Determine whether the termination criterion (44) is satisfied. If it is satisfied, stop the iteration and output the result. Otherwise, update the droop slope boundary and penalty factor, let n = n + 1, and go to step (d).

[0178] Step (5) involves the battery storage and photovoltaic inverter's internal controller dynamically adjusting the active and reactive power outputs of the equipment based on local real-time voltage measurement information and an optimized droop curve. The specific adjustment method is as follows:

[0179] Dead zone power based on the active charge / discharge curves of battery storage and the reactive power droop curves of photovoltaic inverters obtained in step (3). and the slope parameter of the droop curve obtained in step (4) With inflection point voltage Update the PV droop curve and QV droop curve of the local layer device for the current time period, as follows: Figure 4 , Figure 5 As shown, the internal controllers of each battery energy storage and photovoltaic inverter dynamically adjust the active and reactive power outputs of the equipment according to the droop mode based on the actual local voltage deviation.

[0180] To verify the effectiveness of the proposed SoC-based interval management method for co-optimizing the droop curves of battery storage and photovoltaic inverters, a 33-node system was used to test the method. 500 sets of photovoltaic and wind power output and load scenarios in the distribution network were randomly generated using Monte Carlo sampling to simulate real-time uncertainties. Furthermore, two other scheduling and control methods were compared, as follows:

[0181] Method 1: Hierarchical coordination of BESS-PV scheduling and droop control, without considering SoC range management, with the upper and lower limits of SoC fixed at 0.9 and 0.1 respectively.

[0182] Method 2: Under the day-ahead SoC range management, the active and reactive power setpoints of the BESS and PV inverters are optimized, but local droop control is not implemented.

[0183] The comparison results of these three methods during the 10:00-11:00 and 18:00-19:00 periods are given in Table 1 and Table 2, respectively.

[0184] Table 1. Comparison results of different methods during the 10:00-11:00 time period.

[0185]

[0186]

[0187] Table 2 Comparison results of different methods during the 18:00-19:00 time period

[0188]

[0189] During the 10:00-11:00 period, Method 1 exhibited the highest average power loss, followed by Method 2. However, Method 2 showed a greater average voltage deviation than Method 1. This may be due to the limited reactive power capacity and lack of droop control during periods of high photovoltaic power generation. The method proposed in this invention achieves the lowest average power loss and average voltage deviation.

[0190] Furthermore, during the period from 18:00 to 19:00, Method 1 exhibited the highest average power loss and average voltage deviation, while Method 2's performance was at an intermediate level. Finally, the method proposed in this invention achieved the lowest average voltage deviation and average power loss.

[0191] Over the entire day, the total operating cost of Method 1 was $1697.17, Method 2 was $1674.33, and the method proposed in this invention was $1648.55. The results demonstrate that the method proposed in this invention has a significant advantage in terms of economic efficiency.

[0192] Based on the same technical concept as the method embodiments, the present invention also provides a SoC-range managed battery energy storage and photovoltaic inverter droop curve co-optimization system, comprising:

[0193] The SoC interval management module is used to establish a battery energy storage state of charge (SoC) interval management model during the day-ahead phase. Based on the generated day-ahead source-load uncertainty scenario of the distribution network, the module uses a stochastic optimization method to solve the battery energy storage SoC interval management model to obtain the optimal SoC interval for battery energy storage during the day. Based on the obtained optimal SoC interval, the module calculates the maximum charging and discharging power of battery energy storage during the day-ahead scheduling period.

[0194] The droop curve model construction module is used to establish an intraday scheduling model of five droop curves for battery energy storage active power-voltage (PV) and photovoltaic inverter reactive power-voltage (QV) based on binary counting and the big M method, considering the uncertainty of the source and load of the distribution network during the intraday stage.

[0195] The model simplification module is used to identify and discard unnecessary segments in the five-segment droop curve intraday scheduling model for each source load uncertainty scenario using a segmented reduction method, thereby reducing the number of segments in the droop curve intraday scheduling model corresponding to each uncertainty scenario.

[0196] The droop curve optimization solution and distribution module is used to reconstruct the bilinear constraints in the intraday scheduling model of the droop curve using a penalized linear approximation method. Iteratively, it obtains the droop curves of the battery storage PV and the photovoltaic inverter QV during the intraday scheduling period, and distributes them to the battery storage and photovoltaic inverter devices. The device controllers then dynamically adjust the active and reactive power outputs based on the real-time measurement information of the local node voltage and the optimized droop curves, achieving efficient response to changes in the power distribution network source load and node voltage.

[0197] It should be understood that the SoC interval management battery energy storage and photovoltaic inverter droop curve co-optimization system in the embodiments of the present invention can realize all the technical solutions in the above method embodiments. The functions of each functional module can be specifically implemented according to the methods in the above method embodiments. The specific implementation process can be referred to the relevant descriptions in the above embodiments, which will not be repeated here.

[0198] The present invention also provides a computer device, comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, wherein when the programs are executed by the processors, they implement the steps of the SoC range-managed battery energy storage and photovoltaic inverter droop curve co-optimization method as described above.

[0199] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the SoC range management method for co-optimizing battery energy storage and photovoltaic inverter droop curves as described above are implemented.

[0200] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus (systems), computer devices, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0201] This invention is described with reference to a flowchart of a method according to embodiments of the invention. It should be understood that each step in the flowchart and combinations thereof can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the process. Figure 1 A device for a function specified in one or more processes.

[0202] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 The function specified in one or more processes.

[0203] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 Steps of a specified function in one or more processes.

Claims

1. A method for co-optimizing the droop curves of battery energy storage and photovoltaic inverters under SoC range management, characterized in that, Includes the following steps: In the day-ahead phase, a battery energy storage state of charge (SoC) interval management model is established. Based on the generated day-ahead source-load uncertainty scenario of the distribution network, a stochastic optimization method is used to solve the battery energy storage SoC interval management model to obtain the optimal SoC interval for battery energy storage during the day. Based on the obtained optimal SoC interval, the maximum charging and discharging power of battery energy storage during the day-ahead scheduling period is calculated. During the intraday phase, considering the uncertainty of the distribution network's source and load during the day, an intraday scheduling model based on the five-segment droop curves of battery energy storage active power-voltage (PV) and photovoltaic inverter reactive power-voltage (QV) is established. The segmented reduction method is adopted to identify and discard unnecessary segments in the five-segment droop curve intraday scheduling model for each source load uncertainty scenario, thereby reducing the number of segments in the droop curve intraday scheduling model corresponding to each uncertainty scenario. A penalized linear approximation solution method is adopted to reconstruct the bilinear constraints in the intraday scheduling model of the droop curve. The droop curves of battery energy storage (PV) and photovoltaic inverters (QV) during the intraday scheduling period are obtained iteratively and sent to the battery energy storage and photovoltaic inverter devices. The device controllers dynamically adjust the active and reactive power outputs based on the real-time measurement information of the local node voltage and the optimized droop curves, so as to achieve efficient response to changes in the power grid source load and node voltage.

2. The method according to claim 1, characterized in that, The battery energy storage state of charge (SoC) interval management model aims to minimize network power loss, node voltage deviation, and distribution network operating costs. The constraints include battery energy storage operation constraints, battery energy storage SoC interval constraints, distribution network power flow constraints, node voltage constraints, and line transmission capacity constraints. The objective function of the battery energy storage state of charge (SoC) interval management model is as follows: In the formula, N s Let S be the total number of scenarios, ω1, ω2, and ω3 be the set of scenarios s, and ω1, ω2, and ω3 be the weight coefficients for different objectives, respectively. For the power loss of the distribution network in scenario s, The average voltage offset of the distribution network in scenario s. To determine the transaction costs between the distribution network and the upstream power grid in scenario s, The calculation formulas for the battery energy storage operating cost in scenario s are as follows: Where T is the set of time period indices t, J is the set of distribution network nodes j, L is the set of branches jk, and r jk Let P be the resistance of branch jk. jk,t,s and Q jk,t,s Let N be the active power and reactive power flowing through branch jk in scenario s during time period t, respectively, and let V0 be the reference voltage of the root node. j Let V be the total number of distribution network nodes j. j,t,s This represents the voltage at node j in scenario s during time period t. and These represent the unit purchase price and selling price of electricity in the distribution network during time period t, respectively. and These represent the power purchased and power sold by the distribution network in scenario s during time period t, respectively, where τ is the charging and discharging time. The unit operating cost of battery energy storage, This represents the battery energy storage discharge power or charging power in scenario s during time period t at node j. and Let represent the battery energy storage charging efficiency and discharging efficiency at node j, respectively. The operating constraints of battery energy storage are as follows: The constraints for the battery energy storage SoC range are as follows: in, The rated charge and discharge power of the battery energy storage at node j, For the remaining battery energy stored in scenario s during time period t at node j, The initial electrical energy stored in the battery at node j. The remaining energy of the battery storage in scenario s at node j after 24 hours of charge and discharge. and These represent the lower and upper bounds of the battery energy storage SoC range at node j and time period t, respectively. For the rated battery energy storage capacity at node j, SOC min Minimum SoC value preset for battery energy storage, SOC max The maximum SoC value preset for battery energy storage. β t and These represent the minimum and maximum widths of the battery energy storage SoC range during time period t, respectively.

3. The method according to claim 2, characterized in that, The process involves using a stochastic optimization method to solve the battery storage SoC interval management model to obtain the optimal SoC interval for battery storage within a day, and then calculating the maximum charging and discharging power of battery storage during the day's scheduling period based on the obtained optimal SoC interval. This includes: By introducing slack variables and polygon approximation to linearize the model, the battery energy storage state of charge (SoC) interval management model is transformed into a mixed-integer linear model. Then, a stochastic optimization method is used to solve the mixed-integer linear model to obtain the optimal SoC upper limit for intraday battery energy storage operation. and lower limit And the remaining electrical energy stored in the battery during the time period t-1 at node j. Based on the obtained and The value is used to calculate the maximum charging and discharging power of the battery energy storage during the daily scheduling period, as shown below: in, This represents the maximum charging power of battery energy storage during node j at time t within the day. This represents the maximum discharge power of the battery energy storage during the intraday phase at node j and time period t.

4. The method according to claim 2, characterized in that, A five-segment droop curve intraday scheduling model for battery energy storage (PV) and photovoltaic inverters (QV) based on binary counting and the Big M method is established, including: Establish droop curve models for active charging and discharging of battery energy storage and reactive power of photovoltaic inverters, in order to... Indicates the active power of battery energy storage discharge. Negative value of active power for battery energy storage charging Reactive power of photovoltaic inverters For any one of the following, the intraday scheduling model with five downward-sloping curves is shown below: in, Let be the voltage value at the inflection point of the droop curve of the corresponding device during time period t at node j. For the upper / lower bound of the power of the device droop curve model at node j and time period t, when express hour, and when express hour, and when express hour, and This represents the minimum reactive power / maximum reactive power of the photovoltaic inverter. The dead zone power of the corresponding device droop curve at node j during time period t. and These are the adjustment powers for the second and fourth segments of the droop curve corresponding to the equipment. The voltage of node j in time period t and the voltage at the inflection point of the droop curve. The difference, and These are the slopes of the second and fourth segments of the drooping curve at node j and time period t, respectively. and These represent the minimum and maximum slopes of the droop curve for the corresponding device during time period t at node j. The five-segment droop curve model is reconstructed based on binary counting and the Big M method. First, a vector consisting of three binary decision variables is defined. Five possible binary combinations in this vector are selected to represent the five segments of the droop curve. The binary variable constraints of the droop curve are as follows: in, Let be the binary decision variables for the drooping curve. A vector consisting of three binary decision variables, obtained by... Different binary combinations are used to represent the five segments of the drooping curve; Then, define five coefficient matrices as follows: γ1=[1,1,1], γ2=[-1,1,1], γ3=[1,-1,1], γ4=[-1,-1,1], γ5=[1,1,-1] Finally, the power and voltage corresponding to each segment of the droop curve are represented by multiple inequality constraints. The droop curve reconstruction constraints are shown below: Where M is a specified constant, and the superscript T denotes the matrix transpose symbol; Ultimately, a five-segment droop curve intraday scheduling model A for battery energy storage (PV) and photovoltaic inverter (QV) based on binary counting and the Big M method was formed. The objective of model A is to minimize the power loss of the distribution network, the average voltage deviation of the distribution network, the transaction cost between the distribution network and the upper-level grid, and the operating cost of battery energy storage. The main constraints include battery energy storage operation constraints, distribution network power flow constraints, node voltage constraints, line transmission capacity constraints, binary variable constraints of the droop curve, and droop curve reconfiguration constraints.

5. The method according to claim 1, characterized in that, Using a segmented reduction method, unnecessary segments in the five-segment droop curve intraday scheduling model for each source load uncertainty scenario are identified and discarded, including: A model B for optimal scheduling of active power of battery storage and reactive power of photovoltaic inverter under droop-free control is established. Its objective is to minimize the power loss of distribution network, the average voltage deviation of distribution network, the transaction cost between distribution network and the upper-level grid, and the operating cost of battery storage. The constraints include battery storage operation constraints, distribution network power flow constraints, node voltage constraints, and line transmission capacity constraints. All uncertain scenarios correspond to only one set of optimal decision variables. Solving model B yields a set of optimal active power of battery storage and reactive power output of photovoltaic inverter, which is used as the dead zone power of the droop curve. Based on the dead zone power and the upper and lower limits of equipment power as boundaries, the droop curve is divided into three regions: above the dead zone power, dead zone power, and below the dead zone power. An optimal scheduling model C for the active power of battery storage and the reactive power of photovoltaic inverters without droop control is established. The objective is to minimize distribution network power loss, average voltage deviation, transaction costs between the distribution network and the upstream grid, and operating costs of battery storage. Constraints include battery storage operation constraints, distribution network power flow constraints, node voltage constraints, and line transmission capacity constraints. Each set of uncertainties has a corresponding set of optimal decision variables. Based on the power output values ​​of the corresponding devices obtained from model C, their positions within the defined regions are determined, and unnecessary segments of the droop curve are identified. The judgment criteria are as follows: When the power value is above the dead zone power, the fourth and fifth segments of the droop curve are unnecessary segments; when the power value is equal to the dead zone power, the first and fifth segments of the droop curve are unnecessary segments; when the power value is below the dead zone power, the first and second segments of the droop curve are unnecessary segments.

6. The method according to claim 4, characterized in that, The penalized linear approximation solution method is used to reconstruct the bilinear constraints in the intraday scheduling model of the droop curve. Iterative solutions are then used to obtain the droop curves of the battery storage PV and the photovoltaic inverter QV during the intraday scheduling period, including: (a) Ignoring variables The superscripts and subscripts in the text will be used to represent the bilinear constraints in model A. In short, ΔG = η·ΔV, and the McCormick envelope is used to relax this bilinear constraint as shown below: ΔG≥η min ·ΔV+η·ΔV min -or min ·ΔV min ΔG≥η max ·ΔV+η·ΔV max -or max ·ΔV max ΔG≤η max ·ΔV+η·ΔV min -or max ·ΔV min ΔG≤η min ·ΔV+η·ΔV max -or min ·ΔV max Where, η min and η max ΔV represents the minimum and maximum values ​​of the slope of the sag curve, respectively. min and ΔV max Let represent the minimum and maximum values ​​of the voltage offset, respectively. After relaxation, model D is obtained, and the slope of the droop curve and the initial values ​​of the voltage offset are solved to obtain the initial values. At this point, n = 0; (b) Based on the first-order Taylor expansion formula at point A linear approximation of the bilinear terms in the vicinity is expressed as: in, It is a linearized function; (c) Introduce an auxiliary variable μ to represent With ΔG n The absolute difference between them is expressed as: (d) Add the auxiliary variable as a penalty term to the objective function to establish a penalized linear approximation model E, whose objective function is expressed as: Its constraints include battery energy storage operation constraints, distribution network power flow constraints, node voltage constraints, line transmission capacity constraints, droop curve binary variable constraints, droop curve reconstruction constraints, relaxation constraints based on McCormick envelopes, linear approximation constraints based on first-order Taylor expansion in step (b), and related inequality constraints of auxiliary variable μ in step (c), where π n It is a penalty factor that gradually increases with each iteration; (e) Solve for model E to obtain the optimal droop slope result. Replace the constraints in step (c) of model E with: A model F with a fixed slope is obtained, and the optimization result of its objective equation is set as obj2. Solving this model yields the optimal voltage offset result. Check if the termination criterion is met; if it is, stop the iteration and output the result. The termination criteria are as follows: Where δ is a preset threshold; The sag slope boundary for the next iteration is updated based on the current optimal sag slope value, expressed as: in, and These are the maximum and minimum values ​​of the droop slope in the (n+1)th iteration, respectively. Let ψ be the droop slope obtained in the nth iteration. n It is a parameter greater than 0, used to gradually tighten the boundary of the sag slope; Update penalty factor π n The method is as follows: p n+1 =min{θπ n ,p max } Where θ is a coefficient greater than 1, and π max The maximum value of the pre-set penalty factor.

7. The method according to claim 6, characterized in that, The solution process of the penalized linear approximation method includes: (a) initializationδ,θ,1>ψ>0,n=0,π max ≥π n >0, (b) Establish models A, B, and C, and perform piecewise reduction to identify unnecessary line segments of the drooping curve and reduce model complexity; (c) Based on the McCormick envelope, relax the bilinear constraints in model A to obtain model D, and solve for the initial value of the slope of the sag curve. (d) Based on Establish a penalized linear approximation model E, and solve it to obtain... (e) Based on droop slope Establish a slope-fixed model F and solve for the... (f) Determine if the termination criterion is met. If it is met, stop the iteration and output the result. Otherwise, update the droop slope boundary and penalty factor, let n = n + 1, and go to step (d).

8. A SoC-range managed battery energy storage and photovoltaic inverter droop curve co-optimization system, characterized in that, include: The SoC interval management module is used to establish a battery energy storage state of charge (SoC) interval management model during the day-ahead phase. Based on the generated day-ahead source-load uncertainty scenario of the distribution network, the module uses a stochastic optimization method to solve the battery energy storage SoC interval management model to obtain the optimal SoC interval for battery energy storage during the day. Based on the obtained optimal SoC interval, the module calculates the maximum charging and discharging power of battery energy storage during the day-ahead scheduling period. The droop curve model construction module is used to establish an intraday scheduling model of five droop curves for battery energy storage active power-voltage (PV) and photovoltaic inverter reactive power-voltage (QV) based on binary counting and the big M method, considering the uncertainty of the source and load of the distribution network during the intraday stage. The model simplification module is used to identify and discard unnecessary segments in the five-segment droop curve intraday scheduling model for each source load uncertainty scenario using a segmented reduction method, thereby reducing the number of segments in the droop curve intraday scheduling model corresponding to each uncertainty scenario. The droop curve optimization solution and distribution module is used to reconstruct the bilinear constraints in the intraday scheduling model of the droop curve using a penalized linear approximation method. Iteratively, it obtains the droop curves of the battery storage PV and the photovoltaic inverter QV during the intraday scheduling period, and distributes them to the battery storage and photovoltaic inverter devices. The device controllers then dynamically adjust the active and reactive power outputs based on the real-time measurement information of the local node voltage and the optimized droop curves, achieving efficient response to changes in the power distribution network source load and node voltage.

9. A computer device, characterized in that, include: One or more processors; Memory; And one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, wherein when the programs are executed by the processors, they implement the steps of the battery energy storage and photovoltaic inverter droop curve adaptive optimization method for distribution network hierarchical scheduling as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the SoC range management method for co-optimization of battery energy storage and photovoltaic inverter droop curves as described in any one of claims 1-7.

Images