Double-layer optimization scheduling method, device and equipment for power distribution network and storage medium

Abstract

The invention relates to the technical field of power distribution network scheduling, and discloses a power distribution network double-layer optimization scheduling method, device and equipment and a storage medium, and the method comprises the steps: collecting real-time data and prediction data of each distributed power supply in a power distribution network, and constructing a data set; based on the data set, establishing an optimal scheduling model with maximization of new energy consumption as a target, and considering constraint conditions of energy storage equipment; solving the model to obtain a power generation plan and an energy storage charging and discharging plan; and when detecting that the real-time voltage exceeds the limit, starting an emergency control strategy, and adjusting reactive power output by using a recursive least square algorithm. The problem that according to an existing power distribution network dispatching scheme, when the distributed energy connected into the power distribution network is in a dynamic state, the voltage stability of the power distribution network cannot be guaranteed is solved.

Description

Technical Field

[0001] This invention relates to the field of power distribution network dispatching technology, specifically a two-layer optimized dispatching method, device, equipment, and storage medium for power distribution networks. Background Technology

[0002] With the widespread application of distributed power sources (such as solar and wind power) in distribution networks, the operating characteristics and control strategies of distribution networks have undergone significant changes. The high proportion of renewable energy integration presents numerous challenges to distribution networks: First, wind and solar power generation exhibit significant intermittency and volatility, and their output is greatly affected by weather conditions, increasing the difficulty of power balance and voltage regulation in the distribution network; second, the large-scale integration of distributed power sources alters the traditional unidirectional power flow characteristics of distribution networks, forming a complex multi-source power supply network structure; third, while the integration of energy storage devices can mitigate power fluctuations, their charging and discharging strategies are closely related to the grid's operating status, requiring comprehensive consideration of various constraints.

[0003] Current power grid dispatching primarily employs a static dispatching scheme based on the assumptions of centralized generation and fixed loads. This scheme has significant shortcomings: firstly, it cannot effectively address the risk of voltage exceedances caused by fluctuations in distributed power output, especially in scenarios with high renewable energy penetration, where node voltage fluctuations are more pronounced; secondly, traditional dispatching methods do not adequately consider the constraints on energy storage devices, making it difficult to achieve synergistic optimization between maximizing renewable energy consumption and ensuring safe grid operation. Furthermore, existing voltage regulation strategies are mostly reactive, only adjusting passively after voltage exceedances are detected, lacking preventative control measures, resulting in slow response times and unsatisfactory performance. Summary of the Invention

[0004] This invention proposes a two-layer optimized scheduling method, device, equipment, and storage medium for distribution networks, aiming to solve the problem that existing distribution network scheduling schemes cannot guarantee grid voltage stability when distributed energy sources connected to the distribution network undergo dynamic changes.

[0005] In a first aspect, the present invention provides a two-layer optimized scheduling method for a distribution network, comprising:

[0006] Real-time and predicted data of each distributed power source in the distribution network are collected and a dataset is constructed; wherein the predicted data includes predicted power generation data and predicted load data, and the real-time data includes the status parameters of energy storage devices and the connection topology of each distributed power source.

[0007] An optimized scheduling model for the distribution network is established based on the dataset, with the goal of maximizing the absorption of new energy and considering the constraints of each energy storage device. The constraints include line safety constraints and upper and lower limits of SOC and daily SOC clearing constraints for each energy storage device.

[0008] Solving the optimized scheduling model yields the generation plan and energy storage charging and discharging plan for each distributed power source.

[0009] When a real-time voltage over-limit is detected in the distribution network, a preset emergency control strategy is activated. The recursive least squares algorithm is used to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source, so that the voltage is restored to a safe range.

[0010] In an optional implementation, the step of establishing an optimized scheduling model for the distribution network based on the dataset, with the goal of maximizing the absorption of new energy sources and considering the constraints of each energy storage device, includes:

[0011] The power transmitted by the node at each time point is calculated based on the predicted power generation data and the predicted load data.

[0012] An optimal scheduling model for the distribution network is constructed with the objective function of minimizing the total power transmitted from the transmission lines and the optimization direction of maximizing the renewable energy absorption rate.

[0013] Based on the predicted power generation data, the predicted load data, the state parameters, and the connection topology, the line safety constraints and the upper and lower limits of SOC and daily SOC clearing constraints of each energy storage device are calculated.

[0014] In an optional implementation, solving the optimized scheduling model to obtain the generation plan and energy storage charging / discharging plan for each distributed power source includes:

[0015] The particle swarm optimization algorithm is used to solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan for each distributed power source.

[0016] In an optional implementation, the step of employing a particle swarm optimization algorithm to solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan for each distributed power source includes:

[0017] A particle swarm is constructed based on the real-time data and the predicted data. The objective function is used to iteratively solve for each particle in the particle swarm to obtain the corresponding optimal solution.

[0018] Determine whether the optimal solution satisfies the constraints;

[0019] If the constraints are met, then select the solution that is closest to the upper limit of the constraints from the optimal solutions that meet the constraints, and obtain the power generation plan and energy storage charging and discharging plan for each distributed power source.

[0020] In an optional implementation, the activation of a preset emergency control strategy, which uses a recursive least squares algorithm to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source, includes:

[0021] Using the recursive least squares algorithm, a sensitivity matrix is ​​constructed based on the connection topology of each distributed power source and combined with the real-time voltage deviation and reactive power deviation at continuous time intervals.

[0022] The reactive power regulation of the distribution network is calculated based on the sensitivity matrix, and the reactive power output of the corresponding distributed power source is adjusted based on the reactive power regulation.

[0023] In an optional implementation, the step of constructing a sensitivity matrix using a recursive least squares algorithm based on the connection topology of each distributed power source and the real-time voltage deviation at consecutive times includes:

[0024] The recursive least squares parameters are initialized based on the connection topology of each distributed power source. The recursive least squares parameters are then iteratively updated by combining the node voltage deviation ΔV and reactive power deviation ΔQ measured in real time at continuous moments to obtain the sensitivity matrix.

[0025] In an optional implementation, the recursive least squares parameters are initialized based on the connection topology of each distributed power source. These parameters are then iteratively updated using the real-time measured node voltage deviation ΔV and reactive power deviation ΔQ to obtain a sensitivity matrix, including:

[0026] Initialize the initial sensitivity matrix, covariance matrix, and forgetting factor based on the connection topology of each distributed power source;

[0027] Based on the real-time measured node voltage deviation ΔV and reactive power deviation ΔQ at continuous moments, calculate the prediction error, gain matrix, and new covariance matrix;

[0028] Based on the prediction error, the gain matrix, and the new covariance matrix, the initial sensitivity matrix is ​​iteratively updated to obtain the sensitivity matrix.

[0029] Secondly, the present invention provides a two-layer optimized dispatching device for a distribution network, comprising:

[0030] The acquisition module is used to collect real-time and forecast data from each distributed power source in the distribution network and construct a dataset; wherein the forecast data includes forecasted generation data and forecasted load data, and the real-time data includes the status parameters of the energy storage device and the connection topology of each distributed power source.

[0031] The first optimization module is used to establish an optimized scheduling model for the distribution network based on the dataset, with the goal of maximizing the absorption of new energy and considering the constraints of each energy storage device. The constraints include line safety constraints and upper and lower limits of SOC and daily SOC clearing constraints for each energy storage device.

[0032] The solution module is used to solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan of each distributed power source.

[0033] The second optimization module is used to activate a preset emergency control strategy when the real-time voltage of the distribution network is detected to exceed the limit. The module uses a recursive least squares algorithm to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source so that the voltage is restored to a safe range.

[0034] Thirdly, the present invention provides an electronic device including a processor and a memory, the memory storing machine-executable instructions that can be executed by the processor, the processor executing the machine-executable instructions to implement the above-mentioned two-layer optimized scheduling method for power distribution networks.

[0035] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement the aforementioned two-layer optimized scheduling method for power distribution networks.

[0036] The embodiments of the present invention bring the following beneficial effects:

[0037] This invention provides a two-layer optimized scheduling method, device, equipment, and storage medium for a distribution network. It collects real-time and predicted data from various distributed power sources in the distribution network and constructs a dataset. The predicted data includes predicted generation data and predicted load data, while the real-time data includes the state parameters of energy storage devices and the connection topology of each distributed power source. Based on the dataset, an optimized scheduling model for the distribution network is established, aiming to maximize the absorption of new energy sources while considering the constraints of each energy storage device. These constraints include line safety constraints and upper and lower limits of the State of Charge (SOC) and daily SOC clearing constraints for each energy storage device. The optimized scheduling model is solved to obtain the generation plan and energy storage charging and discharging plan for each distributed power source. When a real-time voltage exceedance is detected in the distribution network, a preset emergency control strategy is activated, using a recursive least squares algorithm to adjust the reactive power output of each distributed power source based on its connection topology. This method maximizes the absorption of new energy sources while ensuring the safe operation of the power grid by constructing a two-layer optimization model combined with a real-time voltage dynamic adjustment mechanism. At the same time, it uses a recursive least squares algorithm to achieve rapid voltage correction, thereby improving the energy utilization rate in the distribution network, preventing voltage overruns, and optimizing the coordinated operation of energy storage.

[0038] Other features and advantages of the invention will be set forth in the following description, or some features and advantages may be inferred from the description or determined without doubt, or may be learned by practicing the techniques described above.

[0039] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0040] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0041] Figure 1 A flowchart of a two-layer optimized scheduling method for power distribution networks provided in an embodiment of the present invention;

[0042] Figure 2 A flowchart for constructing a sensitivity matrix provided in an embodiment of the present invention;

[0043] Figure 3 A flowchart for calculating reactive power regulation provided in an embodiment of the present invention;

[0044] Figure 4This is a schematic diagram of a two-layer optimized dispatching device for power distribution networks provided in an embodiment of the present invention;

[0045] Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0047] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0048] In related technologies, the widespread application of distributed generation has exacerbated voltage fluctuations in power distribution networks. Traditional dispatching schemes rely on assumptions of centralized generation and stable loads, making it difficult to adapt to the highly volatile nature of renewable energy output. For example, when photovoltaic power generation experiences intermittent changes due to weather conditions, static dispatching models cannot adjust their operating strategies in a timely manner, leading to insufficient renewable energy absorption and frequent voltage exceedances, thus affecting power supply reliability.

[0049] To address the aforementioned issues, this invention proposes a method for dynamically adapting to fluctuations in renewable energy output and optimizing scheduling strategies in real time. Specifically, it acquires real-time operating parameters and historical data of distributed power sources using data acquisition equipment, constructing a standardized dataset with time alignment. Based on predicted data, it calculates the power transmitted through transmission lines for each time period, establishing an optimization model with the objectives of minimizing transmitted power and maximizing absorption rate, while adding constraints such as line capacity and energy storage state of charge. The model is solved using a particle swarm optimization algorithm to obtain a day-ahead scheduling plan that balances economy and safety. During real-time operation, if a node voltage is detected to exceed the safe range, an emergency control module is immediately triggered. A recursive least squares algorithm is used to identify the system sensitivity online, generate reactive power regulation commands, and send them to the corresponding distributed power source, enabling the voltage to quickly return to normal levels.

[0050] Please see Figure 1 A specific implementation of the two-layer optimized scheduling method for distribution networks provided by the present invention includes the following steps:

[0051] 110. Collect real-time and forecast data from each distributed power source in the distribution network and construct a dataset.

[0052] It should be noted that the predicted data includes predicted power generation data and predicted load data, which can be generated using a time series model combined with historical data, and used as input for constructing the day-ahead dispatch plan. The real-time data includes the state parameters of energy storage devices and the connection topology of each distributed power source. For example, data acquisition devices are set up in the distribution network to collect the real-time SOC value, line parameters, and connection relationships with adjacent nodes for each node; using historical load curves and meteorological data, a time series prediction model (such as LSTM, ARIMA, or Prophet model) is used to predict the power generation and load demand for the next hour; the collected data undergoes time alignment and standardization processing to unify data from different sampling frequencies to a fixed time step (such as 15 minutes), forming a standardized dataset containing {node number, line parameters, predicted power generation, predicted load, SOC}.

[0053] In this embodiment, the step specifically includes: determining each node in the distribution network, and collecting the status parameters of the energy storage device of each node, as well as the connection relationship with other nodes and the line parameters of the line where it is located, through data acquisition equipment; the status parameters include the initial SOC value and the charging and discharging efficiency.

[0054] Based on the connection relationships of each node and the line parameters, a mapping relationship between nodes and devices is established to obtain the connection topology of each distributed power source.

[0055] Historical operating data of the power distribution network is collected, and time series prediction algorithms are used to combine the state parameters and the connection topology to generate predicted power generation data and predicted load data.

[0056] The state parameters, connection topology, predicted power generation data, and predicted load data are time-aligned and sorted and integrated according to a preset standardized data structure to obtain a dataset.

[0057] The node refers to the connection point for power transmission in the distribution network, which can be identified using power grid topology analysis tools to locate the access points of distributed power sources and energy storage devices.

[0058] The state parameters include the initial SOC value and charge / discharge efficiency of the energy storage device, which can be collected in real time by the battery management system to characterize the current operating state and energy conversion capability of the energy storage device.

[0059] The connection topology refers to the physical connection relationship between distributed power sources and nodes. Specifically, it can be modeled using a graph database or adjacency matrix to describe the power transmission path and network structure characteristics.

[0060] The time series forecasting algorithm refers to a model that predicts future trends based on historical data. For example, ARIMA or LSTM algorithms can be used to generate power generation and load forecasting data by combining energy storage status and topology relationships.

[0061] The standardized data structure refers to the rules for unifying data format and storage order, such as using JSON or CSV format and aligning by timestamps to ensure compatibility and consistency of multi-source data.

[0062] For example, in the data acquisition phase, real-time parameters such as energy storage SOC and charging / discharging efficiency of each node are first acquired through devices such as smart meters and sensors, while network parameters such as line impedance and inter-node connectivity are also recorded. Subsequently, based on graph theory algorithms, the physical connections between nodes and distributed power sources are mapped into a topology model, clarifying the interaction paths between devices. During the prediction data generation process, historical operating data is cleaned and input into the time series model, combining energy storage status and topology constraints to predict future power generation and load demand. Finally, real-time parameters, the topology model, and prediction data are integrated into a standardized dataset through timestamp alignment, eliminating time dimension differences and providing a unified input for subsequent optimized scheduling.

[0063] Furthermore, when establishing the mapping relationship, it is possible to determine the line to which each node belongs and its adjacent nodes based on the distribution network topology; identify the inverter or energy storage device nodes with reactive power regulation capabilities; and construct a node-equipment mapping matrix to describe the sensitivity of node voltage to the reactive power of controllable equipment.

[0064] Furthermore, after collecting historical operating data of the distribution network and generating predicted power generation data and predicted load data by combining the state parameters and the connection topology using a time series prediction algorithm, the method further includes: calculating confidence intervals and detecting anomalies in the predicted data to remove abrupt changes or missing values; and performing moving average filtering on the data and converting it to a unified per-unit system.

[0065] 120. An optimal scheduling model for the power distribution network is established based on the dataset, with the goal of maximizing the absorption of new energy sources and taking into account the constraints of each energy storage device.

[0066] In this embodiment, the constraints include line safety constraints and upper and lower limits of SOC for each energy storage device, as well as daily SOC clearing constraints. The optimization scheduling model here aims to maximize the absorption of new energy sources, and can introduce the minimization of power transmitted from the line as the objective function. The constraints cover the safe operating range of the line, the state of charge limitations of energy storage, and daily clearing requirements, ensuring that the scheduling scheme is within the physically feasible domain.

[0067] Specifically, based on predicted power generation and load data, the power transmitted by each node through the lines is calculated; with the objective function of "minimizing the total power transmitted by the lines", an optimized scheduling model for the distribution network is established; considering line safety constraints, upper and lower limits of SOC constraints for energy storage devices, and daily SOC clearing constraints, the goal is to ensure that new energy is absorbed within the distribution network as much as possible.

[0068] 130. Solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan for each distributed power source.

[0069] Specifically, by using the Particle Swarm Optimization (PSO) algorithm as the solver, the power generation output and energy storage power of each distributed power source are treated as particle variables; the objective function is updated iteratively with convergence as the criterion to obtain the optimal solution that satisfies the constraints; where the optimal solution is the power generation plan and energy storage charging and discharging plan of each distributed power source.

[0070] 140. When a real-time voltage over-limit is detected in the distribution network, a preset emergency control strategy is activated. The recursive least squares algorithm is used to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source so that the voltage is restored to a safe range.

[0071] In this embodiment, the detection of whether the real-time voltage of the distribution network exceeds the limit is mainly achieved by continuously monitoring the real-time voltage of each node and calculating the difference between the real-time voltage and the initial voltage reference value of the node. When the difference exceeds a preset threshold, the lower-level control is triggered. The lower-level control here is to start a preset emergency control strategy and use the recursive least squares algorithm to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source. The nodes that exceed the limit are identified and a set of controllable devices is determined for voltage recovery control.

[0072] It should be noted that this emergency control strategy is based on the recursive least squares (RLS) algorithm to dynamically estimate the sensitivity matrix and adjust the reactive power output of each distributed power source according to the sensitivity relationship. Specifically, the recursive least squares algorithm is used to update the sensitivity matrix online and dynamically adjust the reactive power output. For example, the adjustment amount is iteratively calculated by combining the voltage deviation and reactive power deviation data at continuous time intervals to quickly suppress voltage fluctuations.

[0073] In practical applications, the recursive least squares parameters are first initialized using the distribution network topology as prior information, including the initial sensitivity matrix H0, covariance matrix G0, and forgetting factor σ. After continuously measuring the node voltage deviation ΔV and reactive power deviation ΔQ, the above parameters are iterated continuously until the residuals satisfy the convergence criterion, thus obtaining the real-time sensitivity matrix. The sensitivity coefficients of each node are extracted from the sensitivity matrix. The reactive power regulation is calculated for the node with the largest voltage deviation. Based on the sensitivity weights, the constrained least squares problem can be further solved to obtain the final reactive power regulation. An adjustment command is generated based on the reactive power regulation and sent to the distributed power source control terminal to adjust its reactive power output. When all node voltages recover to a safe range or the regulation reaches the set limit after real-time monitoring, the emergency control cycle is stopped.

[0074] This embodiment uses a recursive least squares algorithm to update the sensitivity matrix online, which can dynamically track changes in system operating status and improve voltage regulation accuracy. Furthermore, the dual-layer optimization mechanism combines day-ahead planning with real-time correction, ensuring optimal global renewable energy consumption while responding promptly to real-time disturbances, and also addressing the lack of flexibility in static scheduling models.

[0075] In one specific implementation, the step of establishing an optimized scheduling model for the distribution network based on the dataset, with the objective of maximizing the absorption of new energy sources and considering the constraints of each energy storage device, includes:

[0076] The power transmitted by the node at each time point is calculated based on the predicted power generation data and the predicted load data.

[0077] An optimal scheduling model for the distribution network is constructed with the objective function of minimizing the total power transmitted from the transmission lines and the optimization direction of maximizing the renewable energy absorption rate.

[0078] Based on the predicted power generation data, the predicted load data, the state parameters, and the connection topology, the line safety constraints and the upper and lower limits of SOC and daily SOC clearing constraints of each energy storage device are calculated.

[0079] It should be noted that the power transmitted through the transmission line refers to the power transmitted by nodes in the distribution network to the external power grid at different times. Specifically, it can be dynamically estimated using a power flow calculation model combined with predicted generation data and load data to quantify the potential for renewable energy absorption. The formula for this power transmission through the transmission line is as follows: , where N pv N represents the total number of photovoltaic cells; bess N represents the total number of energy storage devices. load P represents the total number of loads; line (t) represents the power transmitted by the line at time t; P pv (i,t) represents the predicted power generation data of photovoltaic i at time t; Pload (j,t) represents the predicted load data for load j at time t; P bess (k,t) represents the data consumed by energy storage device k at time t.

[0080] The objective function is defined as minimizing the total power transmitted through the transmission lines. Reducing the transmitted power indirectly reflects the improvement in the local absorption capacity of renewable energy. Specifically, it can be constructed using linear programming or mixed-integer programming methods. If the number of nodes in the distribution network is 96, the objective function based on the aforementioned formula for transmission power is: .

[0081] The line safety constraint refers to the line transmission power not exceeding its rated capacity. Specifically, it can be modeled by limiting the node voltage amplitude and the line current thermal stability limit to avoid overload risks.

[0082] The SOC upper and lower limit constraints refer to the state of charge of the energy storage device being maintained within a preset range during charging and discharging. Specifically, a piecewise linearization method can be used to express this to prevent overcharging or over-discharging of the battery.

[0083] The SOC daily clearing constraint means that the energy storage device must be restored to its initial state of charge at the end of the scheduling cycle. This can be achieved by setting a charge and discharge balance equation to ensure the sustainable operation capability of the energy storage device.

[0084] If the distribution network has 96 nodes, the line safety constraints should meet the following requirements. P line,rated Indicates the rated power of the line; the upper and lower limits of the SOC of the energy storage device should meet the following requirements. , of which S SOC,min S represents the minimum value of the energy storage device. SOC,max This represents the maximum value of the energy storage device; the daily SOC constraint of the energy storage device should meet the following requirements. .

[0085] In practical applications, when constructing an optimized scheduling model, the first step is to determine the power surplus or deficit of each node at different times based on predicted power generation and load data through power flow calculations, and then calculate the power transmitted through the lines. The objective function focuses on minimizing the total transmitted power, and promotes the priority of local consumption of new energy by adjusting the output of distributed power sources and the charging and discharging plans of energy storage. Regarding constraints, line safety constraints are modeled using the range of node voltage amplitude deviation and line transmission power limits; the upper and lower limits of the State of Charge (SOC) of energy storage devices are expressed by setting the dynamic range of state of charge change; and the daily SOC clearing constraint is constrained by the total charge and discharge balance equation. For example, during peak photovoltaic output periods, if the local load cannot fully absorb the energy, the model will prioritize scheduling energy storage devices to store excess energy, while simultaneously reducing the external transmission demand by minimizing the transmitted power through the lines, thereby improving the new energy consumption rate. At the same time, by dynamically optimizing the transmitted power and energy storage scheduling strategies, the model maximizes the local consumption capacity of new energy while ensuring the safe operation of the distribution network, reducing power dependence on the upper-level grid, thereby reducing the risk of voltage exceeding limits and improving power supply reliability.

[0086] In the process of constructing the optimized scheduling model, line safety constraints limit the power transmission limits of each line to keep line current and node voltage within safe operating ranges. State of Charge (SOC) upper and lower limits constraints for energy storage devices ensure operation within their rated capacity range by setting the state of charge boundaries for each scheduling period. Daily SOC clearing constraints enforce consistency between the initial and final states of charge of the energy storage devices each day by establishing charge-discharge balance equations for the devices over a 24-hour cycle. These three types of constraints collectively constitute the boundary conditions of the optimization model. Mathematical programming methods are used to ensure that all solutions satisfy the aforementioned physical constraints when solving for the generation and energy storage charge-discharge plans.

[0087] In one specific implementation, solving the optimized scheduling model to obtain the generation plan and energy storage charging and discharging plan for each distributed power source includes:

[0088] The particle swarm optimization algorithm is used to solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan for each distributed power source.

[0089] It should be noted that the particle swarm optimization algorithm is an optimization algorithm based on swarm intelligence. It performs iterative search by simulating the foraging behavior of bird flocks. Specifically, it can be implemented by initializing the particle swarm, calculating the fitness value, and updating the particle velocity and position. It is suitable for optimization problems with multiple variables and multiple constraints and can quickly converge to the global optimum.

[0090] The power generation plan refers to the output arrangement of distributed power sources at different time periods. It can be represented by a power allocation table or a time-series curve to guide the consumption of new energy sources and the balance of power in the power grid.

[0091] The energy storage charge and discharge plan refers to the charge and discharge power and timing of the energy storage device within the scheduling cycle. Specifically, it can be generated using a charge and discharge state switching strategy to regulate grid power fluctuations and meet SOC constraints.

[0092] Specifically, a particle swarm is constructed based on real-time and predicted data, with each particle representing a possible combination of power generation plans, energy storage charging / discharging plans, and voltage reference values. The fitness value of each particle is calculated using an objective function, such as a weighted evaluation result of the total transmitted power and the renewable energy absorption rate. During the iteration process, particles update their positions based on individual and swarm optimal solutions, gradually approaching the optimal solution that satisfies the constraints. For example, the particle swarm size can be set to 50-200 particles, the number of iterations can be set to 100-500, and the inertia weight can adopt a linear decreasing strategy. When the particle swarm converges, the optimal solution that satisfies line safety constraints, SOC upper and lower limit constraints, and daily clearing constraints is selected, and the corresponding power generation plan and energy storage charging / discharging plan are output.

[0093] The above-described scheme enables the rapid generation of optimized scheduling plans that meet multiple constraints when distributed generation output fluctuates, reducing the risk of voltage exceedances and improving the efficiency of renewable energy integration. For example, when photovoltaic output drops sharply, adjusting the energy storage discharge power and distributed generation output plans maintains stable node voltage, preventing malfunctions of protection devices due to voltage dips. Furthermore, the parallel computing characteristics of the particle swarm optimization algorithm shorten the solution time, meeting real-time scheduling requirements and thus improving the reliability and power quality of the distribution network.

[0094] In one specific implementation, the step of employing a particle swarm optimization algorithm to solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan for each distributed power source includes:

[0095] A particle swarm is constructed based on the real-time data and the predicted data. The objective function is used to iteratively solve for each particle in the particle swarm to obtain the corresponding optimal solution.

[0096] Determine whether the optimal solution satisfies the constraints;

[0097] If the constraints are met, then select the solution that is closest to the upper limit of the constraints from the optimal solutions that meet the constraints, and obtain the power generation plan and energy storage charging and discharging plan for each distributed power source.

[0098] It should be noted that the optimal solution selected here refers to the scheduling scheme with the optimal objective function value under all constraints. Specifically, it can be obtained by comparing the fitness values ​​of different particles. Its role is to provide the distribution network with a scheduling strategy that takes into account both renewable energy consumption and equipment safety.

[0099] Specifically, in solving the optimal scheduling model, an initial population of multiple particles is first constructed based on real-time collected energy storage device state parameters, distributed power supply connection topology, and predicted power generation and load data. Each particle represents a possible scheduling scheme, including variables such as power generation plan, energy storage charge / discharge plan, and voltage reference value. During the iteration process, the fitness value of each particle is calculated according to the objective function, and the optimal solution is gradually approached by updating the particle's velocity and position. After each iteration, it is necessary to verify whether the current solution meets the line safety constraints, energy storage SOC limit, and daily clearing requirements. If a feasible solution exists, the scheme with the optimal objective function value and close to the upper limit of constraints is selected, and the power generation plan, charge / discharge plan, and voltage reference value are finally output.

[0100] In another feasible implementation, the activation of the preset emergency control strategy, which uses a recursive least squares algorithm to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source, includes:

[0101] Using the recursive least squares algorithm, a sensitivity matrix is ​​constructed based on the connection topology of each distributed power source and combined with the real-time voltage deviation and reactive power deviation at continuous time intervals.

[0102] The reactive power regulation of the distribution network is calculated based on the sensitivity matrix, and the reactive power output of the corresponding distributed power source is adjusted based on the reactive power regulation.

[0103] The recursive least squares algorithm is an online parameter estimation method that can update system parameters through recursion. Specifically, it can be implemented using an iterative formula with a forgetting factor and is suitable for dynamically changing power grid environments.

[0104] The sensitivity matrix refers to the response relationship between node voltage and reactive power output changes. Specifically, it can be obtained by calculating the correlation between real-time voltage deviation and reactive power deviation, and is used to quantify the mapping relationship between regulation amount and voltage change.

[0105] The reactive power regulation refers to the reactive power value that needs to be adjusted to eliminate voltage over-limit. Specifically, it can be obtained through the inverse operation of the sensitivity matrix or by solving a system of linear equations, and is used to generate regulation commands for specific distributed power sources.

[0106] In practical applications, when a voltage exceedance is detected, the system initializes the parameters of the recursive least squares algorithm based on the connection topology, such as the initial sensitivity coefficients and covariance matrix. By continuously collecting node voltage and reactive power deviation data at multiple time points, the algorithm dynamically updates the elements in the sensitivity matrix. For example, after measuring ΔV and ΔQ at each time point, the prediction error is calculated and the covariance matrix is ​​updated, thereby adjusting the sensitivity coefficients. The final generated sensitivity matrix accurately reflects the degree of influence of reactive power output of each node on the voltage. Based on this matrix, solving the linear equations yields the reactive power regulation required for each distributed power source, and control commands are used to adjust its reactive power output, bringing the voltage back to a safe range.

[0107] Furthermore, the method of constructing a sensitivity matrix based on the connection topology of each distributed power source and the real-time voltage deviation at consecutive times using the recursive least squares algorithm includes:

[0108] The recursive least squares parameters are initialized based on the connection topology of each distributed power source. The recursive least squares parameters are then iteratively updated by combining the node voltage deviation ΔV and reactive power deviation ΔQ measured in real time at continuous moments to obtain the sensitivity matrix.

[0109] The recursive least squares parameters include the forgetting factor σ, the initial sensitivity matrix H0, and the covariance matrix G=10. 6 E, where E is the identity matrix. The initial sensitivity matrix refers to the initial estimation matrix reflecting the sensitivity of node voltage to changes in reactive power output. Specifically, it can be initialized using the admittance matrix based on the connection topology, used to establish a preliminary correlation between reactive power regulation and voltage deviation. The covariance matrix refers to the statistical characteristic matrix used to describe the parameter estimation error. Specifically, it can be initialized using the identity matrix or historical data statistics, used to correct the step size of parameter updates during iteration. The forgetting factor refers to the weighting coefficient used to adjust the influence of historical data on the current estimate. Specifically, it can use a value in the range of 0.95 to 0.99, used to balance the algorithm's ability to track system dynamic changes and its resistance to noise interference.

[0110] When a voltage exceedance occurs in the distribution network, an initial sensitivity matrix is ​​first constructed based on the electrical connections between distributed power sources. Simultaneously, the covariance matrix is ​​set to an identity matrix with diagonal elements of 1, and a forgetting factor of 0.98 is selected. When a node voltage deviation ΔV = V is detected between two consecutive time points... t -V t-1 and reactive power deviation ΔQ=Q t -Q t-1At that time, the difference between the actual measured value and the model prediction value is obtained by calculating the prediction error at the current moment. Then, the parameter update direction is adjusted using the gain matrix, and the covariance matrix is ​​dynamically decayed using a forgetting factor, finally generating an updated sensitivity matrix. This process continues until a stable sensitivity coefficient is obtained, ensuring that the parameter estimation can follow the real-time changes in the power grid's operating state.

[0111] Specifically, the sensitivity matrix H, which is associated with the reactive power of each node, is constructed by recursively calculating the sensitivity matrix through least squares estimation.

[0112] According to the power flow equation, we have: ΔV≈HΔQ. Based on this, the formula for the sensitivity matrix H is as follows:

[0113] ,

[0114] ,

[0115] .

[0116] Will[ ... The sensitivity matrix is ​​represented as H1 = [k11, k12, ..., knn], where k11, k12, ..., knn are the sensitivity coefficient vectors corresponding to node 1, and so on for other nodes. The final sensitivity matrix is ​​as follows:

[0117] = .

[0118] Furthermore, the recursive least squares parameters are initialized based on the connection topology of each distributed power source. These parameters are then iteratively updated using the real-time measured node voltage deviation ΔV and reactive power deviation ΔQ to obtain a sensitivity matrix, including:

[0119] Initialize the initial sensitivity matrix, covariance matrix, and forgetting factor based on the connection topology of each distributed power source;

[0120] Based on the real-time measured node voltage deviation ΔV and reactive power deviation ΔQ at continuous moments, calculate the prediction error, gain matrix, and new covariance matrix;

[0121] Based on the prediction error, the gain matrix, and the new covariance matrix, the initial sensitivity matrix is ​​iteratively updated to obtain the sensitivity matrix.

[0122] In other words, an initial sensitivity matrix is ​​first constructed based on the physical connections of the distributed power sources. This matrix reflects the inherent influence of each node's reactive power output on voltage changes. The covariance matrix is ​​initialized as an identity matrix with large diagonal elements to ensure the stability of the parameter update process. The forgetting factor is set to a fixed value to maintain the continuity of data updates. During real-time operation, node voltage deviation and reactive power deviation data are collected at set intervals. The gain matrix is ​​calculated using a recursive formula, and the covariance matrix and sensitivity matrix are updated accordingly. For example, when a continuous increase in the voltage deviation of a node is detected, the system adjusts the weights of the corresponding elements in the sensitivity matrix using the gain matrix, making the subsequent reactive power regulation calculation more closely reflect the actual operating state.

[0123] In another embodiment, the step of calculating the reactive power regulation of the distribution network based on the sensitivity matrix, and adjusting the reactive power output of the corresponding distributed power source based on the reactive power regulation, includes:

[0124] Extract the sensitivity coefficient of each element in the sensitivity matrix, and calculate the reactive power regulation of the distribution network based on each sensitivity coefficient;

[0125] Based on the reactive power regulation, a corresponding regulation command is generated, and the reactive power output of the corresponding distributed power source is adjusted based on the regulation command until the real-time voltage of the distribution network is within a safe range.

[0126] Specifically, upon detecting a voltage exceedance, the sensitivity coefficient for each node is first extracted from the sensitivity matrix. This coefficient characterizes the ability of a specific distributed power source to correct voltage deviations due to changes in reactive power output. Then, based on the product of the current voltage deviation and the sensitivity coefficient, the required reactive power adjustment for each distributed power source is calculated, generating an adjustment command containing the target value. This adjustment command is transmitted to the corresponding distributed power source control unit via a communication module, driving the inverter to adjust the reactive power output. During the adjustment process, voltage changes are continuously monitored. If the voltage still exceeds the safe range, the calculation and adjustment steps are repeated until all node voltages return to the allowable range.

[0127] like Figure 2-3 As shown, assuming an initial forgetting factor σ = 0.95, used to control the algorithm's weighting of historical data, an initial sensitivity matrix H0, and an initial sensitivity covariance matrix G, typically taken as a diagonal matrix with larger diagonal elements, G = 10. 6 E (where E is the identity matrix), with a time step of t=1.

[0128] Online measurement of voltage deviation ΔV=V t -V t-1 reactive power deviation ΔQ=Q t -Q t-1 ;

[0129] Calculate the prediction error: ε = ΔQ - H0' * ΔV;

[0130] Calculate the gain variable: K=(G*ΔV) / (σ+ΔV'*G *ΔV);

[0131] The new covariance matrix is: G = (GK * ΔV') / σ;

[0132] Iteratively update the sensitivity matrix: H = H0 + K * ε;

[0133] If the time interval t > N, output the matrix; otherwise, if t = t + 1, return to continue calculating the parameters and obtain a new sensitivity matrix.

[0134] Calculate the voltage deviation vector ΔV for all nodes:

[0135] , This represents the voltage at node i. Indicates the voltage is above the limit. Indicates the lower limit of voltage;

[0136] Select node m with the largest voltage deviation, select node j containing distributed power sources according to equipment priority, and calculate the reactive power adjustment ΔQ based on the sensitivity coefficient. adj =min(ΔV / k mj Q maxj -Qj), if the voltage does not meet the requirements, select the next distributed power node; N pv Q represents the total number of distributed power sources; maxj k represents the maximum reactive power at node j; mj Let be the voltage-reactive power sensitivity coefficient of node m to node j in the sensitivity matrix H.

[0137] In summary, the two-layer optimization method of this invention features rapid response and can effectively meet the real-time operation requirements of distribution networks. By optimizing the scheduling model to maximize green electricity generation and optimizing reactive power regulation to achieve emergency voltage safety control, this method can accurately assess and adjust the operating status of the distribution network in a short time. Whether facing random fluctuations in distributed power sources or uncertainties in load, it can quickly take measures to ensure that the voltage at each node of the distribution network remains within a stable range. This highly efficient response capability not only improves the operational economy and reliability of the distribution network but also provides strong technical support for the widespread integration of distributed energy resources.

[0138] The above describes the two-layer optimized scheduling method for distribution networks in the embodiments of the present invention. The following describes the two-layer optimized scheduling device for distribution networks in the embodiments of the present invention; please refer to... Figure 4One embodiment of the distribution network two-layer optimized dispatching device in this invention includes:

[0139] The acquisition module 410 is used to acquire real-time data and forecast data of each distributed power source in the distribution network and construct a dataset; wherein the forecast data includes forecast power generation data and forecast load data, and the real-time data includes the status parameters of the energy storage device and the connection topology of each distributed power source.

[0140] The first optimization module 420 is used to establish an optimized scheduling model of the distribution network based on the dataset, with the goal of maximizing the absorption of new energy and considering the constraints of each energy storage device, wherein the constraints include line safety constraints and upper and lower limits of SOC and daily SOC clearing constraints of each energy storage device.

[0141] The solution module 430 is used to solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan of each distributed power source.

[0142] The second optimization module 440 is used to activate a preset emergency control strategy when the real-time voltage of the distribution network is detected to exceed the limit. The strategy uses a recursive least squares algorithm to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source so that the voltage is restored to a safe range.

[0143] Optionally, the first optimization module 420 is specifically used for:

[0144] The power transmitted by the node at each time point is calculated based on the predicted power generation data and the predicted load data.

[0145] An optimal scheduling model for the distribution network is constructed with the objective function of minimizing the total power transmitted from the transmission lines and the optimization direction of maximizing the renewable energy absorption rate.

[0146] Based on the predicted power generation data, the predicted load data, the state parameters, and the connection topology, the line safety constraints and the upper and lower limits of SOC and daily SOC clearing constraints of each energy storage device are calculated.

[0147] Optionally, the solver module 430 is specifically used for:

[0148] The particle swarm optimization algorithm is used to solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan for each distributed power source.

[0149] Optionally, the solver module 430 is specifically used for:

[0150] A particle swarm is constructed based on the real-time data and the predicted data. The objective function is used to iteratively solve for each particle in the particle swarm to obtain the corresponding optimal solution.

[0151] Determine whether the optimal solution satisfies the constraints;

[0152] If the constraints are met, then select the solution that is closest to the upper limit of the constraints from the optimal solutions that meet the constraints, and obtain the power generation plan and energy storage charging and discharging plan for each distributed power source.

[0153] Optionally, the second optimization module 440 is specifically used for:

[0154] Using the recursive least squares algorithm, a sensitivity matrix is ​​constructed based on the connection topology of each distributed power source and combined with the real-time voltage deviation and reactive power deviation at continuous time intervals.

[0155] The reactive power regulation of the distribution network is calculated based on the sensitivity matrix, and the reactive power output of the corresponding distributed power source is adjusted based on the reactive power regulation.

[0156] Optionally, the second optimization module 440 is specifically used for:

[0157] The recursive least squares parameters are initialized based on the connection topology of each distributed power source. The recursive least squares parameters are then iteratively updated by combining the node voltage deviation ΔV and reactive power deviation ΔQ measured in real time at continuous moments to obtain the sensitivity matrix.

[0158] Optionally, the second optimization module 440 is specifically used for:

[0159] Initialize the initial sensitivity matrix, covariance matrix, and forgetting factor based on the connection topology of each distributed power source;

[0160] Based on the real-time measured node voltage deviation ΔV and reactive power deviation ΔQ at continuous moments, calculate the prediction error, gain matrix, and new covariance matrix;

[0161] Based on the prediction error, the gain matrix, and the new covariance matrix, the initial sensitivity matrix is ​​iteratively updated to obtain the sensitivity matrix.

[0162] In this embodiment, real-time and predicted data from various distributed power sources in the distribution network are collected and a dataset is constructed. Based on the dataset, an optimized scheduling model is established with the goal of maximizing the absorption of new energy sources, taking into account the constraints of energy storage devices. The model is solved to obtain the power generation plan and the energy storage charging and discharging plan. When a real-time voltage limit is detected, an emergency control strategy is initiated, and the reactive power output is adjusted using a recursive least squares algorithm. This solves the problem that existing distribution network scheduling schemes cannot guarantee grid voltage stability when distributed energy sources connected to the distribution network are dynamic.

[0163] above Figure 4The distribution network dual-layer optimization scheduling device in this embodiment of the invention will be described in detail from the perspective of modular functional entities. The electronic equipment in this embodiment of the invention will be described in detail from the perspective of hardware processing.

[0164] Figure 5 This is a schematic diagram of the structure of an electronic device 500 provided in an embodiment of the present invention. The electronic device 500 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 510 (e.g., one or more processors) and a memory 520, and one or more storage media 530 (e.g., one or more mass storage devices) for storing application programs 533 or data 532. The memory 520 and storage media 530 can be temporary or persistent storage. The program stored in the storage media 530 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the electronic device 500. Furthermore, the processor 510 may be configured to communicate with the storage media 530 and execute a series of instruction operations on the storage media 530 on the electronic device 500 to implement the two-layer optimized scheduling method for the power distribution network provided in the above embodiment.

[0165] Electronic device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input / output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 5 The illustrated electronic device structure does not constitute a limitation on the electronic device provided by the present invention, and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0166] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the distribution network two-layer optimization scheduling method provided in the above embodiments.

[0167] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system, device, or unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0168] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0169] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A two-layer optimal scheduling method for a distribution network, characterized in that, The method includes: Real-time and predicted data of each distributed power source in the distribution network are collected and a dataset is constructed; wherein the predicted data includes predicted power generation data and predicted load data, and the real-time data includes the status parameters of energy storage devices and the connection topology of each distributed power source. An optimized scheduling model for the distribution network is established based on the dataset, with the goal of maximizing the absorption of new energy and considering the constraints of each energy storage device. The constraints include line safety constraints and upper and lower limits of SOC and daily SOC clearing constraints for each energy storage device. Solving the optimized scheduling model yields the generation plan and energy storage charging and discharging plan for each distributed power source. When a real-time voltage over-limit is detected in the distribution network, a preset emergency control strategy is activated. The recursive least squares algorithm is used to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source, so that the voltage is restored to a safe range.

2. The two-layer optimized scheduling method for distribution networks according to claim 1, characterized in that, The optimized scheduling model for the distribution network established based on the dataset aims to maximize the absorption of new energy sources and considers the constraints of each energy storage device, including: The power transmitted by the node at each time point is calculated based on the predicted power generation data and the predicted load data. An optimal scheduling model for the distribution network is constructed with the objective function of minimizing the total power transmitted from the transmission lines and the optimization direction of maximizing the renewable energy absorption rate. Based on the predicted power generation data, the predicted load data, the state parameters, and the connection topology, the line safety constraints and the upper and lower limits of SOC and daily SOC clearing constraints of each energy storage device are calculated.

3. The two-layer optimized scheduling method for distribution networks according to claim 2, characterized in that, Solving the optimized scheduling model yields the generation plan and energy storage charging / discharging plan for each distributed power source, including: The particle swarm optimization algorithm is used to solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan for each distributed power source.

4. The two-layer optimized scheduling method for distribution networks according to claim 3, characterized in that, The particle swarm optimization algorithm is used to solve the optimized scheduling model, obtaining the generation plan and energy storage charging and discharging plan for each distributed power source, including: A particle swarm is constructed based on the real-time data and the predicted data. The objective function is used to iteratively solve for each particle in the particle swarm to obtain the corresponding optimal solution. Determine whether the optimal solution satisfies the constraints; If the constraints are met, then select the solution that is closest to the upper limit of the constraints from the optimal solutions that meet the constraints, and obtain the power generation plan and energy storage charging and discharging plan for each distributed power source.

5. The two-layer optimized scheduling method for distribution networks according to claim 1, characterized in that, The aforementioned activation of the preset emergency control strategy utilizes a recursive least squares algorithm to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source, including: Using the recursive least squares algorithm, a sensitivity matrix is ​​constructed based on the connection topology of each distributed power source and combined with the real-time voltage deviation and reactive power deviation at continuous time intervals. The reactive power regulation of the distribution network is calculated based on the sensitivity matrix, and the reactive power output of the corresponding distributed power source is adjusted based on the reactive power regulation.

6. The two-layer optimized scheduling method for distribution networks according to claim 5, characterized in that, The method of constructing a sensitivity matrix using a recursive least squares algorithm based on the connection topology of each distributed power source and the real-time voltage deviation at consecutive times includes: The recursive least squares parameters are initialized based on the connection topology of each distributed power source. The recursive least squares parameters are then iteratively updated by combining the node voltage deviation ΔV and reactive power deviation ΔQ measured in real time at continuous moments to obtain the sensitivity matrix.

7. The two-layer optimized scheduling method for distribution networks according to claim 6, characterized in that, The recursive least squares parameters based on the connection topology of each distributed power source are initialized, and then iteratively updated using the node voltage deviation ΔV and reactive power deviation ΔQ measured in real time at continuous moments to obtain a sensitivity matrix, including: Initialize the initial sensitivity matrix, covariance matrix, and forgetting factor based on the connection topology of each distributed power source; Based on the real-time measured node voltage deviation ΔV and reactive power deviation ΔQ at continuous moments, calculate the prediction error, gain matrix, and new covariance matrix; Based on the prediction error, the gain matrix, and the new covariance matrix, the initial sensitivity matrix is ​​iteratively updated to obtain the sensitivity matrix.

8. A two-layer optimized dispatching device for a power distribution network, characterized in that, The device includes: The acquisition module is used to collect real-time and forecast data from each distributed power source in the distribution network and construct a dataset; wherein the forecast data includes forecasted generation data and forecasted load data, and the real-time data includes the status parameters of the energy storage device and the connection topology of each distributed power source. The first optimization module is used to establish an optimized scheduling model for the distribution network based on the dataset, with the goal of maximizing the absorption of new energy and considering the constraints of each energy storage device. The constraints include line safety constraints and upper and lower limits of SOC and daily SOC clearing constraints for each energy storage device. The solution module is used to solve the optimized scheduling model to obtain the power generation plan and energy storage charging and discharging plan of each distributed power source. The second optimization module is used to activate a preset emergency control strategy when the real-time voltage of the distribution network is detected to exceed the limit. The module uses a recursive least squares algorithm to adjust the reactive power output of each distributed power source based on the connection topology of each distributed power source so that the voltage is restored to a safe range.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory storing machine-executable instructions that can be executed by the processor, the processor executing the machine-executable instructions to implement the distribution network two-layer optimized scheduling method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement the distribution network two-layer optimization scheduling method according to any one of claims 1 to 7.

Images