Operation optimization method and system for network construction type photovoltaic distribution and storage system

Abstract

The invention relates to the technical field of operation optimization of a power supply and distribution system, and discloses an operation optimization method and system of a network construction type photovoltaic distribution and storage system, and the method comprises the steps: responding to a power-on instruction, carrying out the decoupling of hardware in a distribution and storage process, analyzing the energy reserve and electric energy supply capability, and obtaining a first operation parameter set; analyzing an energy demand and an energy storage condition through data fusion, performing dynamic partitioning on the photovoltaic distribution and storage system, and predicting a network construction capability state to obtain a dynamic partitioning instruction and a first network construction capability state; analyzing the real-time electric quantity state of each region, and performing region smooth transition processing to obtain a functional region set; dynamically scheduling different functional area combinations to carry out energy distribution and storage, generating corresponding control instructions, and obtaining a first control instruction set; controlling the operation of the photovoltaic distribution and storage system, and optimizing the operation of the photovoltaic distribution and storage system; all the batteries can work in a better state, the effective networking capacity of the system is improved, and the overall service life of the battery pack is prolonged.

Description

Technical Field

[0001] This application relates to the field of power supply and distribution system operation optimization technology, and more specifically to a method and system for optimizing the operation of a grid-type photovoltaic power distribution and storage system. Background Technology

[0002] Currently, the penetration rate of new energy sources, represented by photovoltaics, in the power system is continuously increasing. Traditional photovoltaic power generation systems mostly adopt grid-connected control, relying on the traditional power grid to provide stable voltage and frequency references. However, in scenarios with a high proportion of new energy integration, weak grids, or islanded operation, the strength and inertia of the power grid decrease, and traditional grid-connected power sources cannot provide the necessary voltage and frequency support, which may even lead to system instability. At this time, it is necessary to use grid-connected photovoltaic distribution and energy storage systems. Photovoltaic distribution and energy storage systems simulate the characteristics of synchronous generators and actively provide voltage and frequency support to the grid. However, the existing grid-connected photovoltaic distribution and energy storage systems have low utilization rates of energy storage resources and are difficult to cope with complex operating conditions.

[0003] Existing technologies suffer from the following problems: they treat the battery energy storage system as a whole for scheduling, ignoring the inconsistencies in state of charge, health, and internal resistance among the internal battery packs, leading to some batteries being overcharged, over-discharged, or remaining in extreme states for extended periods, while other batteries are underutilized; they employ predefined, fixed charge and discharge logic, lacking online dynamic adjustment capabilities, failing to optimally allocate power, resulting in decreased dynamic response performance and affecting system stability; and their assessment of the system's grid-connected capability is mostly based on rated parameters or the state at a certain static moment, failing to make dynamic and accurate predictions based on the real-time health and consistency of the energy storage system during operation, leading to a mismatch between the system's support capacity and the actual available capacity, posing operational risks. To address at least one of the above problems, this application proposes an operation optimization method and system for a grid-connected photovoltaic energy storage system. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the purpose of this application is to provide a method and system for optimizing the operation of a grid-type photovoltaic power generation and energy storage system, which can effectively solve the problems in the background technology. The specific technical solution of this application is as follows: A method for optimizing the operation of a grid-connected photovoltaic power generation and energy storage system includes: In response to the power-on command, the hardware of the energy allocation and storage process is decoupled based on the pre-built energy allocation and storage database, and the energy storage and power supply capacity are analyzed to obtain the first set of operating parameters; Based on the real-time collected photovoltaic power distribution and storage system status data and the first set of operating parameters, the energy demand and energy storage status are analyzed through data fusion, the photovoltaic power distribution and storage system is dynamically partitioned, and the grid construction capacity status is predicted to obtain the dynamic partitioning command and the first grid construction capacity status. Analyze the real-time power status of each region according to the dynamic partitioning instructions, and perform smooth transition processing of each region by controlling the charging or discharging of the energy storage battery in each region to obtain a set of functional areas. By combining the first grid capacity status and the set of functional zones, different combinations of functional zones are dynamically scheduled for energy allocation and storage, generating corresponding control commands to obtain the first set of control commands. The photovoltaic power generation and storage system is controlled according to the first set of control instructions. During operation, real-time operation status data is collected, and the first set of control instructions is optimized through a preset instruction optimization model to optimize the operation of the photovoltaic power generation and storage system.

[0005] Specifically, in response to the power-on command, the hardware of the energy distribution and storage process is decoupled based on a pre-built energy distribution and storage database to analyze energy reserves and power supply capacity, thereby obtaining a first set of operating parameters, including: In response to the power-on command, the hardware status of the allocation process is analyzed based on the pre-built allocation database, the hardware is decoupled, and the initial hardware state is obtained. By combining the initial state of the hardware with the real-time state data of the corresponding energy storage hardware, the corresponding energy reserves and power supply capacity are analyzed to obtain the first grid capacity value. The first network capacity value is mapped to the corresponding operating parameters in the distribution and storage database to obtain the first set of operating parameters.

[0006] Specifically, the first network capacity value is mapped to corresponding operating parameters in the distribution and storage database to obtain a first set of operating parameters, including: Based on the first network capacity value and the rated parameters of the supercapacitor in the corresponding hardware, analyze the charging voltage variation of the supercapacitor and set the first voltage value. The first voltage value is matched in the energy storage database to obtain the maximum rate of change of photovoltaic power under the corresponding voltage. By combining the maximum rate of change of photovoltaic power and the first grid capacity value, the corresponding operating parameters are calculated to obtain the first set of operating parameters.

[0007] Specifically, based on the real-time collected photovoltaic power distribution and storage system status data and the first set of operating parameters, the energy demand and energy storage status are analyzed through data fusion to dynamically partition the photovoltaic power distribution and storage system, predict the grid connection capability status, and obtain dynamic partitioning instructions and the first grid connection capability status, including: Based on the analysis of the first set of operating parameters, an energy demand vector is constructed. By combining real-time collected photovoltaic power distribution and storage system status data and energy demand vector, the energy storage situation is analyzed, capacity constraints are constructed, and the photovoltaic power distribution and storage system is dynamically partitioned to obtain dynamic partitioning instructions; Based on the dynamic zoning instructions, the consistency of energy storage conditions in each region is analyzed, the grid construction capacity status is predicted, and the first grid construction capacity status is obtained.

[0008] Specifically, the process involves combining real-time collected photovoltaic power distribution and storage system status data and energy demand vectors to analyze energy storage conditions, construct capacity constraints, dynamically partition the photovoltaic power distribution and storage system, and obtain dynamic partitioning instructions, including: Based on the energy demand vector, the corresponding energy demand is analyzed and the corresponding priority is calculated to obtain the corresponding energy demand capacity. By combining the energy demand capacity and the real-time collected photovoltaic power distribution and storage system status data, the total state of charge, health status and supercapacitor voltage status of the energy storage battery pack are analyzed, and the corresponding safety boundary and energy supply upper limit are constructed to obtain the maximum energy supply value. By comparing energy demand capacity and maximum energy supply value, the energy supply area and reserve area that meet the energy demand capacity are divided from the maximum energy supply value; By combining the power supply area and the backup area, the consistency status of individual cells within the battery pack in different areas is analyzed, and the area boundaries are adjusted and optimized to obtain dynamic partitioning instructions.

[0009] Specifically, the step of analyzing the consistency of energy storage conditions in each region based on dynamic zoning instructions, predicting the grid capacity status, and obtaining the first grid capacity status includes: Based on the dynamic zoning instructions and the real-time collected total state of charge of the energy storage battery pack and the voltage data of each individual battery cell, the first capacity value of each region is calculated. For each region, the voltage difference between individual cells is calculated, and the first capacity value is corrected and optimized using a preset capacity optimization model to obtain the second capacity value. By combining the second capacity value and the voltage state of the supercapacitor, the network construction capability state is predicted through a preset network construction capability analysis model, and the first network construction capability state is obtained.

[0010] Specifically, the process involves analyzing the real-time power status of each region according to dynamic partitioning instructions, and performing smooth transition processing by controlling the charging or discharging of the energy storage batteries in each region to obtain a set of functional zones, including: The real-time power status of each region is analyzed according to the dynamic partitioning instructions, and the difference between the target power value and the current power value of each region is calculated to obtain the replenished energy. The direction of energy transfer is determined by the replenished energy, and the charging or discharging of the energy storage battery in each region is controlled according to the direction of energy transfer to achieve a smooth transition between regions, thus obtaining a set of functional areas.

[0011] Specifically, by combining the first network capacity status and the set of functional zones, different combinations of functional zones are dynamically scheduled for energy allocation and storage, generating corresponding control commands to obtain the first set of control commands, including: Based on the first network capability status and functional area set, decompose the requirements and calculate the corresponding priorities to construct a requirement task sequence; According to the sequence of required tasks, analyze the physical characteristics and real-time status of each functional area, dynamically schedule different combinations of functional areas for energy allocation and storage, generate corresponding control commands, and obtain the first set of control commands.

[0012] Specifically, according to the demand task sequence, the physical characteristics and real-time status of each functional area are analyzed, different combinations of functional areas are dynamically scheduled for energy allocation and storage, and corresponding control commands are generated to obtain the first set of control commands, including: Based on the sequence of required tasks, analyze the physical characteristics of each functional area, match the corresponding required tasks for each functional area, and obtain the first task allocation result. Analyze the real-time status of each functional area, filter out the demand tasks with inconsistent timing between functional areas in the first task allocation result, and reallocate them to obtain the second task allocation result; Based on the second task allocation result, the corresponding control instructions are generated to obtain the first set of control instructions.

[0013] A grid-connected photovoltaic (PV) power grid and energy storage system operation optimization system is provided to implement the aforementioned grid-connected PV power grid and energy storage system operation optimization method, comprising: The operation parameter analysis module, in response to the power-on command, decouples the hardware of the energy distribution and storage process based on the pre-built energy distribution and storage database, analyzes the energy storage and power supply capacity, and obtains the first set of operation parameters; The dynamic partitioning module, based on the real-time collected photovoltaic power distribution and storage system status data and the first set of operating parameters, analyzes energy demand and energy storage status through data fusion, dynamically partitions the photovoltaic power distribution and storage system, predicts the grid construction capacity status, and obtains dynamic partitioning instructions and the first grid construction capacity status. The regional optimization module analyzes the real-time power status of each region according to the dynamic partitioning instructions, and performs smooth transition processing of the region by controlling the charging or discharging of the energy storage battery in each region, thus obtaining a set of functional areas. The control command generation module, in conjunction with the first network capacity status and the set of functional zones, dynamically schedules different combinations of functional zones for energy allocation and storage, generates corresponding control commands, and obtains the first set of control commands. The control command optimization module controls the operation of the photovoltaic power generation and storage system according to the first set of control commands. During operation, it collects real-time operation status data and optimizes the first set of control commands through a preset command optimization model to optimize the operation of the photovoltaic power generation and storage system.

[0014] The beneficial effects of this application are as follows: During the power-on initialization phase, by decoupling the energy storage hardware and combining it with real-time status, the actual operating parameters under the current operating conditions are mapped from a pre-built database, providing an accurate initial benchmark for subsequent optimization; multi-source data is fused to construct the system's power supply safety boundary and upper limit in real time, dynamically dividing the power supply area and the backup area, and optimizing the area boundary based on the consistency status of individual batteries; the system's grid-building capability is analyzed by combining individual battery voltage consistency indicators, enabling accurate prediction of the system's real-time and actual grid-building capability, and providing accurate capability data; grid-building requirements are decomposed into task sequences, the physical characteristics and real-time status of each functional area are analyzed, dynamic task matching and reallocation are performed, optimized scheduling is achieved, and real-time feedback closed-loop optimization is carried out during operation. Through dynamic partitioning and consistency management, all individual batteries can operate in a better and more balanced state, improving the effective grid-building capacity of the system and extending the overall battery pack life; through dynamic task allocation, the system can quickly respond to dynamic environments, improving operational safety and robustness. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the operation optimization method for a grid-type photovoltaic power distribution and storage system according to Embodiment 1 of this application; Figure 2 This is a schematic diagram of the power supply area and the backup area in Embodiment 1 of this application; Figure 3 This is a flowchart of the first network capability status analysis process in Embodiment 1 of this application; Figure 4 This is a schematic diagram of the operation optimization system of a grid-type photovoltaic power distribution and storage system in Embodiment 1 of this application. Detailed Implementation

[0016] The present application will be further described in detail below with reference to the accompanying drawings and embodiments.

[0017] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0018] Hereinafter, the terms "first," "second," and other generic terms are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0019] Example 1 refer to Figure 1 The image shows a specific implementation of a grid-type photovoltaic power generation and storage system operation optimization method according to this application, including: S101. In response to the power-on command, the hardware of the energy distribution and storage process is decoupled based on the pre-built energy distribution and storage database, and the energy storage and power supply capacity are analyzed to obtain the first set of operating parameters. S102. Based on the real-time collected photovoltaic power distribution and storage system status data and the first set of operating parameters, the energy demand and energy storage status are analyzed through data fusion, the photovoltaic power distribution and storage system is dynamically partitioned, and the grid construction capacity status is predicted to obtain the dynamic partitioning instruction and the first grid construction capacity status. S103. Analyze the real-time power status of each region according to the dynamic partitioning instructions, and perform smooth transition processing of the region by controlling the charging or discharging of the energy storage battery in each region to obtain a set of functional areas. S104. Combining the first network capacity status and the set of functional zones, dynamically schedule different combinations of functional zones for energy allocation and storage, generate corresponding control commands, and obtain the first set of control commands. S105. Control the operation of the photovoltaic power generation and storage system according to the first set of control instructions. During the operation, collect the operation status data in real time and optimize the first set of control instructions through a preset instruction optimization model to optimize the operation of the photovoltaic power generation and storage system.

[0020] With the large-scale integration of renewable energy sources, primarily photovoltaics, into the power grid, the power system faces challenges related to inertia and insufficient voltage support capabilities. Grid-based technologies require power sources to actively establish grid voltage and frequency, similar to traditional generators, making them crucial for building new power systems. However, existing grid-based photovoltaic distribution and storage systems generally treat energy storage units as a whole and manage them in a rudimentary manner, failing to address the performance differences within the internal battery packs, photovoltaic fluctuations, and complex grid conditions. This results in reduced effective support capabilities, shortened asset lifespan, and suboptimal dynamic response.

[0021] In this embodiment, after responding to the power-on command, the system accesses a pre-built energy allocation and storage database. This database stores historical performance curves, health status degradation, and parameter mapping relationships under different operating conditions for various hardware components. These hardware components include, but are not limited to, lithium-ion battery clusters, supercapacitor modules, and power converters. The hardware in the energy allocation and storage process is decoupled, and real-time status data reported by each independent hardware unit is read sequentially via the communication bus. This data includes, but is not limited to, the initial voltage and temperature of each battery pack, the current voltage of the supercapacitor pack, and the available status of the power converter. Combining the real-time status data with the characteristics of the corresponding hardware in the database, the system's actual energy reserves and instantaneous power supply capacity are analyzed. The analysis results are quantified into specific parameter values, including but not limited to the system's currently available total energy value, maximum sustainable output power, and upper limit of the power rapid adjustment rate, resulting in a first set of operating parameters.

[0022] It should be noted that by analyzing the first operating parameters, an operating benchmark that accurately reflects the actual health status and initial conditions of the hardware is established during the startup phase. This avoids the initial decision-making bias caused by directly using the rated parameters in traditional methods, and provides an accurate input basis for subsequent optimization decisions. It avoids the risk of overcharging or over-discharging or misjudgment of capabilities caused by inaccurate parameters from the source, thereby improving the accuracy of the overall optimization process and the security of system startup.

[0023] Specifically, the system continuously collects status data, including but not limited to the output power of the photovoltaic array, the voltage and frequency of the grid connection point, and the load demand power. Based on the status data and the first set of operating parameters, it analyzes the comprehensive energy demand and the energy storage status of the energy storage units at the current moment through data fusion analysis. Based on the fused information, it constructs capacity constraints, for example, calculating the upper and lower limits of the available discharge capacity of each battery cluster. The photovoltaic distribution and storage system is dynamically partitioned, dividing the entire energy storage system into a power supply area and a backup area. The power supply area is used to respond immediately to current demand, and the backup area is used to cope with future fluctuations or failures. Based on the dynamic partitioning results, the effective capacity value of each area is analyzed in combination with the voltage status of the supercapacitor to predict the grid construction capability status, and obtain the dynamic partitioning command and the first grid construction capability status.

[0024] For example, the network capability prediction process uses a three-layer feedforward neural network. Its training dataset comes from system response data under different battery states, supercapacitor voltages, and grid disturbance scenarios. During training, the TensorFlow framework is used. The input layer has 9 neurons, corresponding to the SOC, SOH, and temperature of the three battery clusters, as well as the supercapacitor voltage and current power demand. The hidden layer has 15 neurons with ReLU activation. The output layer has 3 neurons, representing the predicted inertia support time constant, primary frequency regulation capacity, and voltage regulation capacity, respectively. The loss function is mean squared error, the optimizer is Adam, the initial learning rate is 0.001, the batch size is 32, and training is conducted for 500 epochs until convergence.

[0025] It is important to emphasize that dynamic partitioning can avoid the limitations of fixed grouping, ensuring that the organization of system resources always optimally matches real-time demands and internal states. Through network capability prediction based on consistency correction, the system can dynamically assess its own ability to support the power grid, providing accurate data support for subsequent scheduling decisions and improving the system's ability to cope with uncertainties and its operational safety.

[0026] Furthermore, the dynamic zoning instructions clearly define the target energy status of each region. The real-time energy status of each region is analyzed according to the dynamic zoning instructions, and the difference between the target energy value and the current energy value is calculated to obtain the supplementary energy that needs to be transferred. The direction of energy transfer is determined based on the sign of the supplementary energy, where positive indicates charging and negative indicates discharging. By controlling the charging or discharging process of the energy storage batteries in each region, a smooth transition is achieved, ensuring stable energy changes across the entire region and avoiding sudden power fluctuations, thus obtaining a set of functional zones. Through proactive and controlled smooth transition processing, the internal units are already at the expected optimized operating point before the system executes specific network construction tasks. This eliminates switching shocks or response delays caused by energy imbalances between or within regions, providing a stable internal energy foundation for subsequent precise scheduling. It also reduces the stress on batteries caused by frequent high-power charging and discharging, helping to extend equipment lifespan.

[0027] Specifically, based on the first grid capacity status and functional area set, the received external energy distribution and storage demands are decomposed. These external energy distribution and storage demands include, but are not limited to, frequency regulation instructions issued by the grid dispatching agency and internal demands for smoothing photovoltaic fluctuations. The demands are decomposed into tasks with timestamps and priority tags. According to the demand task sequence, the physical characteristics and real-time status of each functional area are analyzed. Initial matching is performed based on the physical characteristics, and the temporal coordination is analyzed. The current allocation scheme is simulated. If task conflicts or resource competition are found between functional areas, the tasks are reallocated to obtain the second task allocation result. Based on the second task allocation result, specific, time-synchronized control instructions are generated to form the first control instruction set.

[0028] It should be noted that by deconstructing the network construction task and scheduling tasks based on the characteristics and status of functional areas, the system can be ensured to respond to internal and external demands in the most efficient and coordinated manner, avoiding command conflicts and resource consumption, giving full play to the advantages of different technical routes in energy storage systems, and improving the overall efficiency and coordination of the system in executing complex network construction tasks.

[0029] Specifically, the system issues a first set of control commands to each underlying power converter and battery management system for execution, controlling the actual operation of the photovoltaic power generation and energy storage system. During operation, it collects real-time operating status data through a sensor network, including but not limited to the actual output power of each functional area, DC bus voltage, AC grid connection point voltage and frequency, and key component temperature. The real-time data and the first set of control commands are input into a preset command optimization model. This model includes, but is not limited to, a linear optimization model pre-trained using a large amount of historical operating status data. The model contains state-space equations describing the relationship between the power response and voltage-frequency of the energy storage system. The model inputs are current state measurements and historical command sequences, and the output is the adjustment amount for the currently executing first set of control commands. For example, the model continuously optimizes the command sequence for the next few control cycles, aiming to minimize the frequency deviation integral or bus voltage fluctuation. Combined with the operating constraints of each energy storage unit, it calculates the correction value for the power commands of each functional area in the next cycle. The correction value is then superimposed on the original command value to obtain the optimized command, continuously optimizing the operation of the photovoltaic power generation and energy storage system online.

[0030] For example, the state variables of the instruction optimization model include frequency deviation and its rate of change, DC voltage, etc., and the control input is the power instruction adjustment amount of each functional area. The model parameter matrix includes applying power step disturbances at different operating points, recording the dynamic response of the system frequency and voltage, and fitting the coefficients of the state space equation using the least squares method. In each control cycle (e.g., 100ms), the MPC controller uses the current measurement value as the initial state and uses the state space model to predict the system output trajectory under different control input sequences in the future period (e.g., 2 seconds, i.e., 20 cycles). The constrained quadratic programming problem is solved, and the objective function is to minimize the deviation between the predicted trajectory and the desired trajectory (e.g., the frequency deviation is zero), while minimizing the drastic changes in the control quantity. The constraints include the upper and lower limits of the power of each functional area.

[0031] It should be noted that through the instruction optimization process based on real-time feedback, online self-learning and adaptive adjustment can be performed, which can effectively overcome the impact of uncertainties such as model mismatch and external disturbances, dynamically correct the control strategy, and ensure that the system always runs towards the optimal operating trajectory. This can improve the stability and control accuracy of the system under various complex working conditions, enhance the robustness to changes in equipment status and external environmental interference, and improve system performance.

[0032] During the power-on initialization phase, this application decouples and analyzes the energy storage hardware, combining real-time status data to map actual operating parameters under current conditions from a pre-built database, providing an accurate initial benchmark for subsequent optimization. It integrates multi-source data to construct real-time safety boundaries and upper limits for system power supply, dynamically dividing power supply and backup zones, and optimizing zone boundaries based on the consistency status of individual cells. By analyzing the system's grid-connection capability using individual cell voltage consistency indicators, it achieves accurate prediction of the system's real-time and actual grid-connection capability, providing accurate capability data. The grid-connection requirements are decomposed into task sequences, analyzing the physical characteristics and real-time status of each functional area, performing dynamic task matching and reallocation to achieve optimized scheduling, and conducting real-time feedback closed-loop optimization during operation. Dynamic partitioning and consistency management enable all individual cells to operate in a more optimal and balanced state, improving the system's effective grid-connection capacity and extending the overall battery pack lifespan. Dynamic task allocation allows the system to quickly respond to dynamic environments, improving operational safety and robustness.

[0033] Furthermore, in response to the power-on command, the hardware of the energy distribution and storage process is decoupled based on a pre-built energy distribution and storage database, and the energy storage and power supply capacity are analyzed to obtain a first set of operating parameters, including: S201. In response to the power-on command, analyze the hardware status of the allocation process based on the pre-built allocation database, decouple the hardware, and obtain the initial hardware state. S202. Combining the initial state of the hardware and the real-time state data of the corresponding energy storage hardware, analyze the corresponding energy reserves and power supply capacity to obtain the first grid capacity value. S203. Map the first network capacity value to the corresponding operating parameters in the storage and distribution database to obtain the first set of operating parameters.

[0034] In this embodiment, based on the power-on command, the hardware status of the energy storage process is analyzed in a pre-built energy storage database. The energy storage database stores the topological connection relationships, type identifiers, rated parameters, and historical performance calibration coefficients of all energy storage hardware in the system. Through the internal communication network of the system, status query commands are sent sequentially to the local management units of each hardware unit according to the hardware identifier list in the energy storage database. Each management unit returns the initial status data of the hardware. For battery packs, the initial status data includes, but is not limited to, the pack terminal voltage, average temperature, and insulation resistance; for supercapacitor packs, the initial status data includes, but is not limited to, the pack terminal voltage; for power converters, the initial status data includes, but is not limited to, the self-test status and the DC-side capacitor pre-charge voltage, thus obtaining the hardware initial status set of each hardware unit at the moment of power-on.

[0035] It should be noted that by actively decoupling and collecting the independent initial states of each hardware component, accurate data support is provided for subsequent analysis. This captures changes in hardware state during system inactivity or due to maintenance, ensuring the accuracy of the optimization starting point and reducing the risk of making incorrect judgments based on outdated or aggregated information.

[0036] Specifically, combining the initial hardware state set, the system synchronously reads some real-time state data of the corresponding energy storage hardware. For energy reserve analysis, taking a battery cluster as an example, based on the initial terminal voltage and temperature, it queries the voltage, remaining capacity comparison table, and temperature compensation coefficient of that battery model in the database to estimate the current available energy of the cluster. For supercapacitor banks, it directly calculates the currently stored electrical energy based on the initial voltage and known capacity. For power supply capacity analysis, it calculates the instantaneous maximum discharge power based on the voltage of the supercapacitor and its equivalent series resistance. The energy reserve is obtained by weighted summing of the available energy of all battery clusters and the energy of the supercapacitors, with the weights set according to the system's power supply accuracy requirements. The power supply capacity is obtained by selecting the smaller value between the rated power of all power converters and the maximum discharge power that the current energy storage unit can support. The average value of the energy reserve and power supply capacity is calculated to obtain the first grid capacity value.

[0037] It should be noted that by fusing and analyzing multi-dimensional information, the corresponding energy reserves and power supply capacity can be accurately calculated. The resulting first grid capacity value can accurately reflect the resource boundary of the system at the moment of power-on, providing accurate constraints for the formulation of subsequent operation strategies and ensuring that optimization decisions always operate within the safe range of the system's physical capabilities.

[0038] Specifically, the first network capacity value is mapped to corresponding operating parameters in the resource allocation database, resulting in a first set of operating parameters. Through predefined mapping rules, the operating parameters can be dynamically adjusted based on changes in the system's own resource status, ensuring that the system's control behavior matches the current actual physical capabilities. This allows for the full release of the system's performance potential while ensuring safety. When system capacity is sufficient, more aggressive control parameters are used to improve dynamic performance; when capacity is limited, more conservative parameters are automatically switched to ensure stability, thus enhancing the system's safety and dynamic adaptability.

[0039] Furthermore, the first network capacity value is mapped to corresponding operating parameters in the distribution and storage database to obtain a first set of operating parameters, including: S301. Based on the first network capacity value and the rated parameters of the supercapacitor in the corresponding hardware, analyze the charging voltage variation of the supercapacitor and set the first voltage value. S302. Match the first voltage value in the energy storage database to map the maximum rate of change of photovoltaic power under the corresponding voltage. S303. Combining the maximum rate of change of photovoltaic power and the first grid capacity value, calculate the corresponding operating parameters to obtain the first set of operating parameters.

[0040] In this embodiment, based on the first network capacity value, the rated parameters of the corresponding supercapacitor hardware are obtained from pre-stored parameters, including but not limited to total capacity and rated operating voltage range. Based on the system power requirements, the corresponding voltage value is calculated using the relationship between instantaneous energy, voltage, and equivalent series resistance in the supercapacitor's physical characteristics. The charging voltage variation of the supercapacitor is analyzed, and the maximum and minimum voltage values ​​during the variation process are set as corresponding voltage ranges. The average of the maximum and minimum voltage values ​​is calculated and set as the first voltage value. By calculating the first voltage value, a voltage reference value matching the current overall network capacity target can be set for the supercapacitor. This avoids the problem of insufficient performance reserves or low energy utilization caused by the supercapacitor operating at a fixed single voltage value in traditional methods. It sets the supercapacitor's energy state to the optimal position, ensuring rapid response capability and energy reserves when dealing with dynamic tasks such as power surges and providing inertia support.

[0041] Specifically, the energy storage database includes a mapping relationship between the voltage value of the supercapacitor and the maximum rate of change of photovoltaic power. This mapping relationship is established based on simulation or historical operating data analysis during the system design phase. In establishing this relationship, the control characteristics of the photovoltaic inverter, the DC bus capacitance, the response speed of the supercapacitor, and the thermal management and electrical stress limitations of the entire system are comprehensively considered. Based on the first voltage value, the maximum rate of change of photovoltaic power at the corresponding voltage is calculated according to the mapping relationship. By calculating the maximum rate of change of photovoltaic power, the rate of change limit can be dynamically adjusted according to the current instantaneous buffering capacity of the energy storage system. When the system's buffering capacity is sufficient, the limit can be appropriately relaxed, allowing for a higher rate of change; when the buffering capacity is insufficient, the limit can be tightened to prevent system instability caused by sudden increases or decreases in photovoltaic power exceeding the energy storage system's smoothing capacity, thereby improving system stability.

[0042] Specifically, by combining the maximum rate of change of photovoltaic power and the first grid capacity value, corresponding operating parameters are calculated. These operating parameters include, but are not limited to, power loop controller parameters and energy management threshold parameters. Based on the system's maximum continuous power and the system's inertial time constant, the proportional and integral coefficients of the power loop are calculated to obtain the power loop controller parameters. Based on the first grid capacity value and the maximum rate of change of photovoltaic power, energy management threshold parameters corresponding to different energy dispatching are set according to the system's energy allocation and storage accuracy requirements. The operating parameters are integrated to obtain the first set of operating parameters. By calculating the first set of operating parameters, control parameters that match the system capacity are obtained. This allows for operation based on real-time assessment of the system's capacity boundaries, dynamic optimization of system parameters, and ensures that the system's dynamic response characteristics always match the actual state of the internal physical resources. Thus, under the premise of ensuring stable operation, the grid support performance of the system is improved across the entire operating range.

[0043] Furthermore, based on the real-time collected photovoltaic power distribution and storage system status data and the first set of operating parameters, energy demand and energy storage status are analyzed through data fusion. The photovoltaic power distribution and storage system is then dynamically partitioned, and the grid connection capability status is predicted, resulting in dynamic partitioning instructions and the first grid connection capability status, including: S401. Analyze the energy demand based on the first set of operating parameters and construct an energy demand vector; S402. Combining the real-time collected photovoltaic power distribution and storage system status data and energy demand vector, analyze the energy storage situation, construct capacity constraints, dynamically partition the photovoltaic power distribution and storage system, and obtain dynamic partitioning instructions. S403. Analyze the consistency of energy storage in each region according to the dynamic zoning instructions, predict the grid construction capacity status, and obtain the first grid construction capacity status.

[0044] In this embodiment, based on the parameters related to the operating mode and external commands in the first set of operating parameters, and combined with the real-time acquired external input signals, the energy demand is analyzed. This includes real-time active / reactive power commands or grid frequency regulation demands issued by the external dispatch system, load power supply demands, and photovoltaic fluctuation mitigation demands. Grid frequency regulation demands are high-priority instantaneous power demands at the second to minute level, load power supply demands are continuous basic power demands, and photovoltaic fluctuation mitigation demands are energy throughput demands at the minute to hour level. The corresponding parsed demands are integrated sequentially to obtain an energy demand vector. By constructing the energy demand vector, accurate data input is provided for the subsequent optimization process, enabling coordinated processing of multiple demands and avoiding resource scheduling chaos caused by ambiguous demands.

[0045] Specifically, by combining real-time collected photovoltaic power distribution and storage system status data and energy demand vectors, the energy storage situation is analyzed through data fusion, capacity constraints are constructed, and the photovoltaic power distribution and storage system is dynamically partitioned to obtain dynamic partitioning instructions. Dynamic partitioning can break the fixed nature of physical connections and dynamically construct functionally different areas according to real-time demand and resource status, thereby improving resource utilization and task execution efficiency.

[0046] Specifically, based on the dynamic zoning instructions, the consistency of energy storage conditions in each region is analyzed to predict the grid capacity status and obtain the first grid capacity status. By analyzing the consistency of individual batteries within a zone, the dispatchable capacity of the energy storage system can be accurately predicted, effectively avoiding the problem of available capacity being less than nominal capacity due to inconsistencies within battery packs. Grid interaction or internal scheduling based on the prediction results can ensure that the system's support capacity matches its actual deliverable capacity, improving the controllability and reliability of system operation.

[0047] Furthermore, by combining real-time collected photovoltaic power distribution and energy storage system status data and energy demand vectors, the energy storage situation is analyzed, capacity constraints are constructed, and the photovoltaic power distribution and energy storage system is dynamically partitioned to obtain dynamic partitioning instructions, including: S501. Based on the energy demand vector, analyze the corresponding energy demand and calculate the corresponding priority to obtain the corresponding energy demand capacity. S502. Combining the energy demand capacity and the real-time collected photovoltaic power distribution and storage system status data, analyze the total state of charge, health status and supercapacitor voltage status of the energy storage battery pack, construct the corresponding safety boundary and energy supply upper limit, and obtain the maximum energy supply value. S503. Compare the energy demand capacity and the maximum energy supply value, and divide the energy supply area and the reserve area that meet the energy demand capacity from the maximum energy supply value. S504. Combining the power supply area and the backup area, analyze the consistency status of individual cells in the battery pack in different areas, adjust and optimize the area boundaries, and obtain dynamic partitioning instructions.

[0048] In this embodiment, based on the energy demand vector, the type identifier of each component in the vector is analyzed, including but not limited to frequency regulation, load tracking, planned charging and discharging, and economic optimization. A pre-set priority rule base is invoked for parsing. The priority rule base is established based on power grid operation procedures, system safety criteria, and equipment protection strategies. For example, frequency regulation requirements directly related to power grid stability are given the highest priority, followed by requirements to ensure power supply to critical loads, and economic optimization requirements have the lowest priority. The priorities are fine-tuned based on real-time data. When a power grid frequency deviation is detected to exceed a specific threshold, the priority of the frequency regulation requirement is automatically increased. After parsing out the type and priority of each requirement, the corresponding requirement is quantified into a specific energy demand capacity. For example, for the highest priority frequency regulation requirement, the energy demand capacity includes a continuously adjustable power capacity of ±X kW and a supporting energy of Y kWh required within the next 30 seconds. The specific figures for X and Y are calculated based on the actual demand.

[0049] It should be noted that by analyzing and prioritizing energy demand vectors, demand instructions are transformed into specific capacity tasks, providing accurate data support for subsequent resource allocation. This enables the rapid formulation of safe operational decisions when resources are limited or demand conflicts occur, ensuring that high-priority demands are prioritized and improving the reliability and security of system operation.

[0050] Specifically, the system collects real-time aggregated data from all energy storage battery packs, individual battery data, and voltage and temperature data from the supercapacitor bank. Battery pack data includes, but is not limited to, total state of charge (SBC), health status reflecting capacity degradation, and temperature. This real-time data is input into a pre-defined energy storage state analysis model. This model includes, but is not limited to, a neural network model pre-trained using extensive historical operating data. The model's input consists of various real-time energy storage state data, and its output is an assessment of the current available energy storage capacity of the battery packs and supercapacitors. The relationship between input and output is established through pre-defined state-capacity mapping rules. The model incorporates energy storage capacity degradation coefficients corresponding to different SBCs, battery health statuses, and supercapacitor voltages. When a battery's health status falls below a pre-defined health threshold, its available energy storage capacity is multiplied by the corresponding degradation coefficient to reflect the impact of battery aging on energy storage capacity. The impact of force is considered. Based on the output of the energy storage state analysis model, and combined with the physical characteristics and operating specifications of the energy storage components, a safety boundary and an upper limit for energy supply are constructed. The maximum charge / discharge voltage range and maximum charge / discharge current threshold of the energy storage battery pack are determined according to the manufacturer's parameters. The safe voltage range of the supercapacitor is determined according to its rated parameters. These parameters serve as constraints for safe operation. Under the constraints of the safety boundary, and combining the available capacity corresponding to the total SOC of the energy storage battery pack and the instantaneous power supply capacity corresponding to the voltage of the supercapacitor, the maximum energy output scale that the system can stably output under the current state is calculated through energy balance calculations. This maximum energy supply value is then used as the upper limit for energy supply. In the specific calculation process, loss factors such as the charge / discharge efficiency of the battery pack and the energy conversion efficiency of the supercapacitor need to be considered. For example, by multiplying the available capacity of the battery pack by the charge / discharge efficiency coefficient, the actual output battery energy scale is obtained. This is then superimposed with the output energy scale of the supercapacitor to obtain the maximum energy supply value.

[0051] For example, the energy storage state analysis model maps multi-source, heterogeneous real-time state data into an assessment result of available energy storage capacity that can be directly used for capacity planning and scheduling decisions. The model receives a multi-dimensional feature vector, including the total state of charge, health status, average temperature, voltage range, and terminal voltage and average temperature of each battery cluster, such as [0.65, 0.92, 28.5, 015, 400, 29]. The model outputs a structured assessment result vector, including the current available capacity coefficient and current maximum sustainable discharge power coefficient of each battery cluster, and the current available energy and current maximum discharge power of the supercapacitor group. The training dataset for this model was collected from actual operational data of multiple identical energy storage power stations over a period of more than one year. This data includes all input features and their corresponding actual available capacity and actual safe power limits as label data. The model was trained using supervised learning, specifically using the TensorFlow framework. A fully connected feedforward neural network was constructed, including an input layer with 5 battery cluster features and 2 supercapacitor features. The two hidden layers have 32 and 16 neurons respectively, using ReLU as the activation function. The output layer dimension matches the output vector, and the Sigmoid function is used to ensure the output is within a reasonable range. The loss function is a weighted sum of mean squared error loss and mean absolute error loss. Training parameters included an initial learning rate of 0.001 for the Adam optimizer, a batch size of 64 based on the dataset size, and the dataset being randomly divided into 70% (training set), 15% (validation set), and 15% (test set).

[0052] Accurate assessment of the available energy storage capacity of energy storage components based on real-time collected system status data avoids deviations in energy supply capacity assessment caused by factors such as component aging and environmental impact, ensuring the accuracy of energy supply capacity analysis. By constructing safety boundaries based on the physical characteristics of energy storage components, subsequent energy storage operations in zoned operation can be effectively constrained, preventing damage to energy storage components due to overcharging / discharging, voltage exceeding limits, etc., thus improving the safety and stability of system operation. The maximum energy supply value derived from the safety boundaries and energy storage capacity assessment accurately reflects the current stable energy supply capacity of the system, providing accurate data support for the subsequent division of energy supply and backup zones, and avoiding insufficient energy supply or resource waste due to overestimation or underestimation of energy supply capacity.

[0053] Specifically, the energy demand capacity and the maximum energy supply value are input into a preset supply and demand comparison model. The input of the supply and demand comparison model is the energy demand capacity and the maximum energy supply value, and the output is the supply and demand difference and the supply and demand matching degree. The relationship between the input and output is established through difference calculation and matching quantification rules. The supply and demand difference is the difference between the maximum energy supply value and the energy demand capacity, and the supply and demand matching degree is the ratio of the energy demand capacity to the maximum energy supply value.

[0054] like Figure 2 As shown, based on the output of the supply and demand comparison model and combined with the preset zoning rules, the division range of the energy supply zone and the backup zone is determined. The energy storage capacity of the energy supply zone must be at least equal to the energy demand capacity to ensure that the core energy demand can be met. Specifically, the energy supply zone can be defined as a cluster of continuous energy storage modules that can provide no less than the energy demand capacity, based on the module division of the energy storage battery packs. The energy storage capacity of the backup zone is the difference between the maximum energy supply value and the energy storage capacity of the energy supply zone. The energy storage modules in this area must have rapid response capabilities to cope with sudden changes in energy demand, fluctuations in photovoltaic output, and other emergencies. When dividing the area, priority should be given to energy storage modules with fast charging and discharging response speeds and good individual consistency as backup zones. At the same time, key parameters such as the energy storage module number, location information, and available energy storage capacity corresponding to the energy supply zone and the backup zone are recorded to obtain preliminary zoning results, providing preliminary division results for subsequent boundary optimization.

[0055] It should be noted that dividing the energy supply zone and the backup zone based on the principle of supply and demand matching ensures that the energy storage capacity of the energy supply zone meets the core energy demand and guarantees the reliability of basic energy supply. By setting up the backup zone, redundant energy storage is provided to cope with sudden demand, which improves the system's ability to cope with fluctuations in energy demand and photovoltaic output. By clarifying the scope and key parameters of energy storage modules in the energy supply zone and the backup zone, accurate targets are provided for subsequent regional boundary optimization and operation control, enabling precise regulation of different regions and improving the orderliness and efficiency of energy storage resource scheduling.

[0056] Specifically, based on the preliminary zoning results obtained in step S503, including the energy storage module range of the power supply area and the backup area, the battery management system collects real-time voltage data, internal resistance data, charge / discharge current data, and other consistency evaluation parameters of each individual cell in the energy storage battery pack within each area. The collected consistency evaluation parameters of each individual cell are input into a preset individual cell consistency analysis model. The input to the individual cell consistency analysis model is the voltage, internal resistance, charge / discharge current, and other parameters of each individual cell within the same area. The output is the individual cell consistency evaluation result for that area, including the consistency deviation value and consistency level. The correlation between input and output is established through deviation calculation and preset level classification rules. The level classification rules can be set according to the system accuracy requirements. The consistency deviation value is obtained by calculating the standard deviation of the voltage and the coefficient of variation of the internal resistance of each individual cell within the same area. The consistency level is classified according to a preset deviation threshold. When the voltage standard deviation is less than or equal to the deviation threshold, it is determined to be a high consistency level; when the voltage standard deviation is greater than the deviation threshold, it is determined to be a low consistency level. The deviation threshold can be set according to the system's energy storage accuracy requirements.

[0057] Based on the output of the single-cell consistency analysis model, a boundary optimization algorithm is used to adjust the regional boundaries. Regions with single-cell consistency deviation values ​​exceeding the preset consistency threshold are selected. The energy storage modules containing the poorly consistent cells in these regions are moved to adjacent regions to ensure that the single-cell consistency deviation values ​​in each region are lower than the preset consistency threshold after adjustment. After the regional boundary adjustment is completed, the final energy supply area and standby area range are determined. Key parameters such as the energy storage module list, boundary coordinates, available energy storage capacity, and battery cell voltage consistency index of each region are recorded. The key parameters are integrated to generate dynamic partitioning instructions.

[0058] For example, the input to the boundary optimization algorithm includes the logical partitioning result of step S503, including a list of energy storage modules (such as battery cluster numbers) contained in the power supply area and the backup area, and their initial capacity targets. The output of the individual consistency analysis model includes the individual consistency deviation value and consistency level for each energy storage module in the initial scheme. System constraints and parameters include a preset consistency threshold. The algorithm iterates through each energy storage module in the power supply area and marks modules with consistency deviation values ​​exceeding the preset consistency threshold of the power supply area as modules to be adjusted. For each module to be adjusted, the algorithm evaluates the impact of removing it from the power supply area on the total capacity and total power of the power supply area. In the backup area, it searches for replacement modules that meet the requirements of high consistency level, capacity and power not less than 80% of the module to be adjusted, and whose capacity constraints can still be met in both areas after the exchange.

[0059] It is important to emphasize that optimizing the regional boundaries based on the consistency of individual battery cells ensures consistent operating characteristics of energy storage modules within each region, avoiding issues such as uneven charging and discharging and localized overheating caused by poor cell consistency. This improves the stability and safety of operation in each region. Boundary optimization enhances the accuracy and rationality of zoning results, fully considering the actual operating characteristics of energy storage components and avoiding operational risks associated with traditional zoning based solely on capacity. The generated dynamic zoning instructions contain detailed parameters and control requirements for each region, providing clear instructions for subsequent precise control of each region and improving the overall reliability and efficiency of the system.

[0060] like Figure 3 As shown, based on the dynamic zoning instructions, the consistency of energy storage conditions in each region is analyzed to predict the grid capacity status, resulting in the first grid capacity status, including: S601. Calculate the first capacity value for each region based on the dynamic zoning instructions and the real-time collected total state of charge of the energy storage battery pack and the voltage data of each individual battery cell. S602. For each region, calculate the voltage difference between individual cells, and use a preset capacity optimization model to correct and optimize the first capacity value to obtain the second capacity value. S603. Combining the second capacity value and the supercapacitor voltage state, the network construction capability state is predicted through a preset network construction capability analysis model to obtain the first network construction capability state.

[0061] In this embodiment, the region division information, such as the list of energy storage modules, the number of modules, and the module connection method, corresponding to each region is parsed from the dynamic partitioning instructions to clarify the composition range of energy storage units in each region. The total state of charge data of the energy storage battery packs in each region and the real-time voltage data of each individual battery cell in each region are collected synchronously by the system sensors and the battery management system. The first capacity value of each region is calculated. The input of the calculation process is the total state of charge data of each region, the voltage data of individual batteries, and the rated capacity parameters of the energy storage battery packs corresponding to each region. The output is the first capacity value of each region. The correlation between the input and output is established through capacity conversion rules. In the specific calculation process, the theoretical energy storage capacity of each region is calculated based on the rated capacity and total state of charge data of the energy storage battery packs in each region. The validity is verified by combining the collected individual battery voltage data. The energy storage module capacity corresponding to individual batteries whose voltage exceeds the preset normal voltage range is removed to obtain the first capacity value of each region.

[0062] It should be noted that by combining the clearly defined regional ranges of the dynamic zoning instructions, the energy storage capacity of each region can be accurately calculated by region, which can avoid the problem that traditional overall capacity calculation cannot reflect regional differences. By combining the validity verification of individual battery voltage data and eliminating abnormal energy storage module capacity, the authenticity and reliability of the first capacity value can be improved, providing accurate data support for subsequent capacity optimization and grid construction capability prediction.

[0063] Specifically, for each region, the voltage values ​​of all individual cells within the same region are extracted from the collected individual cell voltage data. The voltage difference between individual cells within the region is calculated. The difference between the highest and lowest individual cell voltages within the region is calculated to obtain the maximum voltage difference. The average voltage difference is obtained by averaging the absolute values ​​of the voltage differences between all individual cells and the average individual cell voltage of the region. A preset capacity optimization model is constructed. The input of this model is the first capacity value, the maximum voltage difference, and the average voltage difference for each region. The output is the corrected second capacity value. The correlation between the input and output includes the capacity decay coefficient corresponding to different voltage difference intervals built into the model. The larger the voltage difference, the larger the capacity decay coefficient, indicating that the impact of cell imbalance on the available capacity is more significant. In the specific correction process, based on the calculated maximum and average voltage differences, the corresponding capacity decay coefficient built into the model is matched, and the first capacity value is multiplied by (1 - capacity decay coefficient) to obtain the corrected second capacity value.

[0064] For example, the capacity optimization model defines the correspondence between voltage inconsistency index and capacity decay coefficient through a capacity decay coefficient mapping function. The capacity decay coefficient is a value between 0 and 1, representing the proportion of capacity that needs to be reduced due to inconsistency. The mapping relationship is determined through battery characteristic experimental data and engineering experience. The maximum voltage difference and average voltage difference in the dataset are used as features, and the capacity decay coefficient is used as the target value. A multiple linear regression algorithm is used for fitting to obtain a continuous function. During training, the loss function adopts mean squared error, that is, minimizing the squared difference between the predicted capacity decay coefficient and the measured capacity decay coefficient. The optimizer can use the least squares method.

[0065] The first capacity value is corrected based on the voltage difference of individual cells, which fully considers the impact of cell imbalance on the actual usable capacity. This solves the problem of overestimation of capacity caused by neglecting cell differences in traditional capacity calculation. The calculated second capacity value is more consistent with the actual operating state of the system. The capacity optimization model improves the accuracy of capacity correction by correcting through dual voltage difference dimensions, avoiding correction deviations caused by a single voltage difference index. This helps to avoid the risk of battery damage caused by cell imbalance in advance and improves the safety of system operation.

[0066] Specifically, the second capacity value of each region is obtained, and real-time voltage data of the supercapacitors in each region is collected by system sensors. The real-time voltage data of the supercapacitors is preprocessed using a moving average filtering method to obtain stable supercapacitor voltage values. A pre-defined network capability analysis model is constructed. This model adopts a multi-input single-output mapping structure. The inputs include the second capacity value of each region, the preprocessed supercapacitor voltage value, and auxiliary parameters such as the charge and discharge response time of the corresponding energy storage battery pack and the supercapacitor power density. The output is the first network capability state, including multi-dimensional evaluation results of voltage support capability level, frequency regulation capability level, continuous power supply duration, and instantaneous power regulation peak value. The correlation between input and output is established through the mapping rules after training. The model training process uses historical operating data as samples, including historical second capacity values, supercapacitor voltage values, and corresponding actual network capability performance data. The model parameters are optimized through gradient descent algorithm so that the model can accurately map the correspondence between input parameters and network capability state.

[0067] In the specific prediction process, the input parameters of each region are input into the network construction capability analysis model. The model evaluates the sub-state of network construction capability of each region, and combines the functional positioning of each region for weighted fusion to obtain the first state of network construction capability of the system as a whole. For example, the continuous power supply duration of the power supply area has a higher weight than that of the backup area, and the instantaneous power regulation peak value of the backup area has a higher weight than that of the power supply area. The final network construction capability evaluation result is obtained through weighted calculation.

[0068] It should be noted that by integrating the second capacity value of the energy storage battery pack and the voltage state of the supercapacitor as dual core parameters, the system's continuous power supply capability and instantaneous response capability are taken into account. This avoids the one-sidedness of traditional grid capacity prediction, which only focuses on a single energy storage component, and improves the comprehensiveness of grid capacity status prediction. The grid capacity analysis model achieves accurate quantitative assessment of grid capacity status through multiple input parameters and trained mapping rules. The output multi-dimensional assessment results provide accurate decision-making basis for subsequent system scheduling. By combining the functional positioning of each region for weighted fusion, the obtained prediction results can accurately match the actual operating needs of the system, avoiding scheduling deviations caused by equalizing the grid capacity assessment of each region, and improving the reliability and efficiency of the system's grid operation.

[0069] Furthermore, the real-time power status of each region is analyzed according to the dynamic zoning instructions. By controlling the charging or discharging of the energy storage batteries in each region, a smooth transition between regions is achieved, resulting in a set of functional zones, including: S701. Analyze the real-time power status of each region according to the dynamic partitioning instruction, calculate the difference between the target power value and the current power value of each region, and obtain the replenished energy. S702. Determine the energy transfer direction based on the replenished energy, and control the charging or discharging of the energy storage battery in each region according to the energy transfer direction to perform a smooth transition of the region, thereby obtaining a set of functional areas.

[0070] In this embodiment, according to the dynamic partitioning instructions, the function, operating priority, safe power range, and other attribute parameters of each region are parsed from the instructions to clarify the operating objectives of each region. Through the battery management system and system energy monitoring unit, real-time power status data of each region is collected, including but not limited to the total state of charge of the energy storage battery pack within the region, real-time charge and discharge power, average voltage of individual cells, and total capacity of the energy storage module. A regional power status analysis model is constructed. The input of this model is the real-time power status data of each region and the regional attribute parameters in the dynamic partitioning instructions. The output is the real-time power status of each region, including power stability determination. The correlation between the input and output is determined by preset power status determination rules. The power status determination rules include: when the real-time total charge status of the power supply area is lower than the preset safety lower limit and the charging and discharging power fluctuation exceeds the threshold, the power is determined to be unstable; according to the functional positioning of the area, the corresponding target power value is set. The target power value of the power supply area needs to be set in combination with the energy demand capacity to ensure a stable range for continuous power supply. The target power value of the standby area needs to be set to a redundant range to cope with sudden demand. The target power value can be set according to the system power supply accuracy requirements; the difference between the target power value and the current power value of each area is calculated to obtain the supplementary energy. If the difference is positive, it means that the area needs to supplement energy. If the difference is negative, it means that the area has excess energy that can be output. The absolute value of the difference is the amount of energy to be transferred.

[0071] It should be noted that combining the analysis of real-time power status with the regional attributes of dynamic partitioning instructions can ensure the matching of power analysis with regional functions, avoiding the problem of traditional power analysis ignoring regional differences; setting target power values ​​based on regional functions makes the calculation process of replenishing energy more accurate, providing accurate direction and quantitative standards for subsequent energy transfer; and judging the validity and stability of real-time power through power status analysis can remove the interference of abnormal power data on replenishing energy calculation, improve the accuracy of replenishing energy calculation, and provide accurate data support for the reliability of regional smooth transition processing.

[0072] Specifically, based on the calculated replenishment energy for each region, the energy transfer direction is determined. The analysis process, based on the replenishment energy, regional location distribution, and energy storage module connection topology of each region, outputs the energy transfer direction planning results, including the energy output region, receiving region, and transfer path. The relationship between input and output is established through preset energy transfer rules, which prioritize matching the nearest and shortest connection path of excess energy regions with insufficient energy regions to avoid losses caused by long-distance energy transfer across regions. Based on the energy transfer direction planning results, a smooth transition control algorithm is constructed. The input of this algorithm is the energy transfer direction, replenishment energy scale, and charge / discharge characteristic parameters of the energy storage batteries in each region. The output is the charge / discharge control parameters of the energy storage batteries in each region, including charge / discharge current, voltage, and duration. The relationship between input and output is maintained by controlling the charge / discharge rate within the allowable smooth range of the energy storage batteries to avoid abrupt rate changes.

[0073] In the specific control process, for areas requiring energy replenishment, their energy storage batteries are controlled to enter charging mode, with the charging current gradually increasing from a preset initial value to the target value, and the rate of increase not exceeding the maximum allowable current change rate of the battery. For areas with excess energy, their energy storage batteries are controlled to enter discharging mode, with the discharging current also gradually adjusted to the target value. Simultaneously, real-time data on the change in charge during the charging and discharging process in each area is collected, and the charging and discharging control parameters are dynamically corrected through feedback to ensure that the rate of change in charge in each area remains stable during the energy transfer process. After the energy transfer is completed, the operating status of each area is verified a second time to confirm that the charge in each area reaches the target range, the charging and discharging power is stable, and the consistency of individual batteries meets the requirements. The areas that pass the verification are classified according to their functional attributes, and information such as the functional type, operating parameters, and energy interaction relationships of each area are recorded to obtain a set of functional areas.

[0074] It should be noted that determining the energy transfer direction based on supplementary energy and combining it with regional topology planning of the transfer path can reduce energy transfer losses and improve energy utilization efficiency. Using a smooth transition control algorithm to control the charging and discharging rate can effectively avoid sudden changes in power and voltage fluctuations during energy transfer, thereby improving the stability of regional operation and reducing the impact on the power grid. After ensuring that the operating status of each region is qualified through secondary verification, a set of functional areas is formed, which improves the operational reliability of the functional areas and provides an accurate functional unit basis for subsequent dynamic scheduling and control command generation, thereby improving the coordination of the entire photovoltaic power distribution and storage system.

[0075] Furthermore, by combining the first network capacity status and the set of functional zones, different combinations of functional zones are dynamically scheduled for energy allocation and storage, generating corresponding control commands to obtain the first set of control commands, including: S801. Based on the first network capability status and functional area set, decompose the requirements and calculate the corresponding priorities to construct a requirement task sequence. S802. According to the required task sequence, analyze the physical characteristics and real-time status of each functional area, dynamically schedule different combinations of functional areas for energy allocation and storage, generate corresponding control commands, and obtain the first set of control commands.

[0076] In this embodiment, network parameters such as the system's voltage support capability level, frequency regulation capability level, continuous power supply duration, and instantaneous power regulation peak value are extracted from the first network capability status to determine the upper limit of the system's demand capacity. Based on the functional area set, attributes such as the functional type, available energy storage capacity, charge and discharge response speed, and operating constraints of each functional area are analyzed. A demand decomposition model is constructed. The input of this model is the overall energy allocation and storage demand of the system, the parameters of the first network capability status, and the attributes of the functional area set. The output is the decomposed sub-demand set. The relationship between the input and output is based on the functional positioning of each functional area to decompose the overall demand into corresponding sub-demands, including but not limited to decomposing the overall energy supply demand into stable energy supply sub-demands and matching them with stable energy supply areas, and decomposing it into emergency energy replenishment sub-demands and matching them with emergency response areas. The total scale of the decomposed sub-demands does not exceed the upper limit limited by the first network capability status.

[0077] Priority is calculated based on the attributes of sub-demands. Attributes include, but are not limited to, urgency, importance of the service object, and correlation with grid stability. The calculation process takes as input the attribute parameters of each sub-demand and the operation priority parameters in the first grid capacity state. The output is the priority weight of each sub-demand. The relationship between input and output is quantified by constructing a sub-demand attribute judgment matrix to quantify the relative importance of each sub-demand. For example, the priority weight of sub-demands directly related to grid frequency regulation is higher than that of ordinary energy supply sub-demands. Sub-demands are sorted from high to low according to their priority weights and adjusted in combination with the time constraints of the sub-demands to construct a demand task sequence, and to determine the execution order, task objectives, and constraints of each sub-demand.

[0078] It should be noted that combining the first network capacity status and functional area attributes for demand decomposition ensures the matching of sub-demands with system capabilities and functional area characteristics, avoiding the problem of tasks being unable to be executed due to neglecting system operating boundaries in traditional demand decomposition. By quantifying the priority of sub-demands and constructing an orderly sequence of demand tasks, the execution of critical demands can be prioritized during functional area scheduling, improving the targeting and reliability of energy allocation and storage. Strict adherence to the upper limit constraint of network capacity during demand decomposition ensures that the total scale of sub-demands is within the system's carrying capacity, avoiding a decrease in system network stability due to demand overload and ensuring the safe operation of the system.

[0079] Specifically, based on the sequence of required tasks, the physical characteristics and real-time status of each functional area are analyzed. Different combinations of functional areas are dynamically scheduled for energy allocation and storage, generating corresponding control commands to obtain the first set of control commands. Dynamic matching based on the physical characteristics and real-time status of functional areas ensures that the combination of functional areas can accurately adapt to each sub-requirement, avoiding resource waste or insufficient capacity problems caused by traditional fixed functional area allocation. Through timing coordination verification and reallocation mechanisms, timing conflicts in functional area scheduling can be resolved, improving the coordination of operation among functional areas and ensuring the smooth execution of the required task sequence. The generated control commands include detailed operating parameters and execution conditions, enabling accurate control of functional areas, improving the precision of energy allocation and storage, and enhancing the orderliness and efficiency of system operation.

[0080] Furthermore, according to the demand task sequence, the physical characteristics and real-time status of each functional area are analyzed, different combinations of functional areas are dynamically scheduled for energy allocation and storage, and corresponding control commands are generated to obtain the first set of control commands, including: S901. Analyze the physical characteristics of each functional area according to the sequence of required tasks, match the corresponding required tasks for each functional area, and obtain the first task allocation result. S902. Analyze the real-time status of each functional area, filter out the demand tasks with incompatible timing between functional areas in the first task allocation result, and reallocate them to obtain the second task allocation result. S903. Based on the second task allocation result, generate the corresponding control instructions to obtain the first control instruction set.

[0081] In this embodiment, according to the demand task sequence, information such as task objectives, constraints, and task types for each sub-demand in the sequence are extracted sequentially. Based on the functional area set, the physical characteristic parameters of each functional area are analyzed, including but not limited to energy storage module type, maximum charge / discharge power, energy conversion efficiency, response delay time, rated capacity, charge / discharge cycle life, and operating environment tolerance range. The physical characteristics of the functional areas are analyzed, and the capability assessment results of each functional area are output based on the physical characteristic parameters, including the type of task it can handle, the upper limit of task scale, and the response performance level. The correlation between input and output includes analyzing when the maximum charge / discharge power of a functional area is higher than a preset threshold and the response delay is lower than a threshold. The system determines whether a function can handle instantaneous frequency regulation tasks; it performs task matching, outputting the first task allocation result based on the task information of each sub-requirement, the capability assessment results of the functional area, and the execution order of the required task sequence. It establishes the input-output relationship by ensuring that the capability assessment results of the functional area meet the task objectives and constraints of the sub-requirements. This includes matching continuous power supply sub-requirements to stable power supply areas with high energy conversion efficiency and sufficient rated capacity, and matching instantaneous frequency regulation sub-requirements to emergency response areas with low response delay and wide power adjustment range. The matching results are then compiled into the first task allocation result, determining the sub-requirement list, preliminary task execution time planning, and task execution parameter requirements for each functional area.

[0082] It should be noted that by analyzing the physical characteristics of functional areas and the information of sub-requirements, the adaptability of task allocation to functional area capabilities is ensured, avoiding the problem of low task execution efficiency or inability to execute tasks caused by ignoring the differences in functional area characteristics in traditional task allocation; based on preset matching rules, the automatic matching of functional areas and required tasks is realized, improving the efficiency and standardization of task allocation; the first task allocation result clarifies the task list and core requirements of each functional area, providing a clear basis for subsequent timing collaborative verification and control instruction generation, ensuring the orderly progress of subsequent processes.

[0083] Specifically, based on the first task allocation result, task allocation information for each functional area is extracted, including but not limited to task execution time period, task load intensity, and task priority. Simultaneously, through the system status monitoring unit and battery management system, real-time status data for each functional area is collected, including real-time available capacity, current charge / discharge power, remaining operating load margin, individual battery consistency status, and the progress of currently executed tasks. Real-time status analysis of the functional areas is performed. Based on the real-time status data of the functional areas and the task information in the first task allocation result, the real-time carrying capacity assessment results for each functional area are obtained, including the remaining load-bearing capacity scale and the time window for adding new tasks. The correlation between input and output is established through preset status carrying capacity mapping rules. These rules include determining that a functional area's remaining carrying capacity is weak when its current charge / discharge power has reached 80% of its rated power and its remaining available capacity is below a preset threshold.

[0084] Specifically, a time-series coordination verification is performed. Based on the first task allocation result, the real-time carrying capacity assessment result of each functional area, and the time constraints of the required task sequence, the time-series coordination assessment result is obtained, including a list of time-series incoordinating tasks and conflict types, such as multiple high-load tasks allocated to the same functional area at the same time, or task execution time exceeding the functional area's carrying time window. The correlation between input and output is obtained by analyzing the conflict situation within the same time period. For the selected time-series incoordinating tasks, a task reassignment algorithm is used for adjustment. The adjustment strategies include: transferring conflicting tasks to other functional areas with sufficient remaining carrying capacity and suitable physical characteristics, splitting the execution time of conflicting tasks into the idle time window of the functional area, and discarding low-priority conflicting tasks according to task priority. The adjusted task allocation result is organized into the second task allocation result to ensure that all tasks are conflict-free in the time dimension and that the real-time status of the functional area can support the smooth execution of tasks.

[0085] It should be noted that by combining real-time status analysis of functional areas to assess temporal coordination, it is possible to accurately identify temporal conflicts in the first task allocation result, avoiding the task execution chaos caused by traditional scheduling ignoring real-time status. By adjusting the matching relationship through task reassignment algorithms, temporal coordination problems can be eliminated, ensuring that each functional area executes tasks in an orderly manner in the time dimension, thereby improving the coordination and reliability of functional area scheduling. By adhering to the principle of prioritizing core demand tasks during the reassignment process, the execution of high-priority tasks can be guaranteed, avoiding the failure of critical tasks to be completed on time due to temporal conflicts, and improving the targeting and effectiveness of energy allocation and storage.

[0086] Specifically, based on the second task allocation result, the task objectives, task execution conditions, and real-time status data of each functional area are extracted; control commands are generated, and initial control commands for each functional area are generated based on the task information, real-time status data, and physical characteristic parameters of the functional areas in the second task allocation result. The relationship between input and output is determined based on the task objectives and functional area status to determine the core parameters of the control commands. For continuous power supply tasks, the control commands must specify the charging and discharging mode, target power, maintenance duration, and total state of charge protection threshold. The target power setting is determined by combining the real-time available power of the functional area and the power required by the task to ensure that it does not exceed the physical characteristic limits of the functional area; the initial control commands are validated for validity, including whether the command parameters are within the allowable range of the physical characteristics of the functional area, whether the command logic is consistent with the task execution conditions, and whether there are any conflicts in the command parameters. Invalid parameters are optimized and adjusted; the validated control commands are encoded according to the preset command format, including the functional area identifier, command type, core control parameters, execution priority, start conditions, and fault feedback mechanism; the control commands corresponding to all functional areas are integrated according to the execution order of the required task sequence and the functional area identifier to obtain the first set of control commands.

[0087] It should be noted that the generation of control instructions based on the second task allocation results and functional area status ensures a precise match between the instructions and task requirements and functional area capabilities, avoiding execution deviations caused by neglecting the compatibility between tasks and functional areas in traditional control instruction generation. Through validity verification and format encoding, the reliability and standardization of control instructions are improved, ensuring that the system execution unit can accurately identify and execute the instructions. The first set of control instructions is organized according to the execution order and functional area classification, which facilitates the orderly execution of each instruction by the system, improves the coordination of functional area scheduling and the efficiency of system operation, and provides accurate instruction basis for the precise operation control of the subsequent photovoltaic power distribution and energy storage system.

[0088] like Figure 4 As shown, a grid-type photovoltaic (PV) power grid and energy storage system operation optimization system is used to implement an operation optimization method for a grid-type PV power grid and energy storage system, including: The operation parameter analysis module, in response to the power-on command, decouples the hardware of the energy distribution and storage process based on the pre-built energy distribution and storage database, analyzes the energy storage and power supply capacity, and obtains the first set of operation parameters; The dynamic partitioning module, based on the real-time collected photovoltaic power distribution and storage system status data and the first set of operating parameters, analyzes energy demand and energy storage status through data fusion, dynamically partitions the photovoltaic power distribution and storage system, predicts the grid construction capacity status, and obtains dynamic partitioning instructions and the first grid construction capacity status. The regional optimization module analyzes the real-time power status of each region according to the dynamic partitioning instructions, and performs smooth transition processing of the region by controlling the charging or discharging of the energy storage battery in each region, thus obtaining a set of functional areas. The control command generation module, in conjunction with the first network capacity status and the set of functional zones, dynamically schedules different combinations of functional zones for energy allocation and storage, generates corresponding control commands, and obtains the first set of control commands. The control command optimization module controls the operation of the photovoltaic power generation and storage system according to the first set of control commands. During operation, it collects real-time operation status data and optimizes the first set of control commands through a preset command optimization model to optimize the operation of the photovoltaic power generation and storage system.

[0089] In this embodiment, the operation parameter analysis module actively decouples and maps the energy storage hardware logic during system power-on initialization. Based on real-time data and a pre-built database, it accurately assesses the system's actual energy reserves and power supply capacity, generating a first set of operation parameters reflecting the current true capacity boundary. This provides an accurate and reliable initial benchmark for the entire optimization process from the source, avoiding decision-making biases caused by using fixed nominal parameters and improving the accuracy of all subsequent optimization stages and the safety of system startup. The dynamic partitioning module integrates system status data and operation parameters in real time, analyzes energy demand, and dynamically divides the physically connected hardware cluster into regions with different functional focuses based on the real-time capacity, health status, and consistency of the energy storage units. This predicts the overall network support capability of the system, enabling on-demand networking and specialized division of energy storage resources. It provides accurate and reliable network capability predictions that match the actual internal state, enhancing the system's ability to cope with uncertainties.

[0090] Specifically, the regional optimization module receives dynamic zoning instructions and actively and smoothly adjusts the charging and discharging states of energy storage batteries in each region through a closed-loop control strategy. This ensures that the energy level of each region smoothly transitions to the target operating point that matches its function. It prepares a stable and optimized internal energy layout for subsequent task execution, eliminates switching shocks that may result from uneven energy distribution between regions, reduces battery stress, extends equipment lifespan, and provides a stable internal foundation for precise scheduling. The control command generation module combines network capability prediction with the state attributes of functional areas to decompose energy allocation and storage requirements and dynamically match them to the most suitable functional area combination, generating a coordinated and conflict-free sequence of underlying control commands. This achieves optimal dynamic matching of tasks and resources, avoids command conflicts and resource consumption, and improves the overall efficiency and coordination of the system in executing complex network tasks. During actual system operation, the control command optimization module, based on real-time feedback state data, uses a preset optimization model to continuously optimize and dynamically correct the set of control commands being executed. This enables the system to have online self-learning and adaptive adjustment capabilities, effectively overcoming model mismatch and external disturbances, and continuously improving the system's control accuracy, stability, and robustness under dynamic operating conditions.

[0091] Example 2 This embodiment illustrates the overall process of this application in conjunction with a specific application scenario. In this example, the system is a 1MW / 2MWh grid-type photovoltaic-storage power station, which includes three parallel lithium iron phosphate battery clusters (cluster 1, cluster 2, and cluster 3, each with a rated capacity of approximately 667kWh), a supercapacitor bank (capacity 1kWh, rated voltage 400V), and corresponding photovoltaic arrays and power conversion equipment.

[0092] Upon receiving the power-on command, the main controller immediately sends a status query command to the BMS and supercapacitor manager of each battery cluster via the CAN bus. Each device replies with real-time data: Battery cluster 1 has a terminal voltage of 752.1V, an average temperature of 28℃, and the BMS calculates a maximum allowable discharge current of 280A based on the internal cell status; Battery cluster 2 has a terminal voltage of 751.8V, an average temperature of 30℃, and a maximum allowable discharge current of 260A; Battery cluster 3 has a terminal voltage of 753.0V, an average temperature of 25℃, and a maximum allowable discharge current of 300A. The supercapacitor bank's terminal voltage is 385V. The controller accesses a pre-built energy storage database, which stores a mapping table of battery voltage, temperature, and corresponding SOC. Through table lookup and interpolation calculations, the SOCs of clusters 1, 2, and 3 are obtained as 65%, 62%, and 70%, respectively. Based on the rated capacity and current SOC of each cluster, the available energy of each cluster is calculated to be 433.4 kWh, 413.5 kWh, and 466.9 kWh, respectively, with a total system available energy of approximately 1313.8 kWh. Regarding power supply capacity, taking the minimum allowable discharge current of 260 A reported by the BMS for the three clusters, multiplied by the system average voltage (approximately 752.3 V), the maximum continuous discharge power of the battery is approximately 195.6 kW. Based on its current voltage of 385 V and the corresponding equivalent series resistance of 0.15 Ω in the database, the maximum instantaneous discharge power of the supercapacitor is calculated to be approximately (385^2) / (4 × 0.15) ≈ 247.0 kW. Combining the above analysis, the first set of operating parameters is obtained: total system available energy 1313.8 kWh, maximum continuous battery power 195.6 kW, and supercapacitor peak power 247.0 kW.

[0093] The controller collects real-time data showing a photovoltaic output of 300kW and a local load of 200kW. Simultaneously, it receives a primary frequency regulation reserve requirement instruction from the grid dispatch center, requesting a reserve of 100kW of regulation capacity. Based on this analysis, the current net power demand is 200kW load minus 300kW photovoltaic, i.e., -100kW. This means the system has 100kW of surplus power available for charging, but simultaneously needs to reserve 100kW of frequency regulation capacity. An energy demand vector is constructed, with the high-priority frequency regulation demand capacity being 100kW of power reserve. Combining the first set of operating parameters, the system's maximum sustainable power supply is 195.6kW (battery) + 247.0kW (supercapacitor, but only used for instantaneous support and not included in continuous power). Therefore, the 195.6kW from the battery is primarily used to meet the continuous power demand. The dynamic partitioning module aims to maximize the fulfillment of the frequency regulation demand, considering the states of each battery cluster (SOC of 65%, 62%, and 70%), and calculates this through an optimization algorithm. A feasible division is as follows: Clusters 1 and 3, with moderate SOC and good response capabilities, are designated as the power supply zone, with a combined usable energy of approximately 900.3 kWh and a maximum continuous discharge power of approximately (280 A from Cluster 1 + 300 A from Cluster 3) × average voltage ≈ 436.4 kW, fully meeting the 100 kW frequency regulation backup requirement with ample margin; Cluster 2, with lower SOC and slightly higher temperature, is designated as the backup zone. Meanwhile, supercapacitors are separately designated as the fast response zone. The grid capacity status is predicted, and based on the zoning results, the cell consistency index of the power supply zones (Clusters 1 and 3) is calculated. The standard deviation of the individual cell voltage in cluster 1 is 0.025V, and that in cluster 3 is 0.018V. The theoretical capacity of the power supply area is corrected by a preset capacity optimization model (for example, a 1% capacity reduction is introduced when the standard deviation of the voltage is greater than 0.02V). The corrected capacity and supercapacitor status are then input into the grid construction capability analysis model (this model is a neural network trained with historical data, with battery capacity, power, consistency index, and supercapacitor voltage as inputs and grid construction index as output) to obtain the first grid construction capability status. For example, it is predicted that the system can provide an inertia support time constant of 4.5 seconds and a primary frequency regulation capacity of 150kW.

[0094] According to the dynamic zoning instructions, the target SOC for the power supply zone is set to 68%, and the target SOC for the standby zone is set to 60%. The weighted average SOC of the current power supply zone (cluster 1 at 65%, cluster 3 at 70%) is calculated to be approximately 67.5%, close to the target value of 68%, requiring only a small amount of charging. The current SOC of the standby zone (cluster 2) is 62%, requiring discharge to 60%. The zone optimization module generates smooth power instructions, controlling the power supply zone to charge slowly at a lower power (e.g., 5kW), while simultaneously controlling the standby zone to discharge at approximately 10kW. During this process, a slightly high voltage was detected in a single cell within cluster 3, triggering the balancing function for fine-tuning. After several minutes of transition, the charge levels in each zone reach the target, forming a stable set of functional zones: power supply zone (cluster 1, cluster 3, SOC ~68%), standby zone (cluster 2, SOC ~60%), and fast response zone (supercapacitor, voltage 385V).

[0095] At this moment, the grid frequency fluctuates, requiring the system to immediately provide +50kW power support for an estimated duration of 10 seconds. The control command generation module dynamically schedules power based on the first grid capability status (confirming system capability) and the status of the functional area set. The decision logic prioritizes this short-term, rapid power support task for the fast response area (supercapacitor), as it has the fastest response speed. The module generates the first set of control commands: immediately instruct the power converter of the supercapacitor bank to increase its output power to 50kW within 1 second and maintain it for 10 seconds. Simultaneously, considering the potential need for energy recovery after the support ends, it pre-generates commands to subsequently put the power supply area into low-power charging.

[0096] The system executed the aforementioned control commands. During execution, the command optimization module collected operational data in real time at a frequency of 100Hz. It found that the actual output power of the supercapacitor was 48kW, slightly lower than the command, and the grid frequency recovery speed was slightly slower than expected. Based on this real-time feedback, the preset command optimization model (based on model predictive control principles) immediately performed rolling optimization calculations. The model predicted that including the batteries in the power supply area would accelerate frequency recovery. Therefore, approximately 2 seconds after the original command was executed, an optimized command was generated: while maintaining the supercapacitor's output of 48kW, cluster 1 in the power supply area was additionally activated, discharging at 10kW for 2 seconds. This optimized command was superimposed and executed in real time. Ultimately, the system actually output approximately 58kW of support power, and the grid frequency recovered to normal more quickly. Throughout the entire response process, the system continuously monitored the status of each functional area to ensure its operation within safe boundaries.

[0097] This embodiment fully demonstrates the entire process of the method of this application, from system power-on initialization, dynamic resource organization, state preparation, task scheduling to closed-loop optimization.

[0098] The above description is merely a preferred embodiment of this application. The scope of protection of this application is not limited to the above embodiments. All technical solutions falling within the scope of this application's concept are within the scope of protection of this application. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of this application should also be considered within the scope of protection of this application.

Claims

1. A method for optimizing the operation of a grid-type photovoltaic power distribution and energy storage system, characterized in that, include: In response to the power-on command, the hardware of the energy allocation and storage process is decoupled based on the pre-built energy allocation and storage database, and the energy storage and power supply capacity are analyzed to obtain the first set of operating parameters; Based on the real-time collected photovoltaic power distribution and storage system status data and the first set of operating parameters, the energy demand and energy storage status are analyzed through data fusion, the photovoltaic power distribution and storage system is dynamically partitioned, and the grid construction capacity status is predicted to obtain the dynamic partitioning command and the first grid construction capacity status. Analyze the real-time power status of each region according to the dynamic partitioning instructions, and perform smooth transition processing of each region by controlling the charging or discharging of the energy storage battery in each region to obtain a set of functional areas. By combining the first grid capacity status and the set of functional zones, different combinations of functional zones are dynamically scheduled for energy allocation and storage, generating corresponding control commands to obtain the first set of control commands. The photovoltaic power generation and storage system is controlled according to the first set of control instructions. During operation, real-time operation status data is collected, and the first set of control instructions is optimized through a preset instruction optimization model to optimize the operation of the photovoltaic power generation and storage system.

2. The method for optimizing the operation of a grid-type photovoltaic power distribution and storage system according to claim 1, characterized in that, In response to the power-on command, the hardware of the energy allocation and storage process is decoupled based on a pre-built energy allocation and storage database. Energy reserves and power supply capacity are analyzed to obtain a first set of operating parameters, including: In response to the power-on command, the hardware status of the allocation process is analyzed based on the pre-built allocation database, the hardware is decoupled, and the initial hardware state is obtained. By combining the initial state of the hardware with the real-time state data of the corresponding energy storage hardware, the corresponding energy reserves and power supply capacity are analyzed to obtain the first grid capacity value. The first network capacity value is mapped to the corresponding operating parameters in the distribution and storage database to obtain the first set of operating parameters.

3. The method for optimizing the operation of a grid-type photovoltaic power distribution and storage system according to claim 2, characterized in that, The first network capacity value is mapped to the corresponding operating parameters in the distribution and storage database to obtain the first set of operating parameters, including: Based on the first network capacity value and the rated parameters of the supercapacitor in the corresponding hardware, analyze the charging voltage variation of the supercapacitor and set the first voltage value. The first voltage value is matched in the energy storage database to obtain the maximum rate of change of photovoltaic power under the corresponding voltage. By combining the maximum rate of change of photovoltaic power and the first grid capacity value, the corresponding operating parameters are calculated to obtain the first set of operating parameters.

4. The method for optimizing the operation of a grid-type photovoltaic power distribution and storage system according to claim 1, characterized in that, The process involves dynamically partitioning the photovoltaic power distribution and storage system based on real-time collected status data and a first set of operating parameters, analyzing energy demand and storage conditions through data fusion, predicting grid connection capability status, and obtaining dynamic partitioning instructions and a first grid connection capability status, including: Based on the analysis of the first set of operating parameters, an energy demand vector is constructed. By combining real-time collected photovoltaic power distribution and storage system status data and energy demand vector, the energy storage situation is analyzed, capacity constraints are constructed, and the photovoltaic power distribution and storage system is dynamically partitioned to obtain dynamic partitioning instructions; Based on the dynamic zoning instructions, the consistency of energy storage conditions in each region is analyzed, the grid construction capacity status is predicted, and the first grid construction capacity status is obtained.

5. The method for optimizing the operation of a grid-type photovoltaic power distribution and storage system according to claim 4, characterized in that, The process involves combining real-time acquired photovoltaic power generation and energy storage system status data and energy demand vectors to analyze energy storage conditions, construct capacity constraints, dynamically partition the photovoltaic power generation and energy storage system, and obtain dynamic partitioning instructions, including: Based on the energy demand vector, the corresponding energy demand is analyzed and the corresponding priority is calculated to obtain the corresponding energy demand capacity. By combining the energy demand capacity and the real-time collected photovoltaic power distribution and storage system status data, the total state of charge, health status and supercapacitor voltage status of the energy storage battery pack are analyzed to construct the corresponding safety boundary and energy supply limit, and the maximum energy supply value is obtained. By comparing energy demand capacity and maximum energy supply value, the energy supply area and reserve area that meet the energy demand capacity are divided from the maximum energy supply value; By combining the power supply area and the backup area, the consistency status of individual cells within the battery pack in different areas is analyzed, and the area boundaries are adjusted and optimized to obtain dynamic partitioning instructions.

6. The method for optimizing the operation of a grid-type photovoltaic power distribution and storage system according to claim 5, characterized in that, The step of analyzing the consistency of energy storage conditions in each region based on dynamic zoning instructions, predicting the grid capacity status, and obtaining the first grid capacity status includes: Based on the dynamic zoning instructions and the real-time collected total state of charge of the energy storage battery pack and the voltage data of each individual battery cell, the first capacity value of each region is calculated. For each region, the voltage difference between individual cells is calculated, and the first capacity value is corrected and optimized using a preset capacity optimization model to obtain the second capacity value. By combining the second capacity value and the voltage state of the supercapacitor, the network construction capability state is predicted through a preset network construction capability analysis model, and the first network construction capability state is obtained.

7. The method for optimizing the operation of a grid-type photovoltaic power distribution and energy storage system according to claim 1, characterized in that, The process involves analyzing the real-time power status of each region according to dynamic partitioning instructions, and performing smooth transition processing by controlling the charging or discharging of the energy storage batteries in each region, resulting in a set of functional zones, including: The real-time power status of each region is analyzed according to the dynamic partitioning instructions, and the difference between the target power value and the current power value of each region is calculated to obtain the replenished energy. The direction of energy transfer is determined by the replenished energy, and the charging or discharging of the energy storage battery in each region is controlled according to the direction of energy transfer to achieve a smooth transition between regions, thus obtaining a set of functional areas.

8. The method for optimizing the operation of a grid-type photovoltaic power distribution and storage system according to claim 1, characterized in that, The first control command set is obtained by combining the first grid capacity status and the functional area set, dynamically scheduling different functional area combinations for energy allocation and storage, generating corresponding control commands, and including: Based on the first network capability status and functional area set, decompose the requirements and calculate the corresponding priorities to construct a requirement task sequence; According to the sequence of required tasks, analyze the physical characteristics and real-time status of each functional area, dynamically schedule different combinations of functional areas for energy allocation and storage, generate corresponding control commands, and obtain the first set of control commands.

9. The method for optimizing the operation of a grid-type photovoltaic power distribution and storage system according to claim 8, characterized in that, The process involves analyzing the physical characteristics and real-time status of each functional area according to the required task sequence, dynamically scheduling different combinations of functional areas for energy allocation and storage, generating corresponding control commands, and obtaining a first set of control commands, including: Based on the sequence of required tasks, analyze the physical characteristics of each functional area, match the corresponding required tasks for each functional area, and obtain the first task allocation result; Analyze the real-time status of each functional area, filter out the demand tasks with inconsistent timing between functional areas in the first task allocation result, and reallocate them to obtain the second task allocation result; Based on the second task allocation result, the corresponding control instructions are generated to obtain the first set of control instructions.

10. A grid-type photovoltaic power distribution and energy storage system operation optimization system, characterized in that, A method for optimizing the operation of a grid-type photovoltaic power distribution and storage system as described in any one of claims 1 to 9, comprising: The operation parameter analysis module, in response to the power-on command, decouples the hardware of the energy distribution and storage process based on the pre-built energy distribution and storage database, analyzes the energy storage and power supply capacity, and obtains the first set of operation parameters; The dynamic partitioning module, based on the real-time collected photovoltaic power distribution and storage system status data and the first set of operating parameters, analyzes energy demand and energy storage status through data fusion, dynamically partitions the photovoltaic power distribution and storage system, predicts the grid construction capacity status, and obtains dynamic partitioning instructions and the first grid construction capacity status. The regional optimization module analyzes the real-time power status of each region according to the dynamic partitioning instructions, and performs smooth transition processing of the region by controlling the charging or discharging of the energy storage battery in each region, thus obtaining a set of functional areas. The control command generation module, in conjunction with the first network capacity status and the set of functional zones, dynamically schedules different combinations of functional zones for energy allocation and storage, generates corresponding control commands, and obtains the first set of control commands. The control command optimization module controls the operation of the photovoltaic power generation and storage system according to the first set of control commands. During operation, it collects real-time operation status data and optimizes the first set of control commands through a preset command optimization model to optimize the operation of the photovoltaic power generation and storage system.

Images