A data-driven based distributed energy storage adaptive sliding mode control method

Abstract

This invention relates to a data-driven adaptive sliding mode control method for distributed energy storage, belonging to the field of energy storage control technology. Specifically, it is a data-driven adaptive sliding mode control method for distributed energy storage used to solve the problem of voltage fluctuation exceeding limits in active distribution networks under high-proportion renewable energy access. In the absence of accurate network parameters and historical operating data, this method calculates the voltage sensitivity of distributed energy storage to each node by real-time acquisition of node voltage and photovoltaic power output data, and performs clustering based on sensitivity similarity. Based on the partitioning results, a voltage control model is established with the objectives of minimizing voltage deviation and energy storage power variation. Combining gradient descent and adaptive sliding mode control, the energy storage charging and discharging strategy is solved online. This invention features fast response, strong adaptability, and high robustness, effectively suppressing voltage fluctuations caused by sudden changes in photovoltaic power output and improving the operational stability of the distribution network.

Description

Technical Field

[0001] This invention belongs to the field of energy storage control technology, and in particular to a data-driven adaptive sliding mode control method for distributed energy storage. Background Technology

[0002] With the high proportion of distributed renewable energy integration, the operating conditions of distribution networks are becoming increasingly complex, system volatility is significantly enhanced, and power flow distribution is more variable, leading to frequent problems such as power quality degradation, equipment overload, voltage exceeding limits, and power backflow. To address these challenges, power electronics technology is constantly advancing, and controllable devices based on power electronic devices, such as distributed power inverters and energy storage converters, are widely used in distribution networks. These devices have rapid response and continuous adjustment capabilities, effectively smoothing out random fluctuations in renewable energy output and enabling real-time, precise control of key parameters such as active / reactive power and voltage, becoming an important means to improve the stability and flexibility of distribution network operation.

[0003] Traditional distribution network control methods primarily rely on precise physical models, using system dynamic equations for controller design and parameter tuning, such as state-feedback-based linear quadratic regulation or model predictive control. However, in actual operation, traditional modeling methods face significant limitations due to the difficulty in accurately obtaining line parameters, frequent changes in load characteristics, and the nonlinearity and uncertainty brought about by the integration of numerous distributed energy sources. Especially in scenarios with high renewable energy penetration, system observability and controllability decrease, and a single static model struggles to accurately describe the dynamic behavior of the power grid, leading to degraded control performance or even control failure.

[0004] In recent years, with the rapid development of intelligent measurement terminals and communication networks, distribution networks have generated massive amounts of real-time operational data, covering information such as node voltage, current, power, load characteristics, and environmental factors. This provides the foundation for the application of data-driven control methods. Compared with traditional methods that rely on physical models, data-driven control does not require a complete system mathematical model and has advantages such as low modeling cost, strong adaptability, and ease of online updates, gradually becoming a hot research area in the operation and control of flexible distribution networks.

[0005] However, most existing data-driven control methods rely on offline training based on historical operating data. For example, they utilize machine learning algorithms to construct input-output mapping relationships, thereby achieving adaptive adjustment of voltage, frequency, or power. These methods perform well when the system structure is stable and the operating mode is relatively fixed. However, when dynamic changes occur in the distribution network, such as topology reconfiguration, fault isolation, or economic reconfiguration, the previously trained models may lose their effectiveness, requiring data re-collection and long-term training, which is difficult to meet real-time requirements. Furthermore, for large-scale, heterogeneous distributed energy storage systems, their charging and discharging behavior is significantly affected by multi-source disturbances, making it difficult to capture the system's transient response characteristics using only historical data.

[0006] Therefore, data-driven control methods based on real-time measurement data have emerged. These methods directly utilize the current system input and output observations for online analysis and decision-making, without relying on historical data training or pre-establishing a system mathematical model. They offer advantages such as fast response speed, strong online adaptability, and high robustness. In particular, in distributed energy storage systems, if a real-time measurement-driven control strategy is adopted, it can fully utilize real-time information such as voltage, current, and power of the power grid in the area where the energy storage device is located. Combined with the strong robustness and finite-time convergence characteristics of sliding mode control theory, an adaptive sliding mode control framework can be constructed. This method can achieve rapid and precise control of the energy storage system under unknown system parameters and external disturbances by adjusting the sliding surface parameters online, effectively suppressing voltage fluctuations and improving system stability. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and propose a data-driven distributed energy storage adaptive sliding mode control method, which can maintain the stable operation of the power grid, quickly and accurately provide the optimal energy storage control strategy, greatly improve the work efficiency of dispatchers, and ensure the safe and stable operation of the power grid.

[0008] The technical problem solved by this invention is achieved through the following technical solution: A data-driven adaptive sliding mode control method for distributed energy storage includes the following steps: Step 1: Based on the selected active power distribution network, set the sampling frequency of the distributed energy storage controller. and control the total duration Input the parameters of the distributed energy storage controller; Step 2: Adjust the charging and discharging power of the distributed energy storage, obtain the node voltage measurement value, calculate the voltage sensitivity of the distributed energy storage to each node, and perform similar clustering partitioning of node voltage sensitivity. Step 3: Collect real-time measurement data of each area of ​​the active power distribution network. The real-time measurement data includes: regional node voltage measurement value and photovoltaic power output measurement value. Calculate the sliding mode adaptive adjustment coefficient based on the photovoltaic power output measurement value. Step 4: Based on the distributed energy storage parameters provided in Step 1 and the real-time measurement data of each region in Step 3, establish a data-driven active distribution network voltage control model with the goal of minimizing the deviation between the measured node voltage value and the node voltage reference value in each region and minimizing the change in the charging and discharging power of distributed energy storage. Step 5: Solve the model using the gradient descent method and add sliding mode control to obtain the distributed energy storage charging and discharging strategy. Send the control strategy to the distributed energy storage to obtain new node voltage measurements and photovoltaic power output measurements for each region.

[0009] Step 6: Obtain the current control duration and determine whether the current control duration is greater than the total control duration. If it is less, return to step 3; otherwise, end the process.

[0010] Furthermore, the parameters of the distributed energy storage controller in step 1 include: node voltage reference value. Weighting coefficient , , , Iterative correction coefficient Initial value of sliding mode adaptive adjustment coefficient for distributed energy storage Photovoltaic power change rate coefficient The timing begins when the distributed energy storage controller is connected to the active power distribution network, and is recorded as follows: .

[0011] Furthermore, the method for calculating the voltage sensitivity of each node in step 2 of the distributed energy storage is as follows: in, Distributed energy storage For nodes Voltage sensitivity; and They represent Time and Time Node The voltage measurement value; and They represent Time and Distributed energy storage at all times The charging and discharging power.

[0012] Furthermore, the specific implementation method of the node voltage sensitivity similarity clustering partitioning in step 2 is as follows: Construct the voltage sensitivity vector of each node: in, Represents a node The voltage sensitivity vector; Distributed energy storage For nodes The voltage sensitivity of the voltage; This indicates the quantity of distributed energy storage.

[0013] The node voltage sensitivity of the distributed energy storage access node is selected as the initial centroid, and then the Euclidean distance is calculated based on the voltage sensitivity vector of each node. Each node is assigned to the nearest regional centroid, and the new centroids of each region are updated: in, Represents a node and center of mass The Euclidean distance; Represents a node The voltage sensitivity vector; Indicates the center of mass , representing the voltage sensitivity vector; For the region ; Represents a node and center of mass The Euclidean distance; For the region The number of nodes is determined through a finite number of iterations until the centroid no longer changes, thus dividing the active distribution network into... Each region has a distributed energy storage controller that controls the region's voltage.

[0014] Furthermore, the sliding mode adaptive adjustment coefficient in step 3 is calculated as follows: in, express Real-time distributed energy storage sliding mode adaptive adjustment coefficient; This represents the initial value of the sliding mode adaptive adjustment coefficient for distributed energy storage; Indicates the rate of change coefficient of photovoltaic power; and express Time and The measured value of photovoltaic power output at any given time.

[0015] Furthermore, the data-driven active distribution network voltage control model in step 4 is as follows: in, express The estimated value of the node voltage at time point. express Voltage measurement at time node. This represents the charging and discharging power vector of distributed energy storage. and They represent Time and Real-time distributed energy storage charging and discharging power vector express The time-matter pseudo-Jacobi matrix, in this model, is used to reflect the sensitivity relationship between the control objective of the active distribution network and the output strategy of the controlled equipment: in, yes Voltage measurement at time node. It is estimated by the distributed energy storage controller. Node voltage at time, This represents the weighting coefficient. express The pseudo-Jacobi matrix at time. express The pseudo-Jacobi matrix at time.

[0016] Furthermore, the objective function for data-driven active distribution network voltage control in step 4 is: in, yes The vector of node voltage reference values ​​at time t; express The vector of node voltage measurements at time t; and They represent Time and Momentary distributed energy storage charging and discharging power vector; This represents the weighting coefficient.

[0017] Furthermore, the specific implementation method of step 5 is as follows: in, and They represent Time and Momentary distributed energy storage charging and discharging power vector; express The pseudo-Jacobi matrix at time; and They represent Time and The vector of node voltage reference values ​​at time t; and They represent Time and The vector of node voltage measurements at time t; This represents the weighting coefficient.

[0018] By employing the gradient descent method and simplifying the matrix inversion process, the pseudo-Jacobi matrix can be obtained. The method is as follows: in, and They represent Time and The pseudo-Jacobi matrix at time; express Change in voltage measurement at time node; express The change in the charging and discharging power vector of distributed energy storage at any given moment; and They represent Time and The vector of node voltage measurements at time t; and They represent Time and Momentary distributed energy storage charging and discharging power vector; This represents the weighting coefficient.

[0019] In specific active distribution network voltage control problems, to enhance the time-varying tracking effect of the control method, the pseudo-Jacobi matrix is ​​corrected. The estimation method is as follows: sign( ) sign( ) in, Represents the pseudo-Jacobi matrix Initial value; express The change in the charging and discharging power vector of distributed energy storage at any given moment; These are the iterative correction coefficients.

[0020] The advantages and positive effects of this invention are: The data-driven adaptive sliding mode control method for distributed energy storage of the present invention comprehensively considers the unknown parameters of active distribution network lines, the uncertainty of distributed power output, and the unknown historical data. It performs clustering and partitioning of similar nodes based on node voltage sensitivity, and constructs a data-driven voltage control model for active distribution networks. This effectively improves the voltage over-limit situation of high-proportion photovoltaic distribution networks and enhances the flexible and efficient operation level of active distribution systems. Attached Figure Description

[0021] Figure 1 This is a flowchart of the data-driven distributed energy storage adaptive sliding mode control method of the present invention; Figure 2 This is a diagram of the active power distribution network topology used in this invention; Figure 3 It is a graph showing the output of photovoltaic power and the output of the load. Figure 4 This is a graph showing the results of node clustering and partitioning. Figure 5 This is a voltage comparison diagram at node 25 within region 1; Figure 6This is a voltage comparison diagram at node 33 within region 2; Figure 7 This is a voltage comparison diagram at node 18 within region 3; Figure 8 It is a distributed energy storage charging and discharging power strategy. Detailed Implementation

[0022] The present invention will be further described in detail below with reference to the accompanying drawings.

[0023] A data-driven adaptive sliding mode control method for distributed energy storage, such as... Figure 1 As shown, it includes the following steps: Step 1: Based on the selected active power distribution network, set the sampling frequency of the distributed energy storage controller. and control the total duration Input the parameters of the distributed energy storage controller.

[0024] Set the sampling frequency of the distributed energy storage controller and control the total duration Input parameters to the distributed energy storage controller, specifically including: node voltage reference values. Weighting coefficient , , , Iterative correction coefficient Initial value of sliding mode adaptive adjustment coefficient for distributed energy storage Photovoltaic power change rate coefficient The timing begins when the distributed energy storage controller is connected to the active power distribution network, and is recorded as follows: .

[0025] Step 2: Adjust the charging and discharging power of the distributed energy storage, obtain the node voltage measurement value, calculate the voltage sensitivity of the distributed energy storage to each node, and perform similar clustering partitioning of node voltage sensitivity.

[0026] The calculation method for the voltage sensitivity of distributed energy storage to each node is as follows: in, Distributed energy storage For nodes Voltage sensitivity; and They represent Time and Time Node The voltage measurement value; and They represent Time and Distributed energy storage at all times The charging and discharging power.

[0027] The specific implementation method of node voltage sensitivity similarity clustering partitioning is as follows: Construct the voltage sensitivity vector of each node: in, Represents a node The voltage sensitivity vector; Distributed energy storage For nodes The voltage sensitivity of the voltage; This indicates the quantity of distributed energy storage.

[0028] The node voltage sensitivity of the distributed energy storage access node is selected as the initial centroid, and then the Euclidean distance is calculated based on the voltage sensitivity vector of each node. Each node is assigned to the nearest regional centroid, and the new centroids of each region are updated: in, Represents a node and center of mass The Euclidean distance; Represents a node The voltage sensitivity vector; Indicates the center of mass , representing the voltage sensitivity vector; For the region ; Represents a node and center of mass The Euclidean distance; For the region The number of nodes is determined through a finite number of iterations until the centroid no longer changes, thus dividing the active distribution network into... Each region has a distributed energy storage controller that controls the region's voltage.

[0029] Step 3: Collect real-time measurement data for each area of ​​the active power distribution network. The real-time measurement data includes: regional node voltage measurement values ​​and photovoltaic power output measurement values. Calculate the sliding mode adaptive adjustment coefficient based on the photovoltaic power output measurement values.

[0030] The sliding mode adaptive adjustment coefficient is calculated as follows: in, express Real-time distributed energy storage sliding mode adaptive adjustment coefficient; This represents the initial value of the sliding mode adaptive adjustment coefficient for distributed energy storage; Indicates the rate of change coefficient of photovoltaic power; and express Time and The measured value of photovoltaic power output at any given time.

[0031] Step 4: Based on the distributed energy storage parameters provided in Step 1 and the real-time measurement data of each region in Step 3, establish a data-driven active distribution network voltage control model with the goal of minimizing the deviation between the measured node voltage value and the node voltage reference value in each region and minimizing the change in the charging and discharging power of distributed energy storage.

[0032] The data-driven active distribution network voltage control model is as follows: in, express The estimated value of the node voltage at time point. express Voltage measurement at time node. This represents the charging and discharging power vector of distributed energy storage. and They represent Time and Real-time distributed energy storage charging and discharging power vector express The time-matter pseudo-Jacobi matrix, in this model, is used to reflect the sensitivity relationship between the control objective of the active distribution network and the output strategy of the controlled equipment: in, yes Voltage measurement at time node. It is estimated by the distributed energy storage controller. Node voltage at time, This represents the weighting coefficient. express The pseudo-Jacobi matrix at time. express The pseudo-Jacobi matrix at time.

[0033] Furthermore, the objective function for data-driven active distribution network voltage control in step 4 is: in, yes The vector of node voltage reference values ​​at time t; express The vector of node voltage measurements at time t; and They represent Time and Momentary distributed energy storage charging and discharging power vector; This represents the weighting coefficient.

[0034] Step 5: Solve the model using the gradient descent method and add sliding mode control to obtain the distributed energy storage charging and discharging strategy. Send the control strategy to the distributed energy storage to obtain new node voltage measurements and photovoltaic power output measurements for each region.

[0035] in, and They represent Time and Momentary distributed energy storage charging and discharging power vector; express The pseudo-Jacobi matrix at time; and They represent Time and The vector of node voltage reference values ​​at time t; and They represent Time and The vector of node voltage measurements at time t; This represents the weighting coefficient.

[0036] By employing the gradient descent method and simplifying the matrix inversion process, the pseudo-Jacobi matrix can be obtained. The method is as follows: in, and They represent Time and The pseudo-Jacobi matrix at time; express Change in voltage measurement at time node; express The change in the charging and discharging power vector of distributed energy storage at any given moment; and They represent Time and The vector of node voltage measurements at time t; and They represent Time and Momentary distributed energy storage charging and discharging power vector; This represents the weighting coefficient.

[0037] In specific active distribution network voltage control problems, to enhance the time-varying tracking effect of the control method, the pseudo-Jacobi matrix is ​​corrected. The estimation method is as follows: sign( ) sign( ) in, Represents the pseudo-Jacobi matrix Initial value; express The change in the charging and discharging power vector of distributed energy storage at any given moment; These are the iterative correction coefficients.

[0038] Step 6: Obtain the current control duration And determine the current control duration. Is it greater than the total control time? ,like If yes, proceed to step 3; otherwise, end the process.

[0039] Based on the aforementioned data-driven distributed energy storage adaptive sliding mode control method, the effectiveness of the invention was verified through testing.

[0040] In an embodiment of the present invention, the active power distribution network includes 33 nodes, and the topology is as follows: Figure 2 As shown, 10 photovoltaic (PV) systems are connected to nodes 4, 7, 10, 13, 17, 18, 21, 25, 30, and 33, with capacities of 200kWp, 400kWp, 400kWp, 500kWp, 800kWp, 400kWp, 200kWp, 200kWp, 500kWp, and 800kWp, respectively. Three distributed energy storage devices are connected to nodes 17, 25, and 33, with rated power and rated capacity of 1.5MW / 6MWh, 0.8MW / 3.2MWh, and 0.5MW / 2MWh, respectively. The PV and load output curves are shown below. Figure 3 As shown. The system voltage is 12.66kV, the base power is 1MVA, and the active and reactive loads are 3715kW and 2300kvar, respectively. Sampling frequency. Control the total duration =1h; The reference value for the node voltage in the active distribution network is set to 1.0pu. Weighting coefficient , 0、 , Iterative correction coefficients Initial value of sliding mode adaptive adjustment coefficient for distributed energy storage Photovoltaic power change rate coefficient A data-driven adaptive sliding mode control method for distributed energy storage is adopted for control. Through the above steps, the charging and discharging power strategy of the distributed energy storage can be obtained. To verify the effectiveness of this method, the following two control scenarios are compared for active distribution networks: Scenario 1: Without controlling distributed energy storage, the initial operating state of the active power distribution network is obtained; Scenario 2: Adopting a data-driven distributed energy storage adaptive sliding mode control method.

[0041] The computer hardware environment for performing the optimized calculations was an Intel(R) Core(TM) CPU i5-13500HX with a clock speed of 2.5GHz and 16GB of memory; the software environment was a Windows 11 operating system.

[0042] The computational topology used in the embodiments of this invention is as follows: Figure 2 As shown in the figure. The changes in photovoltaic and load output curves are as follows. Figure 3 As shown. The node clustering partitioning results are as follows. Figure 4 As shown. The voltage at node 25 within region 1 is as follows: Figure 5 As shown. The voltage at node 33 in region 2 is as follows: Figure 6 As shown. The voltage at node 18 within region 3 is as follows: Figure 7 As shown. The distributed energy storage charging and discharging power strategy is as follows: Figure 8 As shown.

[0043] Combination Figures 5 to 8 It can be seen that, compared with scenario one, scenario two can effectively control the charging and discharging power of distributed energy storage in each region. Through the strong robustness of adaptive sliding mode control, disturbances are suppressed, the impact of sudden changes in photovoltaic output on the control effect is reduced, and efficient regulation of active distribution network node voltage is achieved.

[0044] It should be emphasized that the embodiments described in this invention are illustrative rather than limiting. Therefore, this invention includes, but is not limited to, the embodiments described in the specific implementation. Any other implementations derived by those skilled in the art based on the technical solutions of this invention are also within the scope of protection of this invention.

Claims

1. A data-driven adaptive sliding mode control method for distributed energy storage, characterized in that: Includes the following steps: Step 1: Based on the selected active power distribution network, set the sampling frequency of the distributed energy storage controller. and control the total duration Input the parameters of the distributed energy storage controller; Step 2: Adjust the charging and discharging power of the distributed energy storage, obtain the node voltage measurement value, calculate the voltage sensitivity of the distributed energy storage to each node, and perform similar clustering partitioning of node voltage sensitivity. Step 3: Collect real-time measurement data of each area of ​​the active power distribution network. The real-time measurement data includes: regional node voltage measurement value and photovoltaic power output measurement value. Calculate the sliding mode adaptive adjustment coefficient based on the photovoltaic power output measurement value. Step 4: Based on the distributed energy storage parameters provided in Step 1 and the real-time measurement data of each region in Step 3, establish a data-driven active distribution network voltage control model with the goal of minimizing the deviation between the measured node voltage value and the node voltage reference value in each region and minimizing the change in the charging and discharging power of distributed energy storage. Step 5: Solve the model using the gradient descent method and add sliding mode control to obtain the distributed energy storage charging and discharging strategy. Send the control strategy to the distributed energy storage to obtain new node voltage measurements and photovoltaic power output measurements in each region. Step 6: Obtain the current control duration and determine whether the current control duration is greater than the total control duration. If it is less, return to step 3; otherwise, end the process.

2. The data-driven adaptive sliding mode control method for distributed energy storage according to claim 1, characterized in that: The parameters of the distributed energy storage controller in step 1 include: node voltage reference value. Weighting coefficient , , , Iterative correction coefficient Initial value of sliding mode adaptive adjustment coefficient for distributed energy storage Photovoltaic power change rate coefficient The timing begins when the distributed energy storage controller is connected to the active power distribution network, and is recorded as follows: .

3. The data-driven adaptive sliding mode control method for distributed energy storage according to claim 1, characterized in that: The method for calculating the voltage sensitivity of distributed energy storage to each node in step 2 is as follows: ； in, Distributed energy storage For nodes Voltage sensitivity; and They represent Time and Time Node The voltage measurement value; and They represent Time and Distributed energy storage at all times The charging and discharging power.

4. The data-driven adaptive sliding mode control method for distributed energy storage according to claim 1, characterized in that: The specific implementation method of node voltage sensitivity similarity clustering partitioning in step 2 is as follows: Construct the voltage sensitivity vector of each node: ； in, Represents a node The voltage sensitivity vector; Distributed energy storage For nodes The voltage sensitivity of the voltage; This indicates the quantity of distributed energy storage. The node voltage sensitivity of the distributed energy storage access node is selected as the initial centroid, and then the Euclidean distance is calculated based on the voltage sensitivity vector of each node. Each node is assigned to the nearest regional centroid, and the new centroids of each region are updated: ； in, Represents a node and center of mass The Euclidean distance; Represents a node The voltage sensitivity vector; Indicates the center of mass , representing the voltage sensitivity vector; For the region ; Represents a node and center of mass The Euclidean distance; For the region The number of nodes is determined through a finite number of iterations until the centroid no longer changes, thus dividing the active distribution network into... Each region has a distributed energy storage controller that controls the region's voltage.

5. The data-driven adaptive sliding mode control method for distributed energy storage according to claim 1, characterized in that: The sliding mode adaptive adjustment coefficient in step 3 is calculated as follows: ； in, express Real-time distributed energy storage sliding mode adaptive adjustment coefficient; This represents the initial value of the sliding mode adaptive adjustment coefficient for distributed energy storage; Indicates the rate of change coefficient of photovoltaic power; and express Time and The measured value of photovoltaic power output at any given time.

6. The data-driven adaptive sliding mode control method for distributed energy storage according to claim 1, characterized in that: The data-driven active distribution network voltage control model in step 4 is as follows: ； in, express The estimated value of the node voltage at time point. express Voltage measurement at time node. This represents the charging and discharging power vector of distributed energy storage. and They represent Time and Real-time distributed energy storage charging and discharging power vector express The time-matter pseudo-Jacobi matrix, in this model, is used to reflect the sensitivity relationship between the control objective of the active distribution network and the output strategy of the controlled equipment: ； in, yes Voltage measurement at time node. It is estimated by the distributed energy storage controller. Node voltage at time, This represents the weighting coefficient. express The pseudo-Jacobi matrix at time. express The pseudo-Jacobi matrix at time.

7. The data-driven adaptive sliding mode control method for distributed energy storage according to claim 1, characterized in that: The objective function for data-driven active distribution network voltage control in step 4 is: ； in, yes The vector of node voltage reference values ​​at time t; express The vector of node voltage measurements at time t; and They represent Time and Momentary distributed energy storage charging and discharging power vector; This represents the weighting coefficient.

8. The data-driven adaptive sliding mode control method for distributed energy storage according to claim 1, characterized in that: The specific implementation method of step 5 is as follows: ； in, and They represent Time and Momentary distributed energy storage charging and discharging power vector; express The pseudo-Jacobi matrix at time; and They represent Time and The vector of node voltage reference values ​​at time t; and They represent Time and The vector of node voltage measurements at time t; This represents the weighting coefficient. By employing the gradient descent method and simplifying the matrix inversion process, the pseudo-Jacobi matrix can be obtained. The method is as follows: ； in, and They represent Time and The pseudo-Jacobi matrix at time; express Change in voltage measurement at time node; express The change in the charging and discharging power vector of distributed energy storage at any given moment; and They represent Time and The vector of node voltage measurements at time t; and They represent Time and Momentary distributed energy storage charging and discharging power vector; This represents the weighting coefficient. In specific active distribution network voltage control problems, to enhance the time-varying tracking effect of the control method, the pseudo-Jacobi matrix is ​​corrected. The estimation method is as follows: sign( ) sign( ); in, Represents the pseudo-Jacobi matrix Initial value; express The change in the charging and discharging power vector of distributed energy storage at any given moment; These are the iterative correction coefficients.

Images