Multi-stage cooperative control method and system for optical storage charging supply station and optical storage charging supply station

Abstract

The embodiment of the invention relates to a multi-stage cooperative control method and system for an optical storage and charging supply station and the optical storage and charging supply station, and belongs to the technical field of electric energy storage and power networks. Inputting the meteorological characteristics and the historical power generation data into a first time sequence prediction model based on a long short-term memory network to generate a photovoltaic power generation prediction sequence, inputting the load data and the related characteristics into a second time sequence prediction model to generate a load prediction sequence, and calculating a net power difference based on the two prediction sequences, the method comprises the following steps: establishing and solving an optimization model with minimum operation cost as a target, generating a scheduling strategy, executing the optimization strategy in a grid-connected mode, monitoring power balance in an off-grid mode, and executing progressive load reduction in unbalance to maintain system stability. According to the embodiment of the invention, collaborative optimization and safe operation of the optical storage charging supply station in multiple modes are realized.

Description

Technical Field

[0001] This application relates to a multi-level coordinated control method, system, and photovoltaic energy storage charging and replenishment station, belonging to the technical fields of energy storage and power grids. Background Technology

[0002] As the global energy structure accelerates its transition to clean and low-carbon energy, new energy power generation technologies, represented by photovoltaics, have developed rapidly. Simultaneously, photovoltaic-storage-charging-replenishment stations, which integrate photovoltaic power generation, energy storage systems, charging facilities, and local loads, have become crucial nodes in new power systems and promote the local consumption of new energy. By integrating distributed energy resources with controllable loads, photovoltaic-storage-charging-replenishment stations can theoretically achieve comprehensive functions such as self-generation and self-consumption of energy, surplus energy storage, and emergency backup, which is of positive significance for improving the resilience of the distribution network and reducing carbon emissions.

[0003] However, existing photovoltaic (PV) energy storage charging and replenishment stations still face a series of unresolved technical problems in terms of coordinated control, which restricts the large-scale application and development of PV energy storage charging and replenishment stations. The main shortcomings of existing technologies are as follows: Firstly, in terms of resource scheduling, most photovoltaic-storage charging and replenishment stations can only perform independent optimization for a single site or adopt simple rule-based control strategies, failing to fully utilize the spatiotemporal complementarity between distributed sites. For example, when a site experiences insufficient photovoltaic power generation due to weather conditions, the surplus photovoltaic power from adjacent sites often cannot be supplemented through an effective inter-site mutual assistance mechanism, resulting in low overall resource utilization.

[0004] Second, in terms of optimizing control strategies, existing control methods do not adequately consider the overall balance between the economy and safety of system operation. When operating in grid-connected mode, many control strategies use fixed threshold judgments for grid interaction, failing to fully consider market and environmental factors such as time-of-use pricing and grid dispatch instructions. Furthermore, they often ignore the actual operational constraints of photovoltaic-storage charging and replenishment stations, such as charging and discharging efficiency decay and cycle life limitations. This causes the dispatch plan to deviate from expectations in actual execution, making it impossible to automatically achieve the established goals.

[0005] Third, when the distribution network fails and the photovoltaic-storage charging and replenishment station needs to switch to off-grid operation, there is a lack of rapid mode switching mechanism and autonomous stability control capability. The system has failed to establish a real-time frequency / voltage feedback control architecture and has not carried out refined hierarchical management of the load. This can easily lead to frequency instability and voltage collapse of the off-grid system under major disturbances, or excessive load shedding to avoid instability, which seriously affects the reliability of power supply.

[0006] In conclusion, existing technologies can no longer meet people's needs and urgently need to be improved. Summary of the Invention

[0007] The main objective of this application is to provide a multi-level collaborative control method, system, and optical energy storage charging and replenishment station for optical energy storage charging and replenishment, so as to solve the shortcomings of the prior art.

[0008] The embodiments of this application are implemented using the following technical solutions: According to one aspect of the embodiments of this application, a multi-level coordinated control method for photovoltaic-storage charging and replenishment stations is provided, comprising: acquiring environmental data of different photovoltaic-storage charging and replenishment stations in real time; constructing a power generation time series sample based on the environmental data and a preset photovoltaic output characteristic model, wherein the power generation time series sample includes at least meteorological characteristic photovoltaic power generation data, load curve data, and a multi-dimensional correlation feature vector; inputting the meteorological feature vector and historical photovoltaic power generation data in the multi-dimensional correlation feature vector into a first time series prediction model to calculate a photovoltaic power generation output prediction sequence; inputting the time feature vector and meteorological feature vector in the multi-dimensional correlation feature vector into a second time series prediction model to calculate a power load prediction sequence; calculating the net power difference based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, constructing an optimization model and solving it to generate a power optimization dispatch strategy, wherein the power optimization dispatch strategy includes a power optimization dispatch strategy for grid interaction power; executing the power optimization dispatch strategy in grid-connected mode, or: in off-grid mode, monitoring the power supply and demand balance in real time, and if a power supply and demand imbalance is detected in off-grid mode, performing a progressive load reduction operation.

[0009] According to at least one specific implementation of the embodiments of this application, the real-time acquisition of environmental data of different photovoltaic storage charging and replenishment stations specifically refers to the real-time acquisition of environmental data of different photovoltaic storage charging and replenishment stations based on a distributed multi-station collaborative architecture. In the distributed multi-station collaborative architecture, the environmental data includes light intensity, temperature data, humidity data, and geographical data. The multi-dimensional correlation feature vector includes: meteorological feature vector, time feature vector, and system state feature vector.

[0010] According to at least one specific embodiment of the present application, the first time-series prediction model is a photovoltaic output physical-data fusion model based on a long short-term memory network architecture, which is trained with meteorological feature vectors as input and historical photovoltaic power generation data as supervision targets, and the second time-series prediction model is a gated recurrent unit network based on attention mechanism enhancement.

[0011] According to at least one specific embodiment of the present application, the step of inputting the meteorological feature vector and historical photovoltaic power generation data from the multidimensional correlation feature vector into a first time series prediction model to calculate a photovoltaic power generation output prediction sequence, and inputting the time feature vector and meteorological feature vector from the multidimensional correlation feature vector into a second time series prediction model to calculate a power load prediction sequence, further includes: acquiring historical photovoltaic power generation data, load curve data, meteorological feature vector, and time feature vector of the photovoltaic-storage-charging-replenishment station, constructing a multivariate time series input feature set, inputting the meteorological feature vector and historical photovoltaic power generation data into the first time series prediction model, learning the nonlinear mapping relationship between meteorological conditions and power generation through the long short-term memory network architecture built into the first time series prediction model, and generating a photovoltaic power generation output prediction sequence for future prediction periods; inputting the load curve data, time feature vector, and meteorological feature vector into the second time series prediction model, the second time series prediction model adopting a cascaded structure of gated recurrent unit network and attention mechanism to calculate a power load prediction sequence, and calculating a net power difference sequence for each time period based on the photovoltaic power generation output prediction sequence and the power load prediction sequence.

[0012] According to at least one specific implementation of the embodiments of this application, the power optimization scheduling strategy is specifically as follows: based on the current grid-connected / off-grid status of the photovoltaic-storage charging and replenishment station, execute the corresponding grid-connected optimization scheduling strategy or the off-grid frequency stabilization control strategy.

[0013] According to at least one specific embodiment of the present application, the step of calculating the net power difference based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, constructing an optimization model and solving it to generate a power optimization dispatch strategy, the power optimization dispatch strategy including a power optimization dispatch strategy for grid interaction power, further includes: collecting real-time system operating status data, the real-time system operating status data including photovoltaic array output power, energy storage system state of charge, charging pile operating power and total station load power; calculating the net power deviation value of the photovoltaic-energy storage charging and replenishment station at the current moment based on the power balance principle; using the net power deviation value, the real-time state of charge of the energy storage system and the current time-of-use electricity price period, constructing a real-time power allocation model with minimizing operating cost as the objective function; performing secondary rule solving on the real-time power allocation model to generate a collaborative dispatch scheme; generating a power optimization dispatch strategy according to the collaborative dispatch scheme; and distributing the power optimization dispatch strategy to the local controller of each photovoltaic-energy storage charging and replenishment station through a communication network.

[0014] According to at least one specific embodiment of the present application, operating constraints are set in the real-time power allocation model. The operating constraints include at least: photovoltaic power generation rate of change constraints, energy storage system charging and discharging power and capacity constraints, charging pile power adjustable range constraints, and interruptible load capacity constraints.

[0015] According to at least one specific embodiment of the present application, it further includes: in off-grid mode, disconnecting from the grid, controlling the system voltage and frequency through the energy storage converter, and configuring low-frequency load shedding protection; when the system frequency is lower than a first threshold, progressively reducing the load according to a preset load priority sequence until the frequency recovers to above a second threshold.

[0016] According to another aspect of the embodiments of this application, a multi-level collaborative control system for photovoltaic-storage charging and replenishment stations is provided to implement the aforementioned multi-level collaborative control method for photovoltaic-storage charging and replenishment stations, comprising: a power generation time series sample construction module, which acquires environmental data of different photovoltaic-storage charging and replenishment stations in real time, and constructs power generation time series samples based on the environmental data and a preset photovoltaic output characteristic model, wherein the power generation time series samples include at least meteorological characteristic photovoltaic power generation data, load curve data, and a multi-dimensional correlation feature vector; and a time series prediction model calculation module, which inputs the meteorological feature vector and historical photovoltaic power generation data from the multi-dimensional correlation feature vector into a first time series prediction model to calculate the photovoltaic power generation output prediction sequence. The system inputs the time feature vector and meteorological feature vector from the multidimensional correlation feature vector into the second time-series prediction model to calculate the power load prediction sequence. The power optimization dispatch strategy generation module calculates the net power difference and constructs an optimization model based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, and solves the model to generate a power optimization dispatch strategy, which includes a power optimization dispatch strategy for grid interaction power. The grid-connected and off-grid dispatch strategy execution module executes the power optimization dispatch strategy in grid-connected mode, or: in off-grid mode, it monitors the power supply and demand balance in real time, and if an off-grid power supply and demand imbalance is detected, it performs a progressive reduction operation to interrupt the load.

[0017] According to another aspect of the embodiments of this application, a photovoltaic energy storage charging and replenishment station is provided, wherein the photovoltaic energy storage charging and replenishment station is provided with the aforementioned multi-level collaborative control system.

[0018] The beneficial technical effects of the embodiments of this application are: This application's embodiments construct a control method and system from accurate prediction and optimized decision-making to adaptive execution of modes. First, by integrating environmental and historical data and using a time-series prediction model, future photovoltaic power generation and system load are independently predicted. Then, based on the prediction results, an optimization model considering multiple constraints is dynamically constructed and solved to generate a comprehensive power dispatch strategy covering energy storage dispatch, inter-station mutual assistance, and grid interaction. Finally, according to the grid-connected or off-grid operation mode, an economical dispatch strategy or stability control program is adaptively executed to achieve optimal control of photovoltaic, energy storage, charging, and load under different operating modes. Attached Figure Description

[0019] To more clearly illustrate the specific implementation methods of the embodiments of this application or the technical solutions in the prior art, the drawings used in the description of the specific implementation methods or the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of a multi-level coordinated control method for photovoltaic energy storage charging and replenishment stations.

[0021] Figure 2 This is a flowchart of the optimization technical solutions provided in steps S21 to S23.

[0022] Figure 3 This is a flowchart of the optimization technical solutions provided in steps S31 to S33.

[0023] Figure 4 This is the architecture diagram of a multi-level collaborative control system for a photovoltaic-storage-charging-replenishment station. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the embodiments of this application, and not all embodiments. Based on the specific implementation methods in the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of the embodiments of this application.

[0025] like Figure 1 The multi-level coordinated control method for the photovoltaic storage charging and replenishment station shown includes: Step S1: Real-time acquisition of environmental data from different photovoltaic-storage charging and replenishment stations. Based on the environmental data and a preset photovoltaic output characteristic model, a power generation time series sample is constructed. The power generation time series sample includes at least meteorological characteristic photovoltaic power generation data, load curve data, and multi-dimensional correlation feature vectors. For example, in the distributed multi-station collaborative architecture, the environmental data includes light intensity, temperature data, humidity data, and geographical data; the multi-dimensional correlation feature vectors include: meteorological feature vectors, time feature vectors, and system state feature vectors.

[0026] Step S2: Input the meteorological feature vector and historical photovoltaic power generation data from the multidimensional correlation feature vector into the first time-series prediction model to calculate the photovoltaic power generation output prediction sequence. Input the time feature vector and meteorological feature vector from the multidimensional correlation feature vector into the second time-series prediction model to calculate the power load prediction sequence. In this step, the first time-series prediction model is a photovoltaic output physical-data fusion model based on a long short-term memory network architecture, trained with meteorological feature vectors as input and historical photovoltaic power generation data as the supervision target. The second time-series prediction model is a gated recurrent unit network based on attention mechanism enhancement.

[0027] Step S3: Based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, calculate the net power difference, construct an optimization model and solve it to generate a power optimization dispatch strategy. The power optimization dispatch strategy includes a power optimization dispatch strategy for grid interaction power.

[0028] Step S4: In grid-connected mode, execute the power optimization scheduling strategy; or: in off-grid mode, monitor the power supply and demand balance in real time. If an off-grid power supply and demand imbalance is detected, perform a progressive reduction operation to interrupt the load. For example, if the remaining power of each photovoltaic-storage charging and replenishment station is higher than a preset threshold, power is fed back to the grid; if the remaining power of each photovoltaic-storage charging and replenishment station is lower than the preset threshold, power is received from the grid.

[0029] The technical solutions provided in steps S1 to S4 first perform time-series forecasting of photovoltaic power generation and electricity load, convert the forecast results into net power difference, and construct an optimization model based on the net power difference. A comprehensive dispatch strategy is then generated through the optimization model. Finally, based on the current grid-connected / off-grid operation status of photovoltaic power generation and energy storage, the power optimization dispatch strategy is executed. Alternatively, in off-grid mode, the power supply and demand balance is monitored in real time. If a power supply and demand imbalance is detected in off-grid mode, a progressive reduction operation of interrupted load is executed. Step S1 aggregates heterogeneous data from multiple sources, including environmental data, historical operational data, and system status data, through a distributed multi-site collaborative architecture, constructing a data foundation to describe the current operational status of the photovoltaic-storage-charging-load replenishment station. Step S2 then processes power generation and load forecasting based on this data foundation. The collaborative efforts of Steps S1 and S2 ensure that the predicted sequence input to the optimization model in Step S3 has high accuracy and interpretability, thereby enhancing the reliability of subsequent optimization decisions.

[0030] Step S3 is responsible for decision output. It transforms the high-precision, forward-looking source-load prediction sequence output from Step S2 into a net power difference characterizing the future power surplus or deficit of the photovoltaic-storage charging and replenishment station power generation system. Using this net power difference as input, a corresponding optimization model is constructed and solved. This optimization model simultaneously considers the charging and discharging plans of each photovoltaic-storage charging and replenishment station, inter-station mutual assistance power, and grid interaction power. Step S3 applies this optimization model to charge power when electricity prices are low and there is a predicted power surplus, dispatching electricity from predicted surplus stations to predicted shortage stations, scientifically planning inter-station mutual assistance paths, and formulating the optimal grid interaction strategy.

[0031] Step S4 is responsible for executing decisions. In grid-connected mode, step S4 executes the economically optimized scheduling strategy generated in step S3, realizing interaction with various recharge stations under the main grid and reducing the cost of grid-connected power generation. When photovoltaic and energy storage charging needs to switch to off-grid mode due to a fault, the control logic responds quickly based on real-time monitoring of power supply and demand, executes progressive load control, and realizes an adaptive mode execution mechanism. This provides a safety baseline for photovoltaic power generation, ensuring that the photovoltaic and energy storage charging and recharge stations can maintain stable operation under any extreme conditions, preventing system failure or collapse. It achieves a balance between the economic efficiency of photovoltaic and energy storage operation and the reliability of power supply, improving the stability and reliability of the entire photovoltaic and energy storage charging and recharge station cluster.

[0032] like Figure 2 As shown, preferably, in step S2, the step of inputting the meteorological feature vector and historical photovoltaic power generation data from the multidimensional correlation feature vector into the first time series prediction model to calculate the photovoltaic power generation output prediction sequence, and inputting the time feature vector and meteorological feature vector from the multidimensional correlation feature vector into the second time series prediction model to calculate the power load prediction sequence, further includes: Step S21: Obtain historical photovoltaic power generation data, load curve data, meteorological feature vector, and time feature vector of the photovoltaic-storage charging and replenishment station, construct a multivariate time-series input feature set, input the meteorological feature vector and historical photovoltaic power generation data into the first time-series prediction model, learn the nonlinear mapping relationship between meteorological conditions and power generation through the long short-term memory network architecture built into the first time-series prediction model, and generate a photovoltaic power generation output prediction sequence for future prediction periods.

[0033] Step S22: Input the load curve data, time feature vector and meteorological feature vector into the second time series prediction model. The second time series prediction model adopts a cascaded structure of gated cyclic unit network and attention mechanism. It calculates the power load prediction sequence by dynamically focusing key historical load segments and related features through attention weight. Based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, the net power difference sequence is calculated for each time period.

[0034] Step S23: Using the net power difference sequence and the current real-time time-of-use electricity price signal, construct and solve a mixed-integer linear programming model with the objective of minimizing the total operating cost of the photovoltaic-storage charging and replenishment station, generating a corresponding optimized scheduling strategy. In this step, minimizing the total operating cost includes minimizing the total operating cost of the photovoltaic-storage charging and replenishment station while satisfying power balance, energy storage charging and discharging, and grid interaction constraints. The optimized power scheduling strategy includes energy storage charging and discharging plans, inter-station mutual assistance power, and grid-connected interaction power, as well as:

[0035] The optimized scheduling strategy executes the corresponding grid-connected optimized scheduling strategy or off-grid frequency stabilization control strategy based on the current grid-connected / off-grid status of the optical storage charging and replenishment station.

[0036] The optimization solutions provided in steps S21 to S23 first integrate historical operating data and multi-source features, and use time-series prediction models with different architectures to generate photovoltaic power generation output prediction sequences and power load prediction sequences respectively, deriving future net power demand. Then, combined with real-time electricity price signals, and with minimizing total operating cost as the core objective, a corresponding optimized scheduling strategy is constructed and generated. This optimized scheduling strategy encompasses different optimization strategies such as energy storage scheduling, inter-station mutual assistance, and grid interaction, and belongs to a comprehensive optimization strategy. Among them: Step S21 utilizes a Long Short-Term Memory (LSTM) network to model and calculate the temporal relationship between weather and power generation, capturing the long-term dependent photovoltaic (PV) power generation prediction curve. Step S22 employs a cascaded structure of a gated recurrent unit network (GRN) and an attention mechanism. The GRN network excels at handling short-term fluctuations and nonlinear changes in load sequences, while the attention mechanism identifies historical load patterns and related characteristics that have the greatest impact on the current prediction (such as the impact of specific holidays, sudden weather changes, and other special periods on PV power generation). Through the fusion of the GRN and the attention mechanism, the PV-storage-charging-load replenishment station can use the most suitable model for prediction based on the different characteristics of PV power generation and power load, resulting in high-precision and interpretable output results.

[0037] Step S23 uses the net power difference and real-time time-of-use electricity price signal output from step S22 as inputs to construct a mixed-integer linear programming model with the objective of minimizing total operating cost. The net power difference represents the power variation of the photovoltaic-storage-charging-load system itself, while the real-time time-of-use electricity price signal indicates the quantitative situation of the electricity market environment. Based on the dual collaborative logic of these two variables, the mixed-integer linear programming model outputs the following decisions: when charging / discharging energy storage costs are lowest (utilizing low-priced electricity or avoiding high-priced electricity), and the economic efficiency of inter-station mutual assistance (reducing high-priced grid purchases and achieving mutual assistance in power generation between different photovoltaic-storage-charging-load replenishment stations). Through the above optimization strategies, the integration of power resources in the three dimensions of time, space, and equipment is achieved. It fully utilizes the price advantages of peak and off-peak electricity prices and the advantages of different stations and equipment (energy storage, photovoltaics, and the grid), ensuring that photovoltaic-storage-charging-load replenishment stations can adopt the most appropriate scheduling methods under both normal economic operation and emergency supply conditions, thus improving the ability of photovoltaic-storage-charging-load replenishment stations to cope with emergencies such as grid failures.

[0038] like Figure 3 As shown, preferably, in step S3, the net power difference is calculated based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, and an optimization model is constructed and solved to generate a power optimization dispatch strategy. The power optimization dispatch strategy includes a power optimization dispatch strategy for grid interaction power, and further includes: Step S31: Collect real-time operating status data of the system. The real-time operating status data of the system includes the output power of the photovoltaic array, the state of charge of the energy storage system, the operating power of the charging pile and the total load power of the entire station. Calculate the net power deviation value of the photovoltaic-energy storage-charging-load replenishment station at the current moment based on the power balance principle.

[0039] Step S32: Using the net power deviation value, the real-time state of charge of the energy storage system, and the current time-of-use electricity price period, a real-time power allocation model is constructed with minimizing operating costs as the objective function. The real-time power allocation model includes operational constraints, which at least include: photovoltaic power generation rate of change constraints, energy storage system charging and discharging power and capacity constraints, adjustable power range constraints for charging piles, and interruptible load capacity constraints.

[0040] Step S33: Perform secondary rule solving on the real-time power allocation model to generate a collaborative scheduling scheme. Based on the collaborative scheduling scheme, generate an energy optimization scheduling strategy and distribute the energy optimization scheduling strategy to the local controllers of each photovoltaic-storage charging and replenishment station via the communication network. The collaborative scheduling scheme includes: photovoltaic power adjustment commands, energy storage charging and discharging power commands, charging pile power dynamic allocation commands, and interruptible load switching commands.

[0041] The optimization solutions provided in steps S31 to S33 calculate the net power deviation by collecting key operational data of photovoltaic and energy storage in real time. Then, the net power deviation value is combined with real-time electricity prices and energy storage status to construct an online optimization model. This online optimization model aims to minimize operating costs and comprehensively considers multiple constraints required by the current operating status and operational safety boundaries of different photovoltaic and energy storage charging and replenishment stations. Finally, a collaborative scheduling scheme is generated. This scheme includes scheduling instructions for specific equipment, enabling rapid smoothing of power fluctuations at replenishment stations and real-time optimized allocation of resources. Step S31 collects key state quantities such as photovoltaic array output power, energy storage system state of charge, charging pile operating power, and total load power of the entire station, and calculates the net power deviation value based on the power balance principle. This allows for the rapid acquisition of the power balance status of photovoltaic, energy storage, charging and load at the current moment. A positive net power deviation value indicates overgeneration, while a negative value indicates that the load demand exceeds the generation capacity. Step S31 provides input parameters reflecting the health of photovoltaic, energy storage, charging and load for the optimization model in step S32.

[0042] Step S32 optimizes based on an objective function, which aims to minimize the total operating cost of the refueling station. Step S32 introduces real-time time-of-use electricity pricing, directly linking the optimization objective to the economics of electricity prices. This guides the refueling station to minimize grid-purchased electricity during periods of high prices or increase energy storage charging during periods of low prices. Step S32 also sets multiple operational constraints, defining the safe operating range of each device at the current moment. It seeks the lowest-cost power allocation scheme that satisfies the current power balance, ensuring that the decision meets both economic optimization requirements and strictly guarantees the safety and stable operation of the refueling station's equipment, avoiding over-regulation of power control or overload of electrical equipment. For example, operational constraints could include: photovoltaic power change rate, energy storage power and capacity limits, adjustable charging pile range, load disconnectability capacity, etc. Depending on actual operational needs, operational constraints can be added or removed to meet the requirements of different scenarios.

[0043] Step S33 performs a secondary programming solution on the model constructed in step S32, converting the output of the optimized model into an executable collaborative scheduling scheme. This scheme includes direct instructions issued to the local controllers of the equipment at the replenishment station, such as photovoltaic power regulation, energy storage charging and discharging, charging pile power allocation, and load switching. After the equipment executes the instructions, the current state of the photovoltaic, energy storage, charging, and load changes. Step S33 then monitors the variations in the system's real-time operating status data again, initiating the next control cycle and achieving iterative control.

[0044] In summary, the optimization solutions provided in steps S31 to S33 aim to minimize the operating cost of the replenishment station in each adjustment, and achieve coordination and control of multiple resources such as photovoltaics, energy storage, charging piles, and loads under equipment safety constraints, thereby improving the operational economy and safety of the photovoltaic-energy storage-charging-load replenishment station.

[0045] As a preferred option, in off-grid mode, the connection with the grid is disconnected, and the system voltage and frequency are controlled by the energy storage converter, and low-frequency load shedding protection is configured; when the system frequency is lower than the first threshold, the load is progressively reduced according to the preset load priority sequence until the frequency recovers to above the second threshold.

[0046] In this optimized technical solution, the active voltage and frequency regulation capability of the energy storage converter and the preset low-frequency load shedding protection logic are used together to address power imbalances in off-grid conditions and ensure system safety. This optimized technical solution controls the energy storage converter to proactively assume the responsibility of supporting system voltage and frequency immediately after disconnection from the grid, providing rapid response and smoothing for daily load fluctuations. When a severe power deficit exceeding the energy storage regulation capacity occurs, causing the system frequency to continuously drop to a first threshold (e.g., 49.5Hz), the preset low-frequency load shedding protection is activated according to a preset load priority sequence, progressively cutting off or reducing non-critical loads to restore balance at minimal cost.

[0047] The load reduction action continues until the frequency recovers to above the second threshold (e.g., 49.8Hz), creating a hysteresis effect. This ensures the system remains stable within a safe range while preventing frequent malfunctions of protection devices near the frequency critical point. The coordinated operation of these two threshold settings constitutes a comprehensive defense system for the off-grid system, covering all scenarios from normal operation to emergency failures, significantly improving the system's stability and reliability in off-grid conditions.

[0048] For the method steps disclosed in the above embodiments, the method steps are described as a series of actions for the purpose of simplicity. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily necessary for the embodiments of this application.

[0049] like Figure 4 The multi-level collaborative control system for the photovoltaic energy storage charging and replenishment station shown is used to implement the multi-level collaborative control method for the photovoltaic energy storage charging and replenishment station described in any specific embodiment of this application, including: The power generation time series sample construction module acquires environmental data of different photovoltaic-storage charging and replenishment stations in real time. Based on the environmental data and the preset photovoltaic output characteristic model, it constructs power generation time series samples. The power generation time series samples include at least meteorological characteristic photovoltaic power generation data, load curve data and multidimensional correlation feature vectors. The time-series prediction model calculation module inputs the meteorological feature vector and historical photovoltaic power generation data from the multi-dimensional correlation feature vector into the first time-series prediction model to calculate the photovoltaic power generation output prediction sequence, and inputs the time feature vector and meteorological feature vector from the multi-dimensional correlation feature vector into the second time-series prediction model to calculate the power load prediction sequence. The power optimization dispatch strategy generation module calculates the net power difference and constructs an optimization model based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, and solves the model to generate a power optimization dispatch strategy. The power optimization dispatch strategy includes a power optimization dispatch strategy for grid interaction power. The grid-connected and off-grid scheduling strategy execution module executes the power optimization scheduling strategy in grid-connected mode, or monitors the power supply and demand balance in real time in off-grid mode. If an off-grid power supply and demand imbalance is detected, a progressive reduction operation of interrupted load is executed.

[0050] This application embodiment also provides a photovoltaic energy storage charging and replenishment station, wherein the photovoltaic energy storage charging and replenishment station is provided.

[0051] The implementation methods of the system described above are merely illustrative. For example, the various functional modules, units, or subsystems within the system may or may not be physically separate, or they may or may not be physical units; that is, they may be located in the same place or distributed across multiple different systems and their subsystems or modules. Those skilled in the art can select some or all of the functional modules, units, or subsystems to achieve the objectives of the embodiments of this application according to actual needs. Those skilled in the art can understand and implement the above-described situations without any creative effort.

[0052] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the specification of the embodiments of this application.

[0053] In the description of the embodiments of this application, the reference to terms such as "an embodiment," "example," "specific example," etc., means that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the embodiments of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0054] Furthermore, the technical solutions of the various implementation methods in this application can be combined with each other, but only if they are based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the embodiments of this application.

[0055] All features disclosed in the embodiments of this application, or all steps in the disclosed methods or processes, may be combined in any way, except for mutually exclusive features and / or steps. Any feature disclosed in the specification of the embodiments of this application, unless specifically stated otherwise, may be replaced by other equivalent or similar alternative features. That is, unless specifically stated otherwise, each feature is merely one example of a series of equivalent or similar features. Throughout the specification, the same reference numerals indicate the same elements.

[0056] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification of embodiments (including the corresponding claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification of embodiments (including the corresponding claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of this application, and are not intended to limit them. Although the embodiments of this application have been described in detail with reference to the foregoing specific embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein, and such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the specific embodiments of this application.

Claims

1. A multi-level coordinated control method for a photovoltaic energy storage charging and replenishment station, characterized in that, include: Real-time acquisition of environmental data from different photovoltaic-storage charging and replenishment stations; construction of power generation time series samples based on the environmental data and a preset photovoltaic output characteristic model; the power generation time series samples include at least meteorological characteristic photovoltaic power generation data, load curve data, and multidimensional correlation feature vectors. The meteorological feature vector and historical photovoltaic power generation data in the multidimensional correlation feature vector are input into the first time series prediction model to calculate the photovoltaic power generation output prediction sequence. The time feature vector and meteorological feature vector in the multidimensional correlation feature vector are input into the second time series prediction model to calculate the power load prediction sequence. Based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, the net power difference is calculated, an optimization model is constructed and solved, and a power optimization dispatch strategy is generated. The power optimization dispatch strategy includes a power optimization dispatch strategy for grid interaction power. In grid-connected mode, the power optimization scheduling strategy is executed; or in off-grid mode, the power supply and demand balance is monitored in real time. If an off-grid power supply and demand imbalance is detected, a progressive reduction operation of interrupted loads is executed.

2. The multi-level coordinated control method for photovoltaic storage charging and replenishment stations according to claim 1, characterized in that, The real-time acquisition of environmental data from different photovoltaic and energy storage charging and replenishment stations specifically involves acquiring environmental data from different photovoltaic and energy storage charging and replenishment stations in real time based on a distributed multi-station collaborative architecture. In the distributed multi-station collaborative architecture, the environmental data includes light intensity, temperature data, humidity data, and geographical data. The multi-dimensional correlation feature vector includes meteorological feature vector, time feature vector, and system state feature vector.

3. The multi-level coordinated control method for photovoltaic storage charging and replenishment stations according to claim 1, characterized in that, The first time-series prediction model is a photovoltaic output physical-data fusion model based on a long short-term memory network architecture. It is trained with meteorological feature vectors as input and historical photovoltaic power generation data as the supervision target. The second time-series prediction model is a gated recurrent unit network based on attention mechanism enhancement.

4. The multi-level coordinated control method for photovoltaic energy storage charging and replenishment stations according to claim 2, characterized in that, The step of inputting the meteorological feature vector and historical photovoltaic power generation data from the multidimensional correlation feature vector into the first time series prediction model to calculate the photovoltaic power generation output prediction sequence, and inputting the time feature vector and meteorological feature vector from the multidimensional correlation feature vector into the second time series prediction model to calculate the power load prediction sequence, further includes: Historical photovoltaic power generation data, load curve data, meteorological feature vectors, and time feature vectors of the photovoltaic-storage charging and replenishment station are acquired. A multivariate time-series input feature set is constructed. The meteorological feature vectors and historical photovoltaic power generation data are input into the first time-series prediction model. The nonlinear mapping relationship between meteorological conditions and power generation is learned through the long short-term memory network architecture built into the first time-series prediction model, and a photovoltaic power generation output prediction sequence for future prediction periods is generated. The load curve data, time feature vector, and meteorological feature vector are input into the second time series prediction model. The second time series prediction model adopts a cascaded structure of gated cyclic unit network and attention mechanism to calculate the power load prediction sequence. Based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, the net power difference sequence is calculated for each time period.

5. The multi-level coordinated control method for photovoltaic storage charging and replenishment stations according to claim 1, characterized in that, The power optimization scheduling strategy is specifically as follows: based on the current grid-connected / off-grid status of the photovoltaic-storage charging and replenishment station, execute the corresponding grid-connected optimization scheduling strategy or off-grid frequency stabilization control strategy.

6. The multi-level coordinated control method for photovoltaic energy storage charging and replenishment stations according to claim 1, characterized in that, The process involves calculating the net power difference based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, constructing an optimization model, solving it, and generating a power optimization dispatch strategy. This power optimization dispatch strategy includes a power optimization dispatch strategy for grid interaction power, and further includes: The system collects real-time operating status data, which includes photovoltaic array output power, energy storage system state of charge, charging pile operating power, and total load power of the entire station. Based on the power balance principle, the net power deviation value of the photovoltaic-storage-charging-replenishment station at the current moment is calculated. Using the net power deviation value, the real-time state of charge of the energy storage system, and the current time-of-use pricing period, a real-time power allocation model is constructed with the objective function of minimizing operating costs. The real-time power allocation model is solved by a second rule to generate a collaborative scheduling scheme. Based on the collaborative scheduling scheme, an energy optimization scheduling strategy is generated and the energy optimization scheduling strategy is distributed to the local controllers of each photovoltaic-storage-charging-replenishment station through a communication network.

7. The multi-level coordinated control method for photovoltaic storage charging and replenishment stations according to claim 6, characterized in that, The real-time power allocation model is configured with operational constraints, which include at least the following: photovoltaic power generation rate of change constraints, energy storage system charging and discharging power and capacity constraints, charging pile power adjustable range constraints, and interruptible load capacity constraints.

8. The multi-level coordinated control method for photovoltaic storage charging and replenishment stations according to claim 1, characterized in that, Also includes: In off-grid mode, the connection with the grid is disconnected, and the system voltage and frequency are controlled through the energy storage converter, and low-frequency load shedding protection is configured. When the system frequency is lower than the first threshold, the load is gradually reduced according to the preset load priority sequence until the frequency recovers to above the second threshold.

9. A multi-level collaborative control system for a photovoltaic-storage-charging-replenishment station, used to implement the multi-level collaborative control method for a photovoltaic-storage-charging-replenishment station as described in any one of claims 1 to 7, characterized in that, include: The power generation time series sample construction module acquires environmental data of different photovoltaic-storage charging and replenishment stations in real time. Based on the environmental data and the preset photovoltaic output characteristic model, it constructs power generation time series samples. The power generation time series samples include at least meteorological characteristic photovoltaic power generation data, load curve data and multidimensional correlation feature vectors. The time-series prediction model calculation module inputs the meteorological feature vector and historical photovoltaic power generation data from the multi-dimensional correlation feature vector into the first time-series prediction model to calculate the photovoltaic power generation output prediction sequence, and inputs the time feature vector and meteorological feature vector from the multi-dimensional correlation feature vector into the second time-series prediction model to calculate the power load prediction sequence. The power optimization dispatch strategy generation module calculates the net power difference and constructs an optimization model based on the photovoltaic power generation output prediction sequence and the power load prediction sequence, and solves the model to generate a power optimization dispatch strategy. The power optimization dispatch strategy includes a power optimization dispatch strategy for grid interaction power. The grid-connected and off-grid scheduling strategy execution module executes the power optimization scheduling strategy in grid-connected mode, or monitors the power supply and demand balance in real time in off-grid mode. If an off-grid power supply and demand imbalance is detected, a progressive reduction operation of interrupted load is executed.

10. A photovoltaic energy storage and recharge station, characterized in that, The photovoltaic-storage charging and replenishment station is equipped with the multi-level collaborative control system for photovoltaic-storage charging and replenishment stations as described in claim 9.

Images