Main-distributor micro-multi-level collaborative optimization control method and system adapted to non-smooth characteristics

Abstract

The present application discloses a main-distribution-micro multi-level collaborative optimization control method and system that adapts to non-smooth characteristics. The method includes the following steps: constructing a converter equivalent model and a grid topology relationship based on grid measurement data; obtaining a three-phase reactive power segmentation function of distributed energy based on the converter equivalent model and the droop control principle, and using a fitting function for approximate fitting to obtain a droop control function; realizing microgrid decoupling using a node replication method and microgrid neighboring relationships; constructing a multi-level collaborative optimization scheduling model using droop control functions, microgrid decoupling, and an optimization objective function; outputting a main-distribution-microgrid control strategy using an interior point method based on real-time grid data and a multi-level collaborative optimization scheduling model combined with a timing consistency strategy. The beneficial effects of the present application: while reducing the amount of iterative calculations, ensuring that the final control result is based on the characteristics of the distributed energy itself and the global characteristics of the distributed energy, achieving a balance between pertinence and efficiency.

Description

Technical Field

[0001] The present application relates to the field of distributed energy control, and in particular to a main-distributor-micro-multi-level collaborative optimization control method and system that adapts to non-smooth characteristics. Background Art

[0002] The essence of power system optimization dispatch is to ensure the safety, economy and stability of the power system through scientific dispatch and resource allocation. With the increase of the proportion of distributed energy access, the complexity of dispatch increases. Considering that the control curve of distributed energy converter is non-smooth and the derivative is discontinuous, the derivative calculation of the droop control curve of distributed energy converter fails.

[0003] In the related technologies, the distributed energy regulation is calculated by using complex centralized problems. Although the mutual influence relationship between distributed energy can be obtained, the attention to the characteristics of distributed energy itself is weakened by the influence of global attention, resulting in the neglect of the non-smooth characteristics of distributed energy output in the global optimization process, which leads to errors in the final regulation. Although some technologies use piecewise linear approximation to compensate for the non-smooth characteristics of distributed energy, iterative approximation may face the problems of slow convergence speed and long calculation time when dealing with large-scale power system optimization problems, and it is difficult to meet the needs of real-time or near real-time scheduling.

[0004] The patent "A multi-objective power flow optimization method for a new energy multi-terminal flexible DC system", publication number: CN114928057A, publication date: August 19, 2022, specifically discloses that the DC voltage deviation is minimized, the amount of wind and solar power abandoned is minimized, and the static voltage stability margin is maximized as the optimization objectives, and a multi-objective power flow optimization model for a new energy multi-terminal flexible DC collection system is built, and the interior point method and The hybrid alternating algorithm combined with the algorithm solves the multi-objective optimization model. The system coordinated control strategy modeling comprehensively considers the constraints of renewable energy output and coordinated control of converter stations in the power flow optimization model. However, the hybrid algorithm is difficult, computationally inefficient, and relies on high-speed communication during the coordination process.

[0005] The patent "A method for photovoltaic power generation cluster to participate in dynamic voltage control of power system", publication number: CN107134788A, publication date: September 5, 2017, specifically discloses that the photovoltaic cluster PI control architecture is first established; the cluster coordination layer measures the grid connection point voltage, active and reactive power to complete external equivalent value identification, and generates a photovoltaic cluster output reactive power reference value; according to the power flow linearization method, a reactive output distribution model for each photovoltaic power generation unit is established to calculate the reactive output of each unit. However, this solution involves multiple control levels and complex coordination mechanisms, and is inefficient. Summary of the invention

[0006] In order to solve the problem that the prior art cannot balance the pertinence and efficiency of the main and distribution microgrid control, the present application provides a main and distribution micro-multi-level collaborative optimization control method and system that adapts to non-smooth characteristics, constructs a converter equivalent model and a grid topology relationship through a grid measurement model to show the relevant characteristics between the converter and distributed energy, and between distributed energy and distributed energy, and fits the piecewise function through a fitting function to avoid convergence problems caused by non-smoothness, improves the control accuracy, decomposes the complex grid structure into multiple relatively independent microgrids, and improves the computing efficiency by parallel calculation for each microgrid, and then calls the adjacent initial iterative calculation value for boundary coupling calculation through the interior point method combined with the timing consistency strategy, reduces the iterative calculation amount while ensuring that the final control result is based on the distributed energy's own characteristics and the global characteristics of the distributed energy, achieves a balance between pertinence and efficiency, and ensures the consistency of the final control strategy output.

[0007] In order to achieve the above technical objectives, a technical solution provided by the present application is a main-distribution-microgrid multi-level collaborative optimization control method that adapts to non-smooth characteristics, including the following steps: S1: Acquire grid measurement data, and construct a converter equivalent model and a grid topology relationship based on the grid measurement data; S2: Based on the converter equivalent model and the droop control principle, obtain the three-phase reactive power piecewise function of distributed energy, and use the fitting function to approximate the three-phase reactive power piecewise function of distributed energy to obtain the droop control function; S3: Based on the grid topology relationship, obtain the microgrid adjacent relationship, and use the node replication method and the microgrid adjacent relationship to achieve microgrid decoupling; S4: Construct an optimization objective function with the minimum grid operation cost, and use the droop control function, microgrid decoupling and optimization objective function to construct a multi-level collaborative optimization scheduling model; S5: Acquire real-time grid data, and use the interior point method to output the main-distribution-microgrid control strategy based on the real-time grid data and the multi-level collaborative optimization scheduling model combined with the timing consistency strategy.

[0008] Furthermore, constructing a converter equivalent model based on power grid measurement data includes: obtaining converter measurement equations and constraints based on converter measurement principles; constructing a converter equivalent model based on converter position information in power grid topology information and corresponding power measurement data combined with converter measurement equations and constraints.

[0009] Furthermore, the method of realizing microgrid decoupling by using the node replication method and the microgrid neighbor relationship also includes: obtaining a set of adjacent microgrids according to the microgrid neighbor relationship; using the node replication method to introduce a virtual bus for the boundary tie lines of the adjacent microgrids, performing interval decoupling on the adjacent microgrids, and realizing microgrid decoupling.

[0010] Furthermore, the method of constructing a multi-level collaborative optimization scheduling model using a droop control function, microgrid decoupling and an optimization objective function includes: constructing a microgrid optimization model using a droop control function and an optimization objective function; building a boundary coupling optimization layer between microgrid optimization models based on microgrid decoupling; and constructing a multi-level collaborative optimization scheduling model using a microgrid optimization model and a boundary coupling optimization layer.

[0011] Furthermore, the construction of a boundary coupling optimization layer between microgrid optimization models based on microgrid decoupling includes: obtaining boundary coupling constraints of the microgrid optimization model based on microgrid decoupling; obtaining consistent global variables of adjacent microgrids based on the boundary coupling constraints of the microgrid optimization model and the microgrid adjacent relationship; and constructing a boundary coupling optimization layer based on the consistent global variables of adjacent microgrids and the optimization objective function.

[0012] Furthermore, the method of obtaining consistent global variables of adjacent microgrids based on the boundary coupling constraints of the microgrid optimization model and the adjacent relations of the microgrids includes: constructing an augmented Lagrangian function of the adjacent microgrids based on the optimization objective function and the state variables of the virtual bus; integrating the boundary coupling constraints of the adjacent microgrids based on the augmented Lagrangian function to obtain consistent global variables of the adjacent microgrids.

[0013] Furthermore, the S4 also includes: acquiring historical operation data of the power grid, and acquiring the corresponding relationship between the initial value of the consistent global variable and the time series based on the historical operation data of the power grid and the multi-level collaborative optimization scheduling model.

[0014] Furthermore, the method of outputting the main distribution microgrid control strategy using the interior point method based on the real-time data of the power grid and the multi-level collaborative optimization scheduling model in combination with the timing consistency strategy includes: calling the consistency global variable initial value based on the real-time data of the power grid and the corresponding relationship between the consistency global variable initial value and the timing, and outputting the main distribution microgrid control strategy using the interior point method with the consistency global variable initial value, the real-time data of the power grid and the multi-level collaborative optimization scheduling model.

[0015] Furthermore, the method of obtaining the correspondence between the initial values ​​of consistent global variables and the time series based on the historical operation data of the power grid and the multi-level collaborative optimization scheduling model also includes: obtaining the distributed energy output fluctuation value and the load fluctuation value in the historical operation data of the power grid, and taking the distributed energy output fluctuation value and the load fluctuation value corresponding to the time series as the first relationship; obtaining the consistent global variables corresponding to the time series as the second relationship according to the multi-level collaborative optimization scheduling model and the historical operation data of the power grid; and constructing the correspondence between the initial values ​​of the consistent global variables and the time series with the first relationship and the second relationship.

[0016] Furthermore, the calling of the initial value of the consistency global variable based on the real-time data of the power grid and the correspondence between the initial value of the consistency global variable and the timing also includes: obtaining the first difference between the predicted output value of distributed energy and the current output of distributed energy, the predicted load value and the second difference between the current load value according to the real-time data of the power grid; matching the timing corresponding to the first relationship with the first difference and the second difference; matching the consistency global variable corresponding to the second relationship based on the timing corresponding to the first relationship, and using the consistency global variable corresponding to the second relationship as the current initial value of the consistency global variable.

[0017] Furthermore, it also includes: S6: obtaining the average matching times of the consistency global variable in the preset time sequence. If the average matching times is greater than the preset matching times, data screening is performed according to the number of times the consistency global variable initial value is called and the calling time in the corresponding relationship between the consistency global variable initial value and the time sequence.

[0018] Another technical solution provided by the present application is a main-distribution microgrid multi-level collaborative optimization control system that adapts to non-smooth characteristics and is used to implement the above-mentioned method, including: a data acquisition layer for acquiring power grid data; a microgrid mutual assistance layer for building a microgrid optimization scheduling model according to the corresponding microgrid; a main-distribution collaborative control layer for performing collaborative optimization according to the microgrid optimization scheduling model and obtaining the main-distribution microgrid control strategy.

[0019] The beneficial effects of this application are as follows: 1. The converter equivalent model and the grid topology relationship are constructed through the grid measurement model to show the relevant characteristics between the converter and distributed energy, and between distributed energy and distributed energy, so as to construct a three-phase reactive power segmented function for each distributed energy according to the converter equivalent model and the droop control principle, and fit it through the fitting function to avoid the convergence problem caused by non-smoothness and improve the accuracy of regulation. On this basis, according to the adjacent relationship between distributed energy and distributed energy shown in the grid topology relationship, the node replication method is used for decoupling, and the complex grid structure is decomposed into multiple relatively independent microgrids, so as to improve the computing efficiency by parallel computing for each microgrid. Then, a multi-level collaborative optimization scheduling model is constructed by using the droop control function, microgrid decoupling and optimization objective function. Each microgrid is equipped with an optimization model based on the correlation characteristics between the converter and distributed energy, and between distributed energy and distributed energy. When obtaining real-time data from the power grid, each microgrid calculates in parallel, performs independent optimization scheduling, transfers boundary coupling variables, and combines the timing consistency strategy to call the initial iterative calculation value for iterative calculation to finally obtain the global optimal solution. The main and distribution microgrid control strategy is output with the global optimal solution, thereby improving the pertinence and accuracy of the control strategy while using parallel computing on the microgrid side to improve the output efficiency of the control strategy.

[0020] 2. Use the consistency variables corresponding to the coordination of each microgrid in the historical operation data of the power grid, that is, the consistency variables when the optimal control strategy is achieved, as the initial values ​​of the consistency variables of the time series. In actual control, the corresponding initial values ​​of the consistency global variables are called based on the time series. Since distributed energy sources such as light energy and wind energy are greatly affected by the time series, such as the light intensity during the day and at night, the light intensity in summer and winter, or the influence of seasonal wind, distributed energy sources with the same time series may have similar characteristics. At this time, calling the initial values ​​of the consistency global variables by time series can reduce the number of iterations, improve the efficiency of coordination, and reduce the computing pressure.

[0021] 3. The output fluctuation values ​​of distributed energy and load fluctuation values ​​in the historical time series of the power grid are taken as the corresponding time series with the time series similar to the current output fluctuation and load fluctuation of distributed energy, and the consistent global variables are retrieved based on them, so as to avoid the misleading of the retrieval of consistent global variables by annual fluctuations. When the load increase and decrease demands are similar and the output conditions of distributed energy are similar, the regulation of distributed energy must be similar. Using the consistent global variables in this case as the initial value of the current consistent global variables can reduce the number of iterations and improve the regulation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a flow chart of the main-micro multi-level coordinated optimization control method that adapts to the non-smooth characteristics of this application.

[0023] Figure 2 A schematic diagram of an equivalent model of a converter in an embodiment.

[0024] Figure 3 A schematic diagram of converter power relationship in an embodiment.

[0025] Figure 4 This is the converter droop control curve diagram without fitting.

[0026] Figure 5 The figure is a schematic diagram of the master-distributor multi-level coordinated optimization process in an embodiment. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application is further described in detail below in conjunction with the drawings and examples. It should be understood that the specific implementation method described here is only an optimal embodiment of the present application, which is only used to explain the present application and does not limit the scope of protection of the present application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0028] like Figure 1As shown in the first embodiment of the present application, the main-distribution-micro-multi-level collaborative optimization control method adapted to non-smooth characteristics includes the following steps:

[0029] S1: Obtain grid measurement data, and build a converter equivalent model and grid topology relationship based on the grid measurement data;

[0030] S2: Based on the converter equivalent model and the droop control principle, the three-phase reactive power piecewise function of the distributed energy is obtained, and the three-phase reactive power piecewise function of the distributed energy is approximately fitted by the fitting function to obtain the droop control function;

[0031] S3: Obtain the neighbor relationship of microgrids based on the topological relationship of the power grid, and use the node replication method and the neighbor relationship of microgrids to achieve microgrid decoupling;

[0032] S4: Construct an optimization objective function with the minimum grid operation cost, and use the droop control function, microgrid decoupling and optimization objective function to build a multi-level collaborative optimization scheduling model;

[0033] S5: Obtain real-time data of the power grid, and output the main distribution microgrid control strategy using the interior point method based on the real-time data of the power grid and the multi-level collaborative optimization scheduling model combined with the timing consistency strategy.

[0034] In this embodiment, the converter equivalent model and the grid topology are constructed through the grid measurement model to show the relevant characteristics between the converter and the distributed energy, and between the distributed energy and the distributed energy, so as to construct the three-phase reactive power segmented function for each distributed energy according to the converter equivalent model and the droop control principle, and fit it through the fitting function to avoid the convergence problem caused by non-smoothness and improve the accuracy of regulation. On this basis, according to the adjacent relationship between the distributed energy and the distributed energy shown in the grid topology, the node replication method is used for decoupling, and the complex grid structure is decomposed into multiple relatively independent microgrids, so as to improve the computing efficiency by parallel computing for each microgrid. Then, a multi-level collaborative optimization scheduling model is constructed by using the droop control function, microgrid decoupling and optimization objective function. Each microgrid is equipped with an optimization model based on the correlation characteristics between the converter and distributed energy, and between distributed energy and distributed energy. When obtaining real-time data from the power grid, each microgrid calculates in parallel, performs independent optimization scheduling, transfers boundary coupling variables, and combines the timing consistency strategy to call the initial iterative calculation value for iterative calculation to finally obtain the global optimal solution. The main and distribution microgrid control strategy is output with the global optimal solution, thereby improving the pertinence and accuracy of the control strategy while using parallel computing on the microgrid side to improve the output efficiency of the control strategy.

[0035] The grid measurement data at least includes grid topology information and power measurement data. Constructing a converter equivalent model based on the grid measurement data includes:

[0036] Obtain converter measurement equations and constraints based on converter measurement principles;

[0037] Based on the converter location information in the grid topology information and the corresponding power measurement data, combined with the converter measurement equations and constraints, a converter equivalent model is constructed.

[0038] The inverter is a key device in the DC transmission system, which realizes the mutual conversion between AC and DC. Distributed energy is converted from DC to AC through the inverter, and then connected to the grid through the filter and transformer. Figure 2 In one case shown, the transformer connection group is Dy11, the high voltage side is a delta connection, and the low voltage side is a star connection. For the grid connection point, and In between is the line model, which represents the short line between the high-voltage side outlet of the transformer and the grid connection point, and constructs the converter equivalent model.

[0039] Among them, according to the converter measurement principle, the converter measurement includes the active and reactive power measurement of the branch of the grid-connected line, and the converter measurement equation is constructed as follows:

[0040] ;

[0041] in, Indicates phase sequence; Indicates the grid connection point With Node Between Phase branch active power measurement value, Indicates the grid connection point With Node Between Phase branch reactive power measurement value; Indicates the grid connection point With Node Between Actual value of active power in phase branch, Indicates the grid connection point With Node Between Actual value of phase branch reactive power; Indicates the grid connection point With Node Between The measurement error of the phase branch active power, Indicates the grid connection point With Node Between The measurement error of phase branch reactive power.

[0042] In the power system, the busbar is a key node connecting various electrical components (such as generators, transformers, loads, etc.). At the busbar, the power balance constraint must be met, that is, the power flowing into the busbar is equal to the power flowing out of the busbar. Therefore, the constraint condition includes at least the power balance constraint.

[0043] In this embodiment, Figure 2 Take the converter equivalent model shown as an example:

[0044] The power balance constraint of the busbar at the high-voltage side of the transformer is:

[0045] ;

[0046] in, represents the conjugate of a complex number; Indicates the three-phase voltage of the busbar at the grid connection point, Indicates the three-phase voltage of the transformer high-voltage side outlet bus; It indicates the three-phase active power flowing from the transformer high-voltage side outlet bus to the grid connection point bus. Indicates the three-phase active power flowing from the high-voltage side of the transformer to the low-voltage side; It indicates the three-phase reactive power flowing from the transformer high-voltage side outlet bus to the grid connection point bus. It indicates the three-phase reactive power flowing from the high-voltage side to the low-voltage side of the transformer; It represents the branch admittance matrix between the transformer high-voltage side outlet bus and the grid-connected point bus; j represents the imaginary unit.

[0047] The power balance constraint of the bus connected to the filter is:

[0048] ;

[0049] in, Indicates the three-phase voltage of the bus connected to the filter, Indicates the three-phase voltage of the output bus of the distributed energy inverter; It indicates the three-phase active power flowing from the bus connected to the filter to the high-voltage side outlet bus of the transformer. It indicates the three-phase active power flowing from the bus connected to the filter to the output bus of the distributed energy inverter; It represents the three-phase reactive power flowing from the bus connected to the filter to the bus at the high voltage side of the transformer. It represents the three-phase reactive power flowing from the three-phase voltage of the bus connected to the filter to the output bus of the distributed energy inverter; It indicates the three-phase reactive power flowing from the bus connected to the filter to the grounding point. It represents the earth's electrical susceptance; is the branch admittance matrix between the bus connected to the filter and the output bus of the distributed energy inverter; j represents an imaginary unit.

[0050] The power balance constraint of the distributed energy inverter export bus is:

[0051] ;

[0052] in, Represents the three-phase active power of distributed energy, Represents the three-phase reactive power of distributed energy, It represents the three-phase active power flowing from the output bus of the distributed energy inverter to the bus connected to the filter. It represents the three-phase reactive power flowing from the output bus of the distributed energy inverter to the bus connected to the filter.

[0053] In steady state, the zero-sequence current of the converter is zero, and the three-phase voltages of A, B, and C are completely symmetrical. The three-phase symmetry constraint of the output bus of the distributed energy inverter is:

[0054] ;

[0055] in, Indicates the voltage amplitude of phase A of the output bus of the distributed energy inverter. Indicates the voltage amplitude of phase B of the output busbar of the distributed energy inverter. Indicates the voltage amplitude of phase C of the output bus of the distributed energy inverter; Indicates the voltage phase angle of phase A of the distributed energy inverter output bus. Indicates the voltage phase angle of phase B of the distributed energy inverter output bus. Indicates the phase angle of the C-phase voltage at the outlet bus of the distributed energy inverter.

[0056] like Figure 3 As shown in Figure 2, since distributed energy can only inject active power into the system, and the active power and reactive power meet the arc constraints and capacity limitations, Figure 3 At the overcurrent operating point shown, the distributed renewable energy operates at full active power. At this time, the distributed renewable energy also outputs reactive power, and the total power exceeds the capacity limit, which is an overcurrent operation. At the active power limit operating point, the distributed energy reaches the capacity limit, but at this time, the active power and reactive power are both within the safe operating range.

[0057] Therefore, the constraints also include distributed energy arc and capacity constraints:

[0058] ;

[0059] in, Indicates the capacity of distributed energy; represents the power factor, Represents the power factor angle.

[0060] The inverter has five control modes:

[0061] (1) Ratio fixed value control method:

[0062] ;

[0063] in, It represents the ratio of three-phase active power to three-phase reactive power of distributed energy, a constant.

[0064] (2) Power setting control method:

[0065] ;

[0066] in, Indicates the set value of the three-phase active power of distributed energy. Indicates the set value of the three-phase reactive power of distributed energy.

[0067] (3) Active output control method:

[0068] Add to the objective function:

[0069] ;

[0070] When the voltage is within the safety range, the distributed energy will be fully utilized as much as possible to improve the consumption of distributed energy; when the voltage exceeds the safe operating range, some active power can also be cut off to ensure the safe and stable operation of the system.

[0071] (4) Constant power control

[0072] ;

[0073] in, Indicates the active power of phase A, Indicates the active power of phase B, Indicates the active power of phase C; Indicates the reactive power of phase A, Indicates the reactive power of phase B, Indicates the reactive power of phase C; Indicates the total set value of three-phase active power. Indicates the total set value of three-phase reactive power.

[0074] (5) Droop control method

[0075] like Figure 4As shown, droop control is applied at the outlet bus of the distributed energy inverter to control the total reactive power output of the distributed energy and the positive sequence voltage amplitude to meet the droop control curve. First, the positive sequence, negative sequence, and zero sequence voltages corresponding to the three-phase voltage of the outlet bus of the distributed energy inverter are obtained through the phase sequence conversion matrix:

[0076] ;

[0077] Where H represents the phase sequence conversion matrix, , , It represents the positive-sequence, negative-sequence, and zero-sequence voltages corresponding to the three-phase voltage of the distributed energy inverter outlet bus after a 120-degree phase rotation.

[0078] like Figure 4 As shown, it is a schematic diagram of the droop control curve. It represents the sum of three-phase reactive power of distributed energy; It represents the positive sequence voltage amplitude of the output bus of the distributed energy inverter.

[0079] In this embodiment, the ratio constant value control method is suitable for occasions where a specific power needs to be output stably, but it has poor flexibility and cannot respond quickly to changes in load and grid conditions; the power constant value control method is suitable for occasions where there are strict requirements on power output, but it has high requirements on the control system and is difficult to adapt to rapidly changing load and grid conditions; the active output control method is suitable for occasions where distributed energy is connected to the grid, but the control strategy is relatively complex and requires real-time monitoring and adjustment of the grid voltage; the constant power control method is suitable for equipment or systems that need stable output power, but in the case of large load changes, it may not be possible to maintain absolute stability of the system; the droop control method is suitable for microgrids or distributed energy systems, but the control accuracy may be affected by load and grid conditions, and may cause power oscillations in some cases.

[0080] Therefore, in this embodiment, the three-phase reactive power piecewise function of distributed energy is obtained based on the converter equivalent model and the droop control principle:

[0081] ;

[0082] in, Indicates the total reactive power of three phases The upper limit of Indicates the total reactive power of three phases The lower limit of Indicates the three-phase reactive power when the positive sequence voltage amplitude is within the dead zone; Indicates the upper limit of the voltage dead zone, Indicates the lower limit of the voltage dead zone; represents the upper limit of voltage droop control, Indicates the lower limit of voltage droop control; Indicates the upper limit of the bus voltage amplitude, Indicates the lower limit of bus voltage amplitude; denote the droop control coefficient, and , .

[0083] Through the converter equivalent model and droop control principle, a distributed energy three-phase reactive power piecewise function is constructed according to the actual conditions of each microgrid to improve the accuracy of distributed energy three-phase reactive power calculation. The fitting function is used to approximate the piecewise function of the distributed energy three-phase reactive power to obtain the fitted droop control function:

[0084] ;

[0085] in, is the fitting coefficient.

[0086] In this embodiment, by constructing a converter equivalent model for each microgrid, the system behavior under different control strategies is simulated to adapt to different grid environments and load requirements. The distributed energy droop control function curve is smoothed by an approximate fitting method, which achieves precise control of power distribution, avoids derivative discontinuity and calculation failure, and reduces the error of optimization operation results.

[0087] The use of node replication and microgrid neighbor relationships to achieve microgrid decoupling also includes:

[0088] Obtaining a set of adjacent microgrids according to the adjacent relationship of microgrids;

[0089] The node replication method is used to introduce virtual busbars into the boundary tie lines of adjacent microgrids, and the adjacent microgrids are decoupled to achieve microgrid decoupling.

[0090] By introducing a virtual bus and realizing interval decoupling of adjacent microgrids, relatively independent calculations of each microgrid are achieved while ensuring the coordination and consistency of adjacent microgrids in the optimization scheduling process. This allows each microgrid to not only independently run an optimization algorithm to solve its own optimal operation strategy, but also negotiate through iteration and boundary coupling to achieve a global optimal or near-optimal solution, thereby improving the accuracy of distributed energy control. At the same time, parallel computing of microgrids is used to compensate for the complexity brought about by decoupling, thereby ensuring the efficiency of distributed energy control.

[0091] The grid topology information also includes the microgrid topology relationship, and the microgrid neighbor relationship is obtained based on the geographical proximity of the microgrids. busbars, and by Microgrids are interconnected. For any microgrid , which is coupled with the adjacent microgrid through a tie line.

[0092] Among them, microgrid and Microgrid Through the border line Coupled connection, introducing a pair of "virtual busbars" in the middle of the tie line Adjacent microgrids are decoupled in sections. Each decoupled microgrid contains the original local bus and the newly introduced virtual bus, and can be independently optimized and scheduled.

[0093] The virtual bus can be equivalent to a "virtual distributed power source". Represents microgrid The state variables of the virtual bus, Represents microgrid The state variables of the virtual bus include the three-phase voltage phase angle, amplitude, and the three-phase active and reactive power emitted by the virtual bus. , Represents microgrid The three-phase voltage phase angle of the virtual bus, Represents microgrid The three-phase voltage amplitude of the virtual bus, Represents microgrid The three-phase active power of the virtual bus, Represents microgrid Three-phase reactive power of virtual bus; , Represents microgrid The three-phase voltage phase angle of the virtual bus, express The three-phase voltage amplitude of the microgrid virtual bus, Represents microgrid The three-phase active power of the virtual bus, Represents microgrid Three-phase reactive power of the virtual bus.

[0094] Since the virtual bus Connect the line Split into two parts of equal length, so the line after splitting and The branch admittance becomes twice of , For line The branch admittance, For line The branch admittance, For contact line The branch admittance of the line after the split is doubled to include the newly introduced virtual bus. It is described by an admittance matrix of order.

[0095] The newly introduced “virtual bus” is a boundary coupling bus used by adjacent microgrids for information exchange. In order to achieve the equivalence of distributed optimization scheduling and centralized optimization scheduling, the state variables obtained by solving the coupled bus of adjacent sub-microgrids must be equal. The state variables of the virtual bus are used as the boundary coupling constraints of the microgrid:

[0096] , , , ;

[0097] , ;

[0098] , ;

[0099] , ;

[0100] , ;

[0101] in, Represents microgrid The voltage amplitude of phase A of the virtual bus, Represents microgrid The voltage amplitude of phase B of the virtual bus, Represents microgrid The voltage amplitude of phase C of the virtual bus; Represents microgrid The voltage amplitude of phase A of the virtual bus, Represents microgrid The voltage amplitude of phase B of the virtual bus, Represents microgrid The voltage amplitude of phase C of the virtual bus; Represents microgrid The phase angle of the A phase voltage of the virtual bus, Represents microgrid The phase angle of the virtual busbar's phase B voltage, Represents microgrid The phase angle of the C phase voltage of the virtual bus; Represents microgrid The phase angle of the A phase voltage of the virtual bus, Represents microgrid The phase angle of the virtual busbar's phase B voltage, Represents microgrid The phase angle of the C phase voltage of the virtual bus; Represents microgrid The active power of phase A of the virtual bus, Represents microgrid The B-phase active power of the virtual bus, Represents microgrid The C-phase active power of the virtual bus; Represents microgrid The active power of phase A of the virtual bus, Represents microgrid The B-phase active power of the virtual bus, Represents microgrid The C-phase active power of the virtual bus; Represents microgrid The reactive power of phase A of the virtual bus, Represents microgrid The B-phase reactive power of the virtual bus, Represents microgrid C-phase reactive power of virtual bus; Represents microgrid Phase A reactive power of the virtual bus, Represents microgrid The B-phase reactive power of the virtual bus, Represents microgrid Phase C reactive power of the virtual bus.

[0102] The optimization objective function is constructed with the minimum grid operation cost as follows:

[0103] ;

[0104] ;

[0105] ;

[0106] ;

[0107] ;

[0108] in, Represents microgrid The objective function of Represents microgrid The set of all variables in represents the total number of microgrids, Represents microgrid The equality constraint, Represents microgrid The inequality constraints are Represents microgrid Consistency global variables.

[0109] In this embodiment, the equality constraints of the microgrid are used to ensure the energy or power balance within the system, such as power balance constraints, and the inequality constraints of the microgrid are used to limit the operating status and output power range of each component in the system to ensure the safe and stable operation of the system, such as equipment output constraints, safety constraints, etc.

[0110] The multi-level collaborative optimization dispatch model is constructed by using droop control function, microgrid decoupling and optimization objective function, including:

[0111] The microgrid optimization model is constructed using the droop control function and the optimization objective function;

[0112] Build boundary coupling optimization layer between microgrid optimization models based on microgrid decoupling;

[0113] A multi-level collaborative optimization scheduling model is constructed based on the microgrid optimization model and boundary coupling optimization layer.

[0114] The boundary coupling optimization layer between microgrid optimization models based on microgrid decoupling includes:

[0115] Obtain boundary coupling constraints of microgrid optimization model based on microgrid decoupling;

[0116] The consistent global variables of adjacent microgrids are obtained based on the boundary coupling constraints of the microgrid optimization model and the adjacent relationship of microgrids;

[0117] The boundary coupling optimization layer is constructed with the consistent global variables of adjacent microgrids and the optimization objective function.

[0118] In multi-level collaborative optimization scheduling, each microgrid performs independent optimization scheduling and then transfers boundary coupling variables. Therefore, in the framework of multi-level collaborative optimization scheduling, the boundary coupling constraints are unified into the objective function, and the consistent global variables of adjacent microgrids are obtained by using the boundary coupling constraints of the microgrid optimization model and the adjacent relationship of microgrids, including:

[0119] The augmented Lagrangian function of the adjacent microgrid is constructed based on the optimization objective function and the state variables of the virtual bus.

[0120] Based on the augmented Lagrangian function, the boundary coupling constraints of adjacent microgrids are integrated to obtain the consistent global variables of adjacent microgrids.

[0121] For example, building a microgrid and Microgrid The corresponding augmented Lagrangian function and :

[0122] ;

[0123] ;

[0124] in, Indicates the number of iterations; Represents microgrid No. The augmented Lagrange multiplier vector for the iteration, Represents microgrid No. The augmented Lagrange multiplier vector for the iteration; represents the penalty parameter of the ADMM algorithm; Represents microgrid The objective function of Represents microgrid No. Consistent global variables for iterations; Represents microgrid No. The consistency global variables for each iteration.

[0125] According to the two micro-networks The average value of the boundary coupling constraints obtained by iterations is used to obtain the consistent global variable as a fixed reference value for the adjacent microgrid:

[0126] ;

[0127] in, Represents microgrid No. The virtual bus state variable of the iteration, Represents microgrid No. The virtual bus state variables of the iteration.

[0128] In this embodiment, the boundary coupling constraint information is highlighted by the virtual bus state variable, the real-time data of the power grid is obtained, and the real-time data of the power grid is used as input. The multi-level collaborative optimization scheduling model is used to perform independent calculations of each microgrid. In the boundary coupling optimization layer, the consistency global variables and Lagrange multiplier information are updated based on the boundary coupling constraints output by each microgrid. The boundary coupling constraint information between each microgrid is exchanged to determine whether the convergence requirements are met. If so, the corresponding main and distribution microgrid control strategy is output.

[0129] like Figure 5 As shown, in this embodiment, step S5 also includes:

[0130] Obtain real-time data of the power grid, and perform iterative calculations using the interior point method based on the real-time data of the power grid and the multi-level collaborative optimization scheduling model combined with the timing consistency strategy to obtain the consistent global variable update value;

[0131] The residual of the updated value of the consistency global variable is used to determine whether the convergence requirements are met. If so, the iteration is considered completed and the current main distribution microgrid control strategy is output.

[0132] The residual of the updated value of the consistency global variable includes at least the original residual and the dual residual. The interior point method is an algorithm for solving linear programming or nonlinear convex optimization problems. It improves computational efficiency and stability by finding the optimal solution inside the feasible domain rather than searching along the boundary. In this embodiment, a search path is constructed so that the consistency global variable continuously approaches the optimal point from the initial value of the consistency global variable, and at each point of the search path, the constraints of the optimization objective function are met.

[0133] The original residual is:

[0134] ;

[0135] in, Represents microgrid No. The raw residual of the consistent global variable update value of iteration, Represents microgrid Updated values ​​of consistent global variables at the kth iteration.

[0136] The dual residual is:

[0137] ;

[0138] in, Represents microgrid No. The dual residual of the consistent global variable update value at iteration .

[0139] The convergence requirements are:

[0140] ;

[0141] in, represents the sum of residuals of the consistent global variable update values ​​at the kth iteration, Indicates the convergence accuracy.

[0142] In this embodiment, each microgrid optimizes and calculates its own non-smooth characteristics, i.e., the droop control function, in combination with the optimization objective function, and solves the actual control conditions of each microgrid. At the same time, the constraint conditions of each adjacent microgrid are integrated with the optimization objective function by using the boundary coupling optimization layer, and a global integrated optimization objective function is constructed to highlight the mutual influence and coordination relationship between adjacent microgrids, so as to ensure that the global power balance, voltage stability, and line transmission capacity and other constraint conditions are met. The interior point method is combined with the augmented Lagrangian function for iteration, and the decisions of each microgrid are coordinated through the update of the consistent global variables and the Lagrangian multiplier. When the integrated optimization objective function converges, that is, when the residual of the consistent global variable update value meets the convergence requirement, it is considered that the decisions of each microgrid have been coordinated with each other and have reached global optimality, and the optimization calculation results of each microgrid are output as the control strategy of the microgrid.

[0143] The piecewise function is approximately fitted by fitting functions to achieve smooth control of the function curve; the microgrid is divided by the node replication method to decouple the centralized problem into multiple sub-problems; a multi-level collaborative two-stage distributed robust optimization model is established, and the interior point method is used to solve it. The multiplier information and consistency variable updates are realized for the exchange coupling information of adjacent microgrids, thereby achieving fast and effective collaborative optimization operation, solving the non-smooth limitations of the control curve, and reducing the error of the optimization operation results.

[0144] As a second embodiment of the present application, step S4 includes:

[0145] S41: Construct an optimization objective function with the minimum grid operation cost, and use the droop control function, microgrid decoupling and optimization objective function to build a multi-level collaborative optimization scheduling model;

[0146] S42: Obtain historical operation data of the power grid, and obtain the initial value of the consistent global variable and the corresponding relationship between the time series based on the historical operation data of the power grid and the multi-level collaborative optimization scheduling model.

[0147] Correspondingly, based on the real-time data of the power grid and the multi-level collaborative optimization dispatch model combined with the timing consistency strategy, the main distribution microgrid control strategy output by the interior point method also includes:

[0148] Based on the real-time data of the power grid and the correspondence between the initial values ​​of the consistent global variables and the time series, the initial values ​​of the consistent global variables are called, and the control strategy of the main distribution microgrid is output using the interior point method with the initial values ​​of the consistent global variables, the real-time data of the power grid and the multi-level collaborative optimization scheduling model.

[0149] In this embodiment, the consistency variables corresponding to the coordination of each microgrid in the historical operation data of the power grid, that is, the consistency variables when the optimal control strategy is achieved, are used as the initial values ​​of the consistency variables of the time series. In actual control, the corresponding initial values ​​of the consistency global variables are called based on the time series. Since distributed energy sources such as light energy and wind energy are greatly affected by the time series, such as the light intensity during the day and at night, the light intensity in summer and winter, or the influence of seasonal wind, distributed energy sources with the same time series may have similar characteristics, and a time series consistency strategy is constructed. At this time, calling the initial value of the consistency global variable with the historical time series corresponding to the current time series can reduce the number of iterations, improve the efficiency of coordination, and reduce the computing pressure.

[0150] Since the regulation of distributed energy is usually based on the prediction of distributed energy output and load, the real-time data of the power grid at least includes the predicted output value of distributed energy, the current output of distributed energy, the predicted load value and the current load value. At this time, the corresponding relationship between the initial value of the consistent global variable and the time series obtained based on the historical operation data of the power grid and the multi-level collaborative optimization scheduling model also includes:

[0151] Obtain the output fluctuation value and load fluctuation value of distributed energy in the historical operation data of the power grid, and use the output fluctuation value and load fluctuation value of distributed energy corresponding to the time series as the first relationship;

[0152] According to the multi-level collaborative optimization dispatch model and the historical operation data of the power grid, the consistent global variable corresponding to the time series is obtained as the second relationship;

[0153] The first relationship and the second relationship are used to construct the correspondence between the initial value of the consistent global variable and the time series.

[0154] Calling the consistency global variable initial value based on the real-time data of the power grid and the correspondence between the consistency global variable initial value and the timing also includes:

[0155] Obtaining a first difference between a predicted output value of distributed energy and a current output value of distributed energy, a predicted load value, and a second difference between a current load value according to real-time data of the power grid;

[0156] Matching the timing corresponding to the first relationship with the first difference and the second difference;

[0157] The consistency global variable corresponding to the second relationship is matched based on the time sequence corresponding to the first relationship, and the consistency global variable corresponding to the second relationship is used as the initial value of the current consistency global variable.

[0158] In this case, it does not directly correspond to the current time series of the power grid and the historical time series of the power grid. Instead, the distributed energy output fluctuation value and load fluctuation value in the historical time series of the power grid are taken as the corresponding time series, and the consistent global variables are retrieved based on this, so as to avoid the misleading of the consistent global variable retrieval due to annual fluctuations. When the load increase and decrease demands are similar and the distributed energy output conditions are similar, the regulation of distributed energy must be similar. Using the consistent global variable in this case as the initial value of the current consistent global variable can reduce the number of iterations and improve the regulation efficiency, thereby ensuring the real-time nature of the regulation.

[0159] It is understandable that since the historical state and the actual state cannot be completely consistent, a similarity threshold can be set. When the weighted average of the similarity between the first difference and the distributed energy output fluctuation value and the second difference and the load fluctuation value is higher than the similarity threshold, the two are considered to match. When there are multiple matching results, multiple consistency global variables can be called for weighted average to obtain the final consistency global variable initial value. Through the similarity between the current output regulation and the historical regulation, the historical consistency global variable is correspondingly called as the initial value of the current iteration, which reduces the number of iterations, improves the regulation efficiency, and reduces the calculation pressure.

[0160] As a third embodiment of the present application, the main-distribution-micro-multi-level collaborative optimization control method adapted to non-smooth characteristics also includes:

[0161] S6: Obtain an average number of matches of the consistency global variable within a preset time series. If the average number of matches is greater than the preset number of matches, data filtering is performed based on the number of times the consistency global variable initial value is retrieved and the retrieval time in the corresponding relationship between the consistency global variable initial value and the time series.

[0162] In this embodiment, it is considered that if the amount of data in the correspondence between the initial value of the consistency global variable and the time series is too large, it will also affect the control strategy. Therefore, when the average number of matches of the consistency global variable in the preset time series is greater than the preset number of matches, it is considered that the amount of data in the correspondence between the initial value of the consistency global variable and the time series is too large. The average number of matches can avoid the misleading of the actual situation assessment due to occasional excessive matches. Data screening is performed according to the number of calls and the call time of the initial value of the consistency global variable in the correspondence between the initial value of the consistency global variable and the time series, thereby avoiding the impact of invalid data on matching efficiency, further improving the output efficiency of the control strategy, and reducing data storage pressure.

[0163] The pre-stored consistency global variables are weightedly calculated based on the number of times the consistency global variable initial value is called and the call time in the correspondence relationship between the consistency global variable initial value and the timing, and the data with a weighted calculation value less than a preset important threshold is deleted. The consistency global variable initial value with more calls and closer call time has a higher weight, thereby deleting the consistency global variable initial value with fewer calls or even no calls, and the call time is too far away, thereby improving the consistency global variable matching efficiency. It can be understood that in some cases, the correspondence relationship between the consistency global variable initial value and the timing can also delete the timing data. At this time, if there are multiple first relationships corresponding to the timing corresponding to the consistency global variable in the correspondence relationship between the consistency global variable initial value and the timing, and the timing corresponding to the consistency global variable is the same, then the multiple first relationships are all corresponded to the consistency global variable, and deleting the timing data further reduces the pressure on data storage.

[0164] As a fourth embodiment of the present application, a main-distribution-micro-multi-level collaborative optimization control system adapted to non-smooth characteristics includes:

[0165] Data collection layer, used to obtain power grid data;

[0166] The microgrid mutual assistance layer is used to construct a microgrid optimization dispatch model according to the microgrid correspondence;

[0167] The main and distribution coordinated control layer is used to perform coordinated optimization according to the microgrid optimization scheduling model and obtain the main and distribution microgrid control strategy.

[0168] The data acquisition layer outputs the grid data, including grid measurement data and real-time grid data, to the microgrid mutual assistance layer, such as the grid connection point of the substation, the total real-time output, the total installed capacity, the controllable range, etc. The microgrid mutual assistance layer constructs the converter equivalent model based on the converter position information in the grid topology information and the corresponding power measurement data combined with the converter measurement equation and the constraint conditions, and uses the droop control principle and the optimization objective function to obtain the microgrid optimization model corresponding to each microgrid. In the microgrid mutual assistance layer, the piecewise function is approximated by the fitting function to achieve smooth control of the function curve, and the microgrid is divided by the node replication method to decouple the centralized problem into multiple sub-problems. The main and distribution coordinated control layer constructs the boundary coupling optimization layer according to the boundary coupling constraints and optimization objective functions of each microgrid optimization model, and uses the interior point method to solve it, and realizes the Lagrange multiplier information and consistent global variable update for the exchange coupling information of adjacent microgrids. The microgrid mutual assistance layer transmits boundary coupling constraint information to the main and distribution coordinated control layer through the microgrid adjacent relations and microgrid optimization model constraints, such as microgrid connection points, microgrid controllable upper and lower limits, real-time total output, etc. The main and distribution coordinated control layer transmits the coordinated control results back to the microgrid mutual assistance layer based on the boundary coupling optimization iterative calculation. The microgrid mutual assistance layer outputs the corresponding substation control instructions and substation gateway real-time output setting values ​​according to the coordinated control results to realize the main and distribution microgrid control.

[0169] The specific implementation method described above is a preferred implementation method of the main-micro multi-level collaborative optimization control method and system that adapts to non-smooth characteristics of the present application, and is not intended to limit the specific implementation scope of the present application. The scope of the present application includes but is not limited to the specific implementation method. All equivalent changes made in accordance with the shape and structure of the present application are within the protection scope of the present application.

Claims

1. A multi-level coordinated optimization control method for main and micro components that adapts to non-smooth characteristics, characterized by: The steps include: S1: Obtain grid measurement data, and build a converter equivalent model and grid topology relationship based on the grid measurement data; S2: Based on the converter equivalent model and the droop control principle, the three-phase reactive power piecewise function of the distributed energy is obtained, and the three-phase reactive power piecewise function of the distributed energy is approximately fitted by the fitting function to obtain the droop control function; S3: Obtain the neighbor relationship of microgrids based on the topological relationship of the power grid, and use the node replication method and the neighbor relationship of microgrids to achieve microgrid decoupling; S4: Construct an optimization objective function with the minimum grid operation cost, and use the droop control function, microgrid decoupling and optimization objective function to build a multi-level collaborative optimization scheduling model; S5: Obtain real-time data of the power grid, and output the main distribution microgrid control strategy using the interior point method based on the real-time data of the power grid and the multi-level collaborative optimization dispatch model combined with the timing consistency strategy; The node replication method is used to introduce virtual busbars into the boundary tie lines of adjacent microgrids. The augmented Lagrangian function of the adjacent microgrid is constructed based on the optimization objective function and the state variables of the virtual bus. Based on the augmented Lagrangian function, the boundary coupling constraints of adjacent microgrids are integrated to obtain the consistent global variables of adjacent microgrids. Obtain the output fluctuation value and load fluctuation value of distributed energy in the historical operation data of the power grid, and use the output fluctuation value and load fluctuation value of distributed energy corresponding to the time series as the first relationship; According to the multi-level collaborative optimization dispatch model and the historical operation data of the power grid, the consistent global variable corresponding to the time series is obtained as the second relationship; Constructing the correspondence between the initial value of the consistent global variable and the time sequence based on the first relationship and the second relationship; Obtaining a first difference between a predicted output value of distributed energy and a current output value of distributed energy, a predicted load value, and a second difference between a current load value according to real-time data of the power grid; Matching the timing corresponding to the first relationship with the first difference and the second difference; The consistency global variable corresponding to the second relationship is matched based on the time sequence corresponding to the first relationship, and the consistency global variable corresponding to the second relationship is used as the initial value of the current consistency global variable.

2. The main-distributor-microcontroller multi-level collaborative optimization control method adapted to non-smooth characteristics according to claim 1 is characterized in that: The construction of a converter equivalent model according to the power grid measurement data includes: Obtain converter measurement equations and constraints based on converter measurement principles; Based on the converter location information in the grid topology information and the corresponding power measurement data, combined with the converter measurement equations and constraints, a converter equivalent model is constructed.

3. The main-distributor-microcontroller multi-level collaborative optimization control method adapted to non-smooth characteristics according to claim 1 is characterized in that: The method of realizing microgrid decoupling by using the node replication method and the microgrid neighbor relationship also includes: Obtaining a set of adjacent microgrids according to the adjacent relationship of microgrids; The node replication method is used to introduce virtual busbars into the boundary tie lines of adjacent microgrids, and the adjacent microgrids are decoupled to achieve microgrid decoupling.

4. The main-distributor-microcontroller multi-level collaborative optimization control method adapted to non-smooth characteristics according to claim 1 is characterized in that: The multi-level collaborative optimization scheduling model constructed by using the droop control function, microgrid decoupling and optimization objective function includes: The microgrid optimization model is constructed using the droop control function and the optimization objective function; Build boundary coupling optimization layer between microgrid optimization models based on microgrid decoupling; A multi-level collaborative optimization scheduling model is constructed based on the microgrid optimization model and boundary coupling optimization layer.

5. The main-distributor-microcontroller multi-level coordinated optimization control method adapted to non-smooth characteristics according to claim 4 is characterized in that: The boundary coupling optimization layer between microgrid optimization models constructed based on microgrid decoupling includes: Obtain boundary coupling constraints of microgrid optimization model based on microgrid decoupling; The consistent global variables of adjacent microgrids are obtained based on the boundary coupling constraints of the microgrid optimization model and the adjacent relationship of microgrids; The boundary coupling optimization layer is constructed with the consistent global variables of adjacent microgrids and the optimization objective function.

6. The main-distributor-microcontroller multi-level coordinated optimization control method adapted to non-smooth characteristics according to claim 5 is characterized in that: The method of outputting the main distribution microgrid control strategy by using the interior point method according to the real-time data of the power grid and the multi-level collaborative optimization dispatch model combined with the timing consistency strategy includes: Based on the real-time data of the power grid and the correspondence between the initial values ​​of the consistent global variables and the time series, the initial values ​​of the consistent global variables are called, and the control strategy of the main distribution microgrid is output using the interior point method with the initial values ​​of the consistent global variables, the real-time data of the power grid and the multi-level collaborative optimization scheduling model.

7. The main-distributor-microcontroller multi-level coordinated optimization control method adapted to non-smooth characteristics according to claim 6 is characterized in that: Also includes: S6: Obtain an average number of matches of the consistency global variable within a preset time series. If the average number of matches is greater than the preset number of matches, data filtering is performed based on the number of times the consistency global variable initial value is retrieved and the retrieval time in the corresponding relationship between the consistency global variable initial value and the time series.

8. A main-distributor-microcontroller multi-level coordinated optimization control system adapted to non-smooth characteristics, used to implement the method according to any one of claims 1 to 7, characterized in that: include: Data collection layer, used to obtain power grid data; The microgrid mutual assistance layer is used to construct a microgrid optimization dispatch model according to the microgrid correspondence; The main and distribution coordinated control layer is used to perform coordinated optimization according to the microgrid optimization scheduling model and obtain the main and distribution microgrid control strategy.

Images