Micro-grid secondary system networking based on elastic matching and a regulating method thereof

Abstract

The application discloses a micro-grid secondary system networking based on elastic matching and a regulating method thereof, real-time collection of system data; according to the collected system data, a matching combination of micro-grid power distribution is obtained under the constraint of constraint conditions with the lowest total cost and elasticity measurement as a target, an optimized scheduling instruction is formulated according to the obtained matching combination; the optimized scheduling instruction is executed; the system state after the execution of the optimized scheduling instruction is fed back; and the elasticity measurement in the elastic matching is optimized and corrected through a neural network prediction model according to the fed-back system state. The elasticity measurement based on the neural network correction is used for flexible networking, the controllable distributed power supply, energy storage and other equipment in different modules are combined, optimized and coordinately controlled by using the elasticity measurement, so that energy consumption elasticity and power cost minimization are realized while the micro-grid secondary system is flexibly networked.

Description

Technical Field

[0001] This invention relates to flexible networking of microgrids, specifically to a microgrid secondary system networking and control method based on elastic matching. Background Technology

[0002] The secondary system of a microgrid is the power distribution and control system within the microgrid. It manages energy flow, load scheduling, voltage and frequency control, and typically consists of inverters, distribution equipment, battery storage systems, and smart controllers. It enables the monitoring, control, and optimization of various devices within the microgrid. Compared to distributed generation technology, microgrids offer more significant advantages. They can improve power system reliability while achieving coordinated interaction between power sources, grid, and loads, and multi-energy complementarity among distributed energy sources. They also increase the distribution network's capacity to absorb distributed power sources and improve the utilization rate of clean energy. However, the connection of a large number of constant power loads can affect the stable operation of the microgrid. The switching of the operating states of the grid-connected devices under different modes can also introduce intermittent and random disturbances to the system. Various unstable factors in the system may cause the bus voltage to fluctuate at the critical point of the voltage range, exacerbating system oscillations. Summary of the Invention

[0003] Purpose of the invention: To address the above-mentioned shortcomings, this invention provides a microgrid secondary system networking based on elastic matching that improves the reliability of microgrid operation and control, achieves flexible energy consumption, and minimizes electricity costs.

[0004] The present invention also provides a control method for microgrid secondary system networking based on elastic matching.

[0005] Technical solution: To solve the above problems, this invention adopts a microgrid secondary system networking based on elastic matching, including:

[0006] The information acquisition layer is used to collect system data in real time.

[0007] The optimized scheduling layer is used to obtain a matching combination of microgrid power distribution under constraints, based on the collected system data and with the goal of minimizing total cost and elasticity metrics. Based on the obtained matching combination, optimized scheduling instructions are formulated. Microgrid power distribution includes the purchase of additional electricity, the charging power of energy storage systems, and the discharging power of energy storage systems. The elasticity metrics include supply elasticity, which reflects the response of distributed energy resources to changes in environmental conditions; load balance elasticity, which reflects the system's ability to balance loads under load changes; and demand response elasticity, which reflects the degree of user response to changes in electricity prices.

[0008] The module control layer is used to execute optimized scheduling instructions;

[0009] The information feedback layer is used to provide feedback on the system status after executing optimized scheduling instructions;

[0010] The parameter optimization layer is used to correct the resilience measure based on the feedback system state through a neural network prediction model.

[0011] Furthermore, the information acquisition layer includes an operation monitoring data acquisition module and a microgrid comprehensive monitoring index calculation module;

[0012] The operation monitoring data acquisition module is used to collect operation monitoring parameters of the microgrid's operating status in real time, specifically including: total power demand, total power supply, comprehensive energy efficiency of distributed generation, energy storage device efficiency, microgrid energy storage device data, microgrid key equipment data, and microgrid communication equipment data.

[0013] The microgrid comprehensive monitoring index calculation module is used to calculate the microgrid comprehensive monitoring index data based on the real-time collected operation monitoring parameters. Specifically, it includes: microgrid overall power balance index, distributed energy response speed index, energy storage device power balance index, energy storage device charge and discharge efficiency stability index, key equipment working status monitoring index, and network communication reliability index.

[0014] Furthermore, the module control layer includes: a PCC module, a microgrid AC regulation module, a microgrid AC non-regulatory module, a microgrid DC regulation module, a microgrid DC non-regulatory module, and an interconnection module;

[0015] The PCC module switches between on-grid and off-grid operation of the microgrid using a limited switching power control method;

[0016] The microgrid AC regulation module includes one or more distributed power sources, energy storage systems, and grid connection devices;

[0017] The microgrid's non-adjustable AC module includes AC loads, non-dispatchable distributed power sources, and networking devices.

[0018] The microgrid DC regulation module includes an energy storage system with regulation capability and a bidirectional DC / DC chopper for controlling the DC bus voltage of the microgrid.

[0019] The microgrid's non-adjustable DC module includes DC loads, non-dispatchable distributed power sources, and grid-connected devices;

[0020] The interconnection module includes an interconnection converter, with its two ends connected to the microgrid AC bus and the microgrid DC bus, respectively.

[0021] Furthermore, the parameter optimization layer uses a BO-BiGRU-Attention hybrid neural network prediction model to correct the elasticity metric. The specific steps are as follows:

[0022] Step 51: Obtain historical data of distributed energy resources, time series data of charging and discharging power of energy storage systems, historical data of total power demand of microgrids, and historical data of power demand of each user equipment or system within the microgrid, and perform preprocessing.

[0023] Step 52: Construct the initial BiGRU-Attention model. Train the BiGRU-Attention model using the preprocessed data to obtain the initial BiGRU-Attention model.

[0024] Step 53: Hyperparameter optimization. The learning rate, training window size, number of hidden layer neurons, and number of BiGRU layer neurons of the initial BiGRU-Attention model are optimized using the Bayesian optimization algorithm. The BO-BiGRU-Attention model is obtained based on the optimized parameters of the initial BiGRU-Attention model.

[0025] Step 54: Obtain distributed energy data, energy storage system charging and discharging data, total power demand data of microgrid, and power demand data of each user device or system within the microgrid by predicting the BO-BiGRU-Attention model. Calculate the elasticity metric based on the predicted data to obtain the corrected elasticity metric.

[0026] This invention also employs a control method for microgrid secondary system networking based on elastic matching, comprising the following steps:

[0027] Step 1: Collect system data in real time;

[0028] Step 2: Based on the collected system data, with the goal of minimizing total cost and elasticity metrics, a matching combination of microgrid power distribution is obtained under constraints. Optimized dispatch instructions are then formulated based on the obtained matching combination. Microgrid power distribution includes the purchase of additional electricity, the charging power of the energy storage system, and the discharging power of the energy storage system. The elasticity metrics include supply elasticity (reflecting the distributed energy source's responsiveness to changes in environmental conditions), load balance elasticity (reflecting the system's balancing ability under load changes), and demand response elasticity (reflecting the user's responsiveness to changes in electricity prices).

[0029] Step 3: Execute the optimized scheduling instruction;

[0030] Step 4: Feedback on the system status after executing the optimized scheduling instructions;

[0031] Step 5: Based on the feedback of the system status, the elasticity measure in elastic matching is corrected using a neural network prediction model;

[0032] Step 6: Formulate optimized scheduling instructions based on the matching combination of the corrected elasticity metrics, and return to execute step 3.

[0033] Furthermore, the collected system data includes total power demand, total power supply, comprehensive energy efficiency of distributed generation, energy storage device efficiency, microgrid energy storage device data, microgrid key equipment data, microgrid communication equipment data, microgrid overall power balance index, distributed energy response speed index, energy storage device power balance index, energy storage device charge / discharge efficiency stability index, key equipment operating status monitoring index, and network communication reliability index.

[0034] Furthermore, with power balance, voltage stability, and frequency stability of the microgrid as constraints, the optimization problem of minimizing the total cost and resilience metric of the microgrid secondary system is solved based on the conjugate gradient algorithm:

[0035]

[0036] In the formula, w c As a cost weight, w d For demand response elasticity weights, w L For supply elasticity weights, w s E is the load balancing elastic weight. d For demand response elasticity, E L For load balancing flexibility, E s For supply elasticity, C buy The cost of purchasing additional electricity, C charge The cost of charging an energy storage system, C discharge P represents the cost corresponding to the discharge of the energy storage system, α is the ratio of the charging power to the discharging power of the energy storage system, and P is the cost of discharging the energy storage system. charge The charging power of the energy storage system, P max_charge P is the maximum charging power of the energy storage system. discharge P represents the discharge power of the energy storage system. max_discharge P is the maximum discharge power of the energy storage system. renewable P represents the total electricity generated from renewable energy sources. load P represents the total load demand within the microgrid. buy For the extra battery power purchased, V min V represents the minimum voltage limit for each node in the microgrid. node V represents the voltage at each node within the microgrid. max f represents the maximum voltage limit at each node within the microgrid. min f is the minimum limit value for the power frequency within the microgrid. grid f is the power frequency within the microgrid. max This represents the maximum limit value for the power frequency within the microgrid.

[0037] Furthermore, the relative position L(a) of the intermediate point is calculated based on the rules established by the data shape tree. i a1,a n This data shape tree is mapped to the charging power and discharging power of the energy storage system; the specific rules for establishing the data shape tree are as follows:

[0038] Curve A includes a series of sampling points (a1, a2, ..., a...). n Select a point a on curve A. i As the intermediate point, use L(a) i a1,a n ) represents point a i Relative to point a1 and point a n The position is point a. i Relative to point a1 and point a n Bookstein coordinates, point a i Relative to point a1 and point a n Position L(a) i a1,a n )=(a (1) ,a (2) ), where (a (1) ,a (2) The specific calculation formula is as follows:

[0039]

[0040] In the formula, For a k The two-dimensional plane coordinates, k = 1, 2, ..., n.

[0041] Furthermore, a BO-BiGRU-Attention hybrid neural network prediction model is used to correct the elasticity metric. The specific steps are as follows:

[0042] Step 51: Obtain historical data of distributed energy resources, time series data of charging and discharging power of energy storage systems, historical data of total power demand of microgrids, and historical data of power demand of each user equipment or system within the microgrid, and perform preprocessing.

[0043] Step 52: Construct the initial BiGRU-Attention model. Train the BiGRU-Attention model using the preprocessed data to obtain the initial BiGRU-Attention model.

[0044] Step 53: Hyperparameter optimization. The learning rate, training window size, number of hidden layer neurons, and number of BiGRU layer neurons of the initial BiGRU-Attention model are optimized using the Bayesian optimization algorithm. The BO-BiGRU-Attention model is obtained based on the optimized parameters of the initial BiGRU-Attention model.

[0045] Step 54: Obtain distributed energy data, energy storage system charging and discharging data, total power demand data of microgrid, and power demand data of each user device or system within the microgrid by predicting the BO-BiGRU-Attention model. Calculate the elasticity metric based on the predicted data to obtain the corrected elasticity metric.

[0046] The present invention also employs a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the above method when executing the computer program.

[0047] The present invention also employs a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps of the above-described method.

[0048] Beneficial effects: Compared with the prior art, the significant advantage of this invention is that it uses a flexible matching algorithm based on neural network correction for flexible networking. By using parameters such as the system's power demand, volatility, and priority as indicators, and stability and reliability as constraints, the flexible matching algorithm combines and optimizes the controllable distributed power sources, energy storage and other equipment in different modules for coordinated control. This achieves flexible energy consumption and minimizes electricity costs while flexibly networking the microgrid secondary system. Attached Figure Description

[0049] Figure 1 This is a flowchart of the flexible networking technology for the microgrid secondary system in this invention.

[0050] Figure 2 This is a flowchart of the elastic matching algorithm in this invention.

[0051] Figure 3 This is a structural diagram of the functional modular microgrid in this invention.

[0052] Figure 4 This is a structural diagram of the BO-BiGRU-Attention hybrid neural network prediction model in this invention. Detailed Implementation

[0053] Example 1

[0054] like Figure 1As shown in this embodiment, a microgrid secondary system networking based on flexible matching is implemented. Based on determining the types and control methods of each distributed power source or energy storage within the microgrid, the system is divided into five layers, including:

[0055] The information acquisition layer is used for real-time data acquisition from the system; it includes an operation monitoring data acquisition module and a microgrid comprehensive monitoring index calculation module. The operation monitoring data acquisition module is used to monitor the operation status of the microgrid in real time, specifically including: total power demand, total power supply, distributed generation comprehensive energy efficiency, energy storage device efficiency, microgrid energy storage device data, microgrid key equipment data, and microgrid communication equipment data. The microgrid comprehensive monitoring index calculation module is used to calculate comprehensive monitoring index data of the microgrid based on the real-time acquired operation monitoring parameter data, specifically including: microgrid overall power balance index, distributed energy response speed index, energy storage device power balance index, energy storage device charge / discharge efficiency stability index, key equipment operating status monitoring index, and network communication reliability index.

[0056] The calculation method for the overall power balance index of a microgrid is as follows: the total power demand and total power supply of the microgrid are read at regular intervals, and the reading times are numbered; based on this, a model formula for the overall power balance index of the microgrid is constructed.

[0057] The formula for the overall power balance index of a microgrid is shown in (1):

[0058]

[0059] In the formula, MGEL m0 Let be the microgrid's overall power balance index at the m0th read, where e is the natural constant, m0 is the read number (m0 = 1, 2, ..., m), and m is the total number of reads. ES m0 For the total power supply value read in the m0th reading, ED m0 Let τ be the total power demand value read in the m0th iteration, and τ be the correction factor for the microgrid's overall power balance index. By considering the total power demand and supply of the microgrid, this index can comprehensively assess the power balance of the microgrid, helping the system to have a more comprehensive understanding of the power status within the microgrid and providing basic data for optimizing power dispatch.

[0060] The specific calculation method of the distributed energy response speed index is as follows: extract the comprehensive energy efficiency of distributed generation and the efficiency of energy storage device of microgrid from the operation monitoring parameter data; monitor the solar irradiance and wind speed at the location of microgrid power station at regular intervals and number the monitoring time points; construct the distributed energy response speed index model formula; the specific distributed energy response speed index model formula is shown in (2):

[0061]

[0062] In the formula, DERSI k DEGE is the distributed energy response speed index at the k-th monitoring time point, where k is the number of the monitoring time point (k = 1, 2, ..., k0), and k0 is the total number of monitoring time points. k For the overall energy efficiency of distributed generation, IL k Let IL0 be the solar irradiance at the k-th monitoring time point, σ1 be the weighting coefficient of solar irradiance for the overall energy efficiency of distributed generation, and WS be the solar irradiance at the k-th monitoring time point. k Let λ be the wind speed at the k-th monitoring time point, WS0 be the unit value of wind speed, σ2 be the weighting coefficient of wind speed for the overall energy efficiency of distributed generation, λ1 be the weighting coefficient of the overall energy efficiency of distributed generation in the distributed energy response speed index, and ESR be the wind speed at the k-th monitoring time point. k λ² represents the efficiency of the energy storage device, and λ₂ is the weighting coefficient of the energy storage device efficiency in the distributed energy response speed index. By considering multiple factors such as the overall energy efficiency of distributed generation, solar irradiance, and wind speed, the distributed energy response speed index can more comprehensively assess the response speed of distributed energy within a microgrid to changes in the external environment. This helps to fully understand the impact of various factors on energy production within the microgrid. Introducing solar irradiance and wind speed, and adjusting them through weighting coefficients, can better balance the impact of natural environmental factors in the microgrid's location on the overall energy efficiency of distributed generation. This helps the index more accurately reflect the meteorological and geographical conditions of the area where the microgrid is located.

[0063] The specific calculation method for the energy balance index of the energy storage device is as follows: extract the data of the microgrid energy storage device from the operation monitoring parameter data, and number the microgrid energy storage device; monitor the charging efficiency, discharging efficiency, available power and battery capacity of the microgrid energy storage device at regular intervals; construct the energy balance index model formula of the energy storage device; the specific energy balance index model formula of the energy storage device is shown in (3):

[0064]

[0065] In the formula, ESSEI k Let CE be the energy balance index of the energy storage device at the k-th monitoring time point, where i is the number of the microgrid energy storage device, i = 1, 2, ..., i0, and i0 is the total number of microgrid energy storage devices. k i DE k i AE k i BE ki Let the charging efficiency, discharging efficiency, available power, and battery capacity of the i-th microgrid energy storage device be defined at the k-th monitoring time point. These represent the weighting percentages of charging efficiency, discharging efficiency, available capacity, and battery capacity in the energy balance index of an energy storage device. This is a correction factor for the energy balance index of energy storage devices. Introducing factors such as charging efficiency, discharging efficiency, available power, and battery capacity allows for a more comprehensive and integrated consideration of the state of energy storage devices within the microgrid. This helps to comprehensively reflect various aspects of the charging, discharging, and energy storage processes in the energy balance index, improving the index's comprehensiveness and accuracy. Considering available power and battery capacity helps to determine the energy storage status and energy storage capacity of the energy storage devices. This provides more specific reference information for microgrids when coping with demand peaks and grid fluctuations.

[0066] The specific calculation method for the stability index of the charging and discharging efficiency of energy storage equipment is as follows: at certain intervals, monitor the charging and discharging efficiency, temperature, and ambient temperature of the microgrid energy storage device; construct the model formula for the stability index of the charging and discharging efficiency of energy storage equipment; the specific model formula for the stability index of the charging and discharging efficiency of energy storage equipment is shown in (4):

[0067]

[0068] In the formula, ESESI k CDE is the stability index of the charge and discharge efficiency of the energy storage device at the k-th monitoring time point. k For the charge / discharge efficiency at the k-th monitoring time point, ESST k i Let μ1 be the temperature of the i-th microgrid energy storage device at the k-th monitoring time point, and μ1 be the weighting ratio of the energy storage device temperature in the energy storage device's charge-discharge efficiency stability index. k i Let be the ambient temperature of the i-th microgrid energy storage device at the k-th monitoring time point, and μ2 be the weighting ratio of the ambient temperature in the energy storage device's charge-discharge efficiency stability index. This is a correction factor for the stability index of the charge / discharge efficiency of energy storage devices. Factors such as charge / discharge efficiency, energy storage device temperature, and ambient temperature are incorporated to comprehensively consider the efficiency and temperature variations of the energy storage device during operation. This helps to more comprehensively and accurately evaluate the performance of energy storage devices under different conditions. The integrated assessment of charge / discharge efficiency, energy storage device temperature, and ambient temperature helps to fully understand the stability of energy storage devices under different conditions. This is of great significance for microgrid managers to formulate operational strategies and prevent potential problems.

[0069] The optimized scheduling layer employs appropriate optimized scheduling strategies to achieve resource scheduling and demand management during the network construction process, ensuring system stability and security while realizing efficient operation and optimized energy utilization of the microgrid. It analyzes and calculates collected system data through elastic matching to obtain optimized scheduling instructions; aiming at the lowest total cost, it obtains matching combinations of elastic metrics under constraints, and formulates optimized scheduling instructions based on the obtained matching combinations or optimized and corrected matching combinations of elastic metrics; using an elastic matching algorithm, with parameters such as system power demand, volatility, and priority as indicators, and stability and reliability as constraints, it combines and coordinates controllable distributed power sources, energy storage, and interconnection controllers in different modules for optimized controllability.

[0070] The module control layer is used to execute optimized scheduling instructions; it includes various functional modules, such as... Figure 2 As shown, the functional modular microgrid network includes an isolation transformer 7, an AC bus 8, a DC bus 9, an interconnection module (6), a PCC module (1), an AC microgrid regulation module (2), an AC microgrid non-regulatory module (3), a DC microgrid regulation module (4), and a DC microgrid non-regulatory module (5).

[0071] The PCC module (1) provides an external interface for the AC / DC hybrid microgrid. It is connected to the AC bus 8 on the right and to the main grid through the isolation transformer 7 on the other side. The internal equipment of the module is a three-phase AC circuit breaker. The power exchange between the main grid and the microgrid is limited by the power limiting control method in this module.

[0072] The AC microgrid regulation module (2) comprises four parts: one or more controllable distributed power sources 10, a distributed power source grid-connection device 11, a power-type energy storage system 12, and an energy storage system grid-connection device 13. The controllable distributed power source 10 is connected to the AC bus 8 through the distributed power source grid-connection device 11. The grid-connected inverter included in the distributed power source grid-connection device 11 employs active power-AC bus frequency (Pf) droop control and reactive power-AC bus voltage (Q-Uac) droop control. The power-type energy storage system 13 is connected to the AC bus 8 through the energy storage system grid-connection device 13. The grid-connected inverter included in the energy storage system grid-connection device 13 employs PQ control.

[0073] The AC microgrid non-adjustable module (3) consists of three parts: one or more uncontrollable distributed power sources 14, a distributed power source grid connection device 15, and an AC load 16. The uncontrollable distributed power source 14 is connected to the AC bus 8 through the distributed power source grid connection device 15. The AC load 16 is directly connected to the AC bus 8. The AC bus 8, the PCC module (1), the AC microgrid regulation module (2), and the AC microgrid non-adjustable module (3) constitute the AC microgrid system.

[0074] The DC microgrid regulation module (4) consists of two parts: an energy storage system 17 and a bidirectional DC / DC chopper 18. The energy storage system 17 is connected to the DC bus 9 through the grid-connected bidirectional DC / DC chopper 18. The grid-connected bidirectional DC / DC chopper is controlled by a constant DC voltage.

[0075] The DC microgrid non-adjustable module (5) comprises three parts: one or more uncontrollable distributed power sources 19, a distributed power source grid connection device 20, and a DC load 21. The uncontrollable distributed power source 19 is connected to the DC bus 9 through the distributed power source grid connection device 20. The DC load 21 is directly connected to the DC bus 9. The DC micro bus 9, the DC microgrid regulating module (4), and the DC microgrid non-adjustable module (5) constitute the DC microgrid system.

[0076] The interconnection module (6) is connected to the AC bus 8 and the DC bus 9 on both sides respectively. The interconnection converter inside the module can realize bidirectional power exchange. The interconnection converter adopts active power-DC bus voltage (P-Udc) droop control. When the error-free control of DC bus voltage Udc by the DC microgrid regulation module (4) fails, it can automatically participate in the regulation of DC bus voltage Uae and establish the connection between AC microgrid and DC microgrid.

[0077] Modular structures for AC / DC microgrids can provide a reference for microgrid networking methods. Because different modules play different roles in the overall microgrid, different types of distributed power sources and energy storage systems are required for each module. For AC microgrid regulation modules, stable and easily controllable distributed power sources are selected. These distributed power sources should possess stable output and ease of control, such as diesel generators, internal combustion engines, wind farms equipped with energy storage, or photovoltaic power stations. The energy storage system in this module is mainly responsible for rapidly responding to and restoring system stability during significant disturbances; therefore, power-type distributed energy storage systems such as supercapacitor energy storage systems should be selected. DC microgrid regulation modules are crucial for maintaining the stability of the DC microgrid bus voltage and are responsible for smoothing active power fluctuations throughout the DC microgrid. They need to provide power support for a longer period; therefore, energy-type distributed energy storage systems such as battery energy storage systems should be selected. If losses from multi-stage power conversion are not considered, stable and easily controllable distributed power sources can be used instead of energy storage systems to construct the DC microgrid regulation module.

[0078] To achieve multi-mode operation of the modular AC / DC hybrid microgrid structure based on functional differences, the module control layer employs distributed control, responsible for the internal operation control of each module. Different modules rationally allocate different control tasks according to their functional differences. When a module is detected as failing to meet its control task, the operation between modules is coordinated. Based on the different external characteristics of each module, they are further equivalentized. The PCC module is equivalent to a variable power source or load relative to the microgrid. The AC microgrid regulating module is equivalent to a voltage source with varying voltage and frequency. The DC microgrid regulating module is equivalent to a fixed voltage source in the DC microgrid automatic control mode and a fixed output power source or load in the interconnected module control mode. The droop control in the interconnected module can achieve natural switching between inversion and rectification based on the DC microgrid bus voltage. The non-adjustable modules of the DC microgrid and AC microgrid are equivalent to fixed power sources or loads. The microgrid layer control employs centralized control, responsible for coordinating the operation of each module after judging its operating parameters. The PCC module needs to determine whether the interactive power exceeds the limit; the AC regulation module needs to determine whether the AC bus frequency f and voltage Uac meet the operating requirements; the DC microgrid regulation module and interconnection module determine whether the DC bus voltage Uae meets the operating requirements. According to the above equivalent process, the microgrid layer does not need to collect and process all information such as all loads, energy storage, and distributed power sources at the same time. It only needs to determine some operating parameters based on the operating characteristics of different modules to coordinate the operation between modules.

[0079] The information feedback layer is used to provide feedback on the system state after executing optimized scheduling instructions; a Kalman filter is used for recursive filtering, which ensures accurate estimation of the system state even in the presence of noise.

[0080] The parameter optimization layer is used to optimize and correct the elasticity measure in elasticity matching based on the feedback system state through a BO-BiGRU-Attention hybrid neural network prediction model.

[0081] Example 2

[0082] The control method for microgrid secondary system networking based on flexible matching in this embodiment includes the following steps:

[0083] Step 1: Real-time collection of system data; specifically including: total power demand, total power supply, comprehensive energy efficiency of distributed generation, efficiency of energy storage devices, data of microgrid energy storage devices, data of key microgrid equipment, data of microgrid communication equipment, power balance index of the entire microgrid network, response speed index of distributed energy, power balance index of energy storage devices, stability index of charging and discharging efficiency of energy storage devices, working status monitoring index of key equipment, and network communication reliability index.

[0084] Step 2: Based on the collected system data, analyze and calculate using elastic matching to obtain optimized scheduling instructions; aiming at optimal total cost, system stability, and response speed, obtain the matching combination of microgrid power distribution under constraints, and formulate optimized scheduling instructions based on the obtained matching combination; the specific steps are as follows:

[0085] Step A, Data Shape Tree Construction:

[0086] Using an open and ordered radial curve, A represents a series of sampling points (a1, a2, ..., a...). n The curve is composed of points A and A, and a point a is selected on A. i As an intermediate point, i is generally taken as (1+n) / 2. Using L(a i a1,a n ) represents a i relative to point a1 and point a n The location of point a in this article is... i Relative to point a1 and point a n The Bookstein coordinates are obtained by mapping the two initial and final sampling points to fixed positions to determine their coordinates in a coordinate system. Since the positions of the initial and final sampling points are already determined, the Bookstein coordinates of only one intermediate sampling point are sufficient to accurately represent the relative positions of the three points. Point a i Relative to point a1 and point a n Location

[0087] L(a i a1,a n )=(a (1) ,a (2) ), where (a (1) ,a (2) The specific calculation formula can be expressed by the following formula:

[0088]

[0089] In the formula, For a k Two-dimensional plane coordinates of (k = 1, 2, ..., n).

[0090] Step B, Definition of elasticity measure:

[0091] Microgrid resources: Solar power generation P solar Wind power capacity P wind The energy storage system has a charge / discharge power P storage ;

[0092] Microgrid demand: Total electricity demand P load And the electricity demand P of each user, device or system userP device P system ;

[0093] Supply elasticity E s Supply elasticity measures the responsiveness of distributed energy resources to changes in environmental conditions (such as solar radiation and wind speed), i.e., changes in electricity supply caused by changes in environmental conditions. The specific formula is shown below:

[0094]

[0095] Where ΔG is the change in power generation capacity, G is the initial power generation output, ΔE is the change in environmental conditions (such as solar radiation intensity, wind speed, etc.), and E is the initial environmental conditions.

[0096] Load balancing elasticity E L Load balance resilience measures a system's ability to maintain balance under load changes, i.e., the change in system state caused by load changes. The specific calculation formula is as follows:

[0097]

[0098] Where ΔB is the change in the system equilibrium state (such as frequency deviation, voltage deviation, etc.), B is the initial equilibrium state (such as standard frequency, standard voltage, etc.), ΔL is the load change, and L is the initial load.

[0099] Demand Response Elasticity E d Demand response elasticity measures how well users respond to changes in electricity prices, i.e., the change in electricity demand caused by changes in electricity prices. The specific formula is shown below:

[0100]

[0101] Where ΔD is the change in electricity demand, D is the initial electricity demand, ΔC is the change in electricity price, and C is the initial electricity price.

[0102] Step C, Matching Strategy Design:

[0103] Step C1, calculate the total renewable energy generation P renewable :

[0104] P renewable =P solar +P wind (9)

[0105] Step C2: Calculate the total load demand P within the microgrid. load :

[0106] P load =P user +P device +P system(10)

[0107] Step C3: Determine whether the renewable energy generation is sufficient to meet the total load demand.

[0108] If P renewable ≥P load In this way, renewable energy can meet the load demand within the microgrid without the need to purchase additional electricity from the main grid;

[0109] If P renewable <P load In this case, additional electrical energy needs to be purchased from the main grid to meet the load demand within the microgrid.

[0110] Step C4, calculate the cost C corresponding to purchasing additional electricity. buy :

[0111] C buy =(P renewable -P load )×R grid (11)

[0112] In the formula, R grid This represents the cost of purchasing electricity from the main power grid.

[0113] Step C5: Set the charging power P of the energy storage system. charge With discharge power P discharge :

[0114] P charge =α×(P renewable -P load (12)

[0115] P discharge = (1-α)×(P) renewable -P load (13)

[0116] In the formula, α is the ratio of charging and discharging power.

[0117] Step C6: Calculate the relative position L(a) of the intermediate point based on the rules established by the data shape tree. i |a1,a n This is mapped to the charging and discharging power of the energy storage system.

[0118] Specifically, taking multiple power points in time series data as an example, suppose the following power data exists:

[0119] P1 = 100kW, t1 = 1 hour; P2 = 150kW, t2 = 2 hours; P3 = 200kW, t3 = 3 hours. These data points can be mapped on a conventional two-dimensional plane as (1, 100), (2, 150), (3, 200).

[0120] Fix the two endpoints at specific positions in the coordinate system: assume that the power point P1 at time t1 is mapped to (-0.5, 0), and the power point P3 at time t3 is mapped to (0.5, 0);

[0121] Calculate the relative position of the midpoint P2: Determine the position of P2 in this coordinate system through standardization and geometric transformation:

[0122]

[0123] Therefore, the position of P2 in the Bookstein coordinate system is (0.5, 0).

[0124] This method allows power data from a series of time points to be mapped onto a standardized planar coordinate system, thereby better capturing and analyzing the charging and discharging behavior of energy storage systems. This helps to optimize power dispatch strategies, balance system supply and demand, and improve system stability and efficiency.

[0125] Step C8: Calculate the cost C corresponding to charging and discharging the energy storage system. charge and C discharge :

[0126] C charge =P charge ×R charge (16)

[0127] C discharge =P discharge ×R discharge (17)

[0128] In the formula, R charge For the charging cost of energy storage systems, R discharge This refers to the discharge cost of the energy storage system.

[0129] Step C9: Under the constraints of power balance, voltage stability, and frequency stability of the power grid, solve the problem of optimizing the total cost, system stability, and response speed of the microgrid secondary system based on the conjugate gradient algorithm.

[0130]

[0131] In the formula, w c As a cost weight, w d For demand response elasticity weights, w L For supply elasticity weights, w s E is the load balancing elastic weight. d For demand response elasticity, E L For load balancing flexibility, E s For supply elasticity, C buyThe cost of purchasing additional electricity, C charge The cost of charging an energy storage system, C discharge P represents the cost corresponding to the discharge of the energy storage system, α is the ratio of the charging power to the discharging power of the energy storage system, and P is the cost of discharging the energy storage system. charge The charging power of the energy storage system, P max_charge This represents the maximum charging power of the energy storage system.

[0132] P discharge P represents the discharge power of the energy storage system. max_discharge P is the maximum discharge power of the energy storage system. renewable P represents the total electricity generated from renewable energy sources. load P represents the total load demand within the microgrid. buy For the extra battery power purchased, V min V represents the minimum voltage limit for each node in the microgrid. node V represents the voltage at each node within the microgrid. max f represents the maximum voltage limit at each node within the microgrid. min f is the minimum limit value for the power frequency within the microgrid. grid f is the power frequency within the microgrid. max This represents the maximum limit value for the power frequency within the microgrid.

[0133] Based on the above optimization results, the optimal total cost can be obtained by re-iterating the purchase of additional electrical energy, charging power, and discharging power.

[0134] Step 3: Execute optimized scheduling instructions; execute optimized scheduling instructions through functional modular microgrid networking;

[0135] Step 4: Feedback on the system state after executing the optimized scheduling instructions; use a Kalman filter for recursive filtering to ensure accurate estimation of the system state even in the presence of noise.

[0136] The specific steps of recursive filtering using the Kalman filter are as follows:

[0137] Step A: Based on the state estimate of the previous moment and the control input of the system, predict the state of the system at the next moment. This prediction process will estimate the expected value and covariance matrix of the system state, representing the uncertainty of the state estimate.

[0138] Step B: Calculate the Kalman gain by comparing the difference between the predicted value and the actual observed value, and use this gain to correct the predicted value.

[0139] Step 5: Based on the feedback system status, optimize and correct the elasticity measure in elastic matching using a neural network prediction model; the specific steps are as follows:

[0140] Step A: Obtain historical data on solar and wind power capacity, historical time-series data on the charging and discharging power of energy storage systems, historical data on the total power demand of the microgrid, and historical data on the power demand of each user device or system within the microgrid. Perform data preprocessing: To avoid excessive numerical differences, use the Min-Max Normalization method to normalize the data, and divide the dataset into training and testing sets in a 9:1 ratio. The specific normalization formula is as follows:

[0141]

[0142] In the formula: x normalized This represents the normalized data, where x represents the original data. max and x min These represent extremely large and extremely small data, respectively.

[0143] Step B, Constructing the Initial Model: By using a bidirectional GRU layer, both historical and future capacity and load data are considered simultaneously, thus defining the elasticity metric more comprehensively and improving the accuracy of elasticity matching. Specifically, the output of the BiGRU neural network is shown in equations (20), (21), and (22).

[0144]

[0145] In the formula: This represents the state of the forward hidden layer at time t; w represents the state of the reverse hidden layer at time t; t The weights of the forward hidden layer state at time t are represented by v. t b represents the weights of the reverse hidden layer state at time t; t This represents the bias of the hidden layer state at time t.

[0146] By combining the Attention mechanism, the attention score is converted into attention weights using the softmax function, and the most important features are extracted to make the model more robust. The specific calculation formulas are shown in (23), (24), and (25):

[0147] s(x i ,q)=V T tanh(Wx i +b) (23)

[0148]

[0149] In the formula: a i and x iLet x and q represent the attention score and weight of capacity or load, respectively. attention(x,q) represents the output of the attention layer at time i, and W and b represent the learned weight matrix and bias, respectively.

[0150] Step C, Hyperparameter Optimization: The learning rate, training window size, number of hidden layer neurons, and number of BiGRU layer neurons of the BiGRU-Attention model are optimized using the Bayesian optimization algorithm.

[0151] Furthermore, the Bayesian optimization method maximizes or minimizes a black-box objective function. Based on Bayes' theorem and the Gaussian process equiprobability model, it continuously selects the next sampling point in the search space and estimates the optimal value of the function based on the existing sampling data, gradually converging to the global optimum. The specific expression is shown in (26):

[0152]

[0153] In the formula: f(h) represents the prior distribution model, h * Let f(h) represent the optimal parameter values ​​in the constraint domain, and H represent the candidate set.

[0154] Step D, Prediction and Parameter Correction: The BiGRU-Attention model is trained using the training set, and the network model parameters are adjusted and optimized. The BO-BiGRU-Attention model is used to make predictions using the test set. The predicted data include distributed energy data, energy storage system charging and discharging data, total power demand data of the microgrid, and power demand data of each user device or system within the microgrid. The predicted data is then normalized and corrected using elasticity metrics.

[0155] Step 6: Formulate optimized scheduling instructions based on the optimized and corrected elasticity metric matching combination, and return to execute step 2.

[0156] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application 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.

[0157] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, 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 apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0158] 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 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0159] 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 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0160] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A microgrid secondary system networking method based on flexible matching, characterized in that, include: The information acquisition layer is used to collect system data in real time. The optimized scheduling layer is used to obtain a matching combination of microgrid power distribution under constraints, based on the collected system data and with the goal of minimizing total cost and elasticity metrics. Based on the obtained matching combination, optimized scheduling instructions are formulated. Microgrid power distribution includes the purchase of additional electricity, the charging power of energy storage systems, and the discharging power of energy storage systems. The elasticity metrics include supply elasticity, which reflects the response of distributed energy resources to changes in environmental conditions; load balance elasticity, which reflects the system's ability to balance loads under changing load conditions; and demand response elasticity, which reflects the degree of user response to changes in electricity prices. The module control layer is used to execute optimized scheduling instructions; The information feedback layer is used to provide feedback on the system status after executing optimized scheduling instructions; The parameter optimization layer is used to correct the elasticity measure based on the feedback system state through a neural network prediction model; The parameter optimization layer uses a BO-BiGRU-Attention hybrid neural network prediction model to correct the elasticity metric. The specific steps are as follows: Step 51: Obtain historical data of distributed energy resources, time series data of charging and discharging power of energy storage systems, historical data of total power demand of microgrids, and historical data of power demand of each user equipment or system within the microgrid, and perform preprocessing. Step 52: Construct the initial BiGRU-Attention model. Train the BiGRU-Attention model using the preprocessed data to obtain the initial BiGRU-Attention model. Step 53: Hyperparameter optimization. The learning rate, training window size, number of hidden layer neurons, and number of BiGRU layer neurons of the initial BiGRU-Attention model are optimized using the Bayesian optimization algorithm. The BO-BiGRU-Attention model is obtained based on the optimized parameters of the initial BiGRU-Attention model. Step 54: Obtain distributed energy data, energy storage system charging and discharging data, total power demand data of microgrid, and power demand data of each user device or system within the microgrid by predicting the BO-BiGRU-Attention model. Calculate the elasticity metric based on the predicted data to obtain the corrected elasticity metric.

2. The microgrid secondary system networking based on flexible matching according to claim 1, characterized in that, The information acquisition layer includes an operation monitoring data acquisition module and a microgrid comprehensive monitoring index calculation module; The operation monitoring data acquisition module is used to collect operation monitoring parameters of the microgrid's operating status in real time, specifically including: total power demand, total power supply, comprehensive energy efficiency of distributed generation, energy storage device efficiency, microgrid energy storage device data, microgrid key equipment data, and microgrid communication equipment data. The microgrid comprehensive monitoring index calculation module is used to calculate the microgrid comprehensive monitoring index data based on the real-time collected operation monitoring parameters. Specifically, it includes: microgrid overall power balance index, distributed energy response speed index, energy storage device power balance index, energy storage device charge and discharge efficiency stability index, key equipment working status monitoring index, and network communication reliability index.

3. The microgrid secondary system networking based on flexible matching according to claim 1, characterized in that, The module control layer includes: a PCC module, a microgrid AC regulation module, a microgrid AC non-regulatory module, a microgrid DC regulation module, a microgrid DC non-regulatory module, and an interconnection module; The PCC module switches between on-grid and off-grid operation of the microgrid using a limited switching power control method; The microgrid AC regulation module includes one or more distributed power sources, energy storage systems, and grid connection devices; The microgrid's non-adjustable AC module includes AC loads, non-dispatchable distributed power sources, and networking devices. The microgrid DC regulation module includes an energy storage system with regulation capability and a bidirectional DC / DC chopper for controlling the DC bus voltage of the microgrid. The microgrid's non-adjustable DC module includes DC loads, non-dispatchable distributed power sources, and grid-connected devices; The interconnection module includes an interconnection converter, with its two ends connected to the microgrid AC bus and the microgrid DC bus, respectively.

4. A control method for microgrid secondary system networking based on elastic matching, characterized in that, Includes the following steps: Step 1: Collect system data in real time; Step 2: Based on the collected system data, with the goal of minimizing total cost and elasticity metrics, a matching combination of microgrid power distribution is obtained under constraints. Optimized dispatch instructions are then formulated based on the obtained matching combination. Microgrid power distribution includes the purchase of additional electricity, the charging power of the energy storage system, and the discharging power of the energy storage system. The elasticity metrics include supply elasticity (reflecting the distributed energy source's responsiveness to changes in environmental conditions), load balance elasticity (reflecting the system's balancing ability under load changes), and demand response elasticity (reflecting the user's responsiveness to changes in electricity prices). Step 3: Execute the optimized scheduling instruction; Step 4: Feedback on the system status after executing the optimized scheduling instructions; Step 5: Based on the feedback of the system status, the elasticity measure in elastic matching is corrected using a neural network prediction model; The elasticity metric is optimized and corrected using a BO-BiGRU-Attention hybrid neural network prediction model. The specific steps are as follows: Step 51: Obtain historical data of distributed energy resources, time series data of charging and discharging power of energy storage systems, historical data of total power demand of microgrids, and historical data of power demand of each user equipment or system within the microgrid, and perform preprocessing. Step 52: Construct the initial BiGRU-Attention model. Train the BiGRU-Attention model using the preprocessed data to obtain the initial BiGRU-Attention model. Step 53: Hyperparameter optimization. The learning rate, training window size, number of hidden layer neurons, and number of BiGRU layer neurons of the initial BiGRU-Attention model are optimized using the Bayesian optimization algorithm. The BO-BiGRU-Attention model is obtained based on the optimized parameters of the initial BiGRU-Attention model. Step 54: Obtain distributed energy data, energy storage system charging and discharging data, total power demand data of microgrid, and power demand data of each user equipment or system within the microgrid by predicting the BO-BiGRU-Attention model. Calculate the elasticity metric based on the predicted data to obtain the corrected elasticity metric. Step 6: Formulate optimized scheduling instructions based on the matching combination of the corrected elasticity metrics, and return to execute step 3.

5. The control method according to claim 4, characterized in that, The collected system data includes total power demand, total power supply, comprehensive energy efficiency of distributed generation, energy storage device efficiency, microgrid energy storage device data, microgrid key equipment data, microgrid communication equipment data, microgrid overall power balance index, distributed energy response speed index, energy storage device power balance index, energy storage device charge and discharge efficiency stability index, key equipment working status monitoring index, and network communication reliability index.

6. The control method according to claim 4, characterized in that, With power balance, voltage stability, and frequency stability of the microgrid as constraints, this paper uses the conjugate gradient algorithm to solve the optimization problem that minimizes the total cost and resilience of the microgrid secondary system. ； In the formula, As a cost weight, For demand response elasticity weight, To supply elastic weights, For load balancing elastic weight, For demand response elasticity, For load balancing flexibility, For supply elasticity, The cost of purchasing additional electricity. The cost of charging an energy storage system The cost corresponding to the discharge of the energy storage system. This refers to the ratio of the charging power to the discharging power of the energy storage system. The charging power for the energy storage system, This represents the maximum charging power of the energy storage system. The discharge power of the energy storage system. This represents the maximum discharge power of the energy storage system. Total renewable energy generation, For the total load demand within the microgrid, For the extra battery power purchased, These are the minimum voltage limits for each node within the microgrid. This represents the voltage at each node within the microgrid. This represents the maximum voltage limit for each node within the microgrid. This represents the minimum limit for the power frequency within the microgrid. The power frequency within the microgrid. This represents the maximum limit value for the power frequency within the microgrid.

7. The control method according to claim 4, characterized in that, Calculate the relative position of the intermediate point based on the rules established by the data shape tree. This data shape tree is mapped to the charging power and discharging power of the energy storage system; the specific rules for establishing the data shape tree are as follows: Curve A includes a series of sampling points Select a point on curve A As the midpoint, use Point Relative to point and points The position is point Relative to point and points The Bookstein coordinates of the point Relative to point and points Location ,in, The specific calculation formula is as follows: ； In the formula, , for The two-dimensional plane coordinates, k=1, 2, …,n.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 4 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 4 to 7.

Images