Control method and system of photovoltaic energy storage system

Abstract

The invention relates to a control method and system for a photovoltaic energy storage system, and the method comprises the steps: obtaining the real-time irradiance data of the photovoltaic energy storage system, carrying out the power generation prediction of an irradiance region, and obtaining a dynamic power generation prediction value; acquiring charge state parameters of a photovoltaic energy storage system and aggregated power load demand data of a target power utilization area, and performing energy storage scheduling analysis in combination with the dynamic power generation power prediction value to obtain a reference energy storage scheduling strategy; acquiring energy consumption equipment parameters of the target power utilization area, and performing equipment scheduling control with the reference energy storage scheduling strategy to obtain an equipment-level control instruction set; and performing grid-connected mode and off-grid mode collaborative analysis on the dynamic generation power prediction value, the reference energy storage scheduling strategy and the equipment-level control instruction set to obtain an energy storage global control strategy. According to the invention, multi-level collaborative optimization can be realized, and the economical efficiency and sustainability of the system are improved while stable power supply is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic energy storage systems, and particularly relates to a control method and system for a photovoltaic energy storage system. Background Art

[0002] As an important technical means for renewable energy utilization, with the continuous improvement of the penetration rate of distributed energy and the in-depth promotion of the power market reform, how to optimize the scheduling of the energy storage resources of a photovoltaic energy storage system and intelligently match the electricity load. At the energy storage scheduling level, most solutions only perform charge and discharge control based on fixed thresholds, lacking dynamic optimization of the coupling relationship between the state of charge and real-time load demand, which is likely to lead to insufficient energy storage utilization or load matching deviation. Traditional control strategies often treat the grid-connected / off-grid mode switching and device-level scheduling separately, making it difficult to meet the multi-objective coordination requirements in complex electricity usage scenarios, restricting the improvement of the economy and reliability of the system. Summary of the Invention

[0003] The main object of the present invention is to provide a control method and system for a photovoltaic energy storage system, which can achieve multi-level collaborative optimization, improve the economy and sustainability of the system while ensuring stable power supply.

[0004] To achieve the above object, the present invention provides a control method for a photovoltaic energy storage system, including: Obtaining real-time irradiance data of the photovoltaic energy storage system, performing irradiance regional power generation prediction, and obtaining a dynamic power generation prediction value; Obtaining the state of charge parameter of the photovoltaic energy storage system and the aggregated power load demand data of the target power consumption area, and combining the dynamic power generation prediction value to perform energy storage scheduling analysis to obtain a reference energy storage scheduling strategy; Obtaining the energy-consuming device parameters of the target power consumption area, and performing device scheduling control with the reference energy storage scheduling strategy to obtain a device-level control instruction set; Performing grid-connected mode and off-grid mode collaborative analysis on the dynamic power generation prediction value, the reference energy storage scheduling strategy and the device-level control instruction set to obtain an energy storage global control strategy.

[0005] Further, the obtaining real-time irradiance data of the photovoltaic energy storage system, performing irradiance regional power generation prediction, and obtaining a dynamic power generation prediction value includes: Obtaining the real-time irradiance data of the photovoltaic power generation array of the photovoltaic energy storage system, and performing spatial grid processing on the real-time irradiance data to obtain an irradiance distribution matrix; Performing cloud movement trajectory analysis on the irradiance distribution matrix according to a preset historical meteorological database to obtain a dynamic irradiance correction coefficient; Perform unit power generation efficiency conversion on the irradiance distribution matrix according to the dynamic irradiance correction coefficient to obtain a unit power generation power matrix; Obtain the target meteorological conditions of the historical meteorological database, and perform unit power prediction on the unit power generation power matrix to obtain a unit power generation power prediction result; Perform dynamic spatial integration on the unit power generation power prediction result according to the irradiance distribution matrix to obtain the dynamic power generation power prediction value.

[0006] Furthermore, obtain the state of charge parameter of the photovoltaic energy storage system and the aggregated power load demand data of the target power consumption area, and perform energy storage scheduling analysis in combination with the dynamic power generation power prediction value to obtain a reference energy storage scheduling strategy, including: Perform charge and discharge capacity evaluation processing on the state of charge parameter to obtain a charge and discharge power constraint range; Perform real-time difference calculation on the dynamic power generation power prediction value and the aggregated power load demand data to obtain a real-time net power difference sequence; Perform frequency domain decomposition on the real-time net power difference sequence to obtain a basic power component and an adjustable power component; Construct charge and discharge instructions according to the basic power component and the charge and discharge power constraint range to obtain an initial charge and discharge instruction set; Perform collaborative optimization processing on the initial charge and discharge instruction set and the adjustable power component to obtain an optimized charge and discharge scheduling sequence; Perform boundary verification processing on the optimized charge and discharge scheduling sequence according to a preset scheduling safety threshold to obtain the reference energy storage scheduling strategy.

[0007] Furthermore, the constructing charge and discharge instructions according to the basic power component and the charge and discharge power constraint range to obtain an initial charge and discharge instruction set includes: Perform time domain segmentation processing on the basic power component to obtain basic power sub-components in multiple continuous time intervals; Perform charge and discharge feasibility determination on each basic power sub-component according to the charge and discharge power constraint range to obtain charge and discharge permission flag bits for each time interval; Perform linear programming solution on the charge and discharge permission flag bits according to the basic power sub-component and the charge and discharge power constraint range to obtain a set of candidate charge and discharge power values; Construct instructions for the set of candidate charge and discharge power values according to a preset state of charge safety interval to obtain the initial charge and discharge instruction set.

[0008] Further, obtaining the energy-consuming equipment parameters of the target power consumption area, and performing equipment scheduling control with the reference energy storage scheduling strategy to obtain a device-level control instruction set, including: Decompose the power response characteristics of the energy-consuming equipment parameters to obtain a device power feature set; Perform dynamic association matching on the reference energy storage scheduling strategy according to the device power feature set to obtain a device energy storage association matrix; Perform multi-device scheduling analysis based on the device energy storage association matrix to obtain a device scheduling solution set; Construct device-level instructions for the device scheduling solution set according to the reference energy storage scheduling strategy to obtain the device-level control instruction set.

[0009] Further, performing collaborative analysis of grid-connected mode and off-grid mode on the dynamic power generation prediction value, the reference energy storage scheduling strategy, and the device-level control instruction set to obtain an energy storage global control strategy, including: Calculate the grid-connected and off-grid power distribution of the dynamic power generation prediction value to obtain a grid-connected power reference value and an off-grid power reference value; Check the energy storage charge and discharge constraints for the grid-connected power reference value according to the state of charge parameter to obtain a grid-connected power correction value; Evaluate the reliability of off-grid power supply for the device-level control instruction set to obtain a load guarantee report; <X Perform power matching degree analysis according to the off-grid power reference value and the load guarantee report to obtain an off-grid operation feasibility determination result; Calculate the power deviation of the grid-connected power correction value according to a preset grid dispatching instruction to obtain a grid-connected regulation instruction; Make an operation mode decision on the reference energy storage scheduling strategy according to the off-grid operation feasibility determination result and the grid-connected regulation instruction to obtain the energy storage global control strategy.

[0010] Further, making an operation mode decision on the reference energy storage scheduling strategy according to the off-grid operation feasibility determination result and the grid-connected regulation instruction to obtain the energy storage global control strategy, including: Perform binary processing on the off-grid operation feasibility determination result according to a preset off-grid operation feasibility threshold to obtain a mode switching trigger flag; Perform dynamic priority division on the grid-connected regulation instruction according to the mode switching trigger flag to obtain a grid-connected regulation weight coefficient; Perform power distribution optimization calculation on the grid-connected regulation weight coefficient and the off-grid power reference value to obtain a mode collaborative power instruction; Dynamically modify and update the reference energy storage scheduling strategy according to the pattern collaborative power instruction to obtain an optimized energy storage charge-discharge strategy; Conduct an off-grid power supply margin test on the optimized energy storage charge-discharge strategy according to the load guarantee report to obtain off-grid support capacity information; Conduct a global optimization of the operating mode of the optimized energy storage charge-discharge strategy according to the off-grid support capacity information and the pattern collaborative power instruction to obtain the global energy storage control strategy.

[0011] Further, based on the device-level control instruction set, analyze the power supply requirements of the energy-consuming device parameters to obtain an emergency power supply demand sequence, and combine the global energy storage control strategy to perform a load priority reallocation on the photovoltaic energy storage system to obtain an emergency control plan: Parse the device-level control instruction set and map the device status to obtain a device operation priority matrix; Conduct a power supply continuity analysis on the energy-consuming device parameters according to the device operation priority matrix to obtain a target load identification set; Calculate the supply-demand gap for the target load identification set according to the dynamic power generation prediction value to obtain an emergency power supply demand sequence; Conduct a mode adaptability correction process on the emergency power supply demand sequence according to the global energy storage control strategy to obtain a corrected demand sequence; Conduct a load-energy storage emergency allocation on the photovoltaic energy storage system according to the corrected demand sequence and the global energy storage control strategy to obtain the emergency control plan.

[0012] The present invention also provides a control system for a photovoltaic energy storage system, which is applied to the control method of the photovoltaic energy storage system described in any one of the above, and includes: An acquisition module, which is used to acquire real-time irradiance data of the photovoltaic energy storage system, conduct irradiance regional power generation prediction, and obtain a dynamic power generation prediction value; An analysis module, which is used to acquire the state of charge parameter of the photovoltaic energy storage system and the aggregated power load demand data of the target power consumption area, and combine the dynamic power generation prediction value to conduct energy storage scheduling analysis to obtain a reference energy storage scheduling strategy; An association module, which is used to acquire the energy-consuming device parameters of the target power consumption area, and conduct device scheduling control with the reference energy storage scheduling strategy to obtain a device-level control instruction set; A processing module, which is used to conduct grid-connected mode and off-grid mode collaborative analysis on the dynamic power generation prediction value, the reference energy storage scheduling strategy, and the device-level control instruction set to obtain a global energy storage control strategy.

[0013] A control method and system for a photovoltaic energy storage system provided by the present invention have the following beneficial effects: Through dynamic irradiance area power generation prediction, it can more accurately reflect the real-time power generation capacity of the photovoltaic system, reduce prediction errors, provide a reliable basis for energy storage scheduling, and thus improve the overall energy utilization efficiency of the system. Combining the state of charge and load demand data, dynamically adjusting the energy storage charge and discharge strategy, avoiding overcharging or over-discharging problems, improving the service life of the energy storage system, enhancing the power grid supply-demand matching ability at the same time, and reducing the light curtailment rate. Generating device-level control instructions based on the parameters of energy-consuming devices, realizing precise load management, optimizing the operation efficiency of devices, reducing unnecessary energy waste, and enhancing the economic efficiency of the system. Through the collaborative analysis of grid-connected and off-grid modes, ensuring that the system can operate stably under different working conditions, improving power supply reliability, reducing dependence on the power grid at the same time, and enhancing the autonomous regulation ability of the system. Integrating power generation prediction, energy storage scheduling and device control, realizing multi-level collaborative optimization, while ensuring stable power supply, improving the economic efficiency and sustainability of the system. Brief Description of the Drawings

[0014] Figure 1 is a flowchart of a control method for a photovoltaic energy storage system provided by the present invention; Figure 2 is a structural diagram of a control system for a photovoltaic energy storage system provided by the present invention.

[0015] The realization, functional characteristics and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. Detailed Embodiments

[0016] In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, and are not used to limit the present invention.

[0017] Next, with reference to the accompanying drawings and specific implementation manners, the present invention will be further described.

[0018] Referring to Figure 1 as described above, the present invention also provides a control method for a photovoltaic energy storage system, including: Step S1: Obtain the real-time irradiance data of the photovoltaic energy storage system, perform irradiance area power generation prediction, and obtain the dynamic power generation power prediction value; Step S2: Obtain the state of charge parameter of the photovoltaic energy storage system and the aggregated power load demand data of the target power consumption area, and perform energy storage scheduling analysis in combination with the dynamic power generation power prediction value to obtain the benchmark energy storage scheduling strategy; Step S3: Obtain the parameters of the energy-consuming devices in the target power consumption area, perform device scheduling control with the benchmark energy storage scheduling strategy, and obtain the device-level control instruction set; Step S4: Conduct collaborative analysis of the dynamic power generation prediction value, the reference energy storage scheduling strategy, and the device-level control instruction set for grid-connected and off-grid modes to obtain the global energy storage control strategy.

[0019] Based on the above steps, the detailed step process is as follows: Step S1: Real-time irradiance data is the core input for power generation prediction of the photovoltaic energy storage system, and its accuracy directly affects the reliability of the dynamic power generation prediction value. During the implementation process, high-precision irradiance sensors need to be deployed or meteorological satellite data interfaces need to be accessed to collect the solar radiation intensity of the target area at a second-level or minute-level frequency. In the data preprocessing link, outliers need to be removed and the moving average or Kalman filter algorithm is used to smooth the noise. Irradiance-based regional power generation prediction is achieved through physical models or machine learning methods. Physical models calculate the theoretical power generation based on parameters such as the conversion efficiency, tilt angle, and azimuth angle of photovoltaic modules; machine learning models rely on the mapping relationship between historical irradiance and power generation to train algorithms such as LSTM or random forest. Dynamic power generation prediction needs to consider spatio-temporal characteristics such as cloud movement and seasonal changes. Short-term prediction (15 minutes to 1 hour) uses time series analysis, and medium- and long-term prediction (more than 1 day) combines numerical weather forecasting. The prediction results are output in the form of probability distribution or interval estimation to provide a flexible adjustment space for subsequent scheduling.

[0020] Step S2: The state of charge parameter reflects the current energy reserve of the energy storage system. It is necessary to monitor the single-cell voltage, temperature, and charge-discharge cycle times in real time through the battery management system to correct the SOC estimation value. The aggregated power load demand data comes from smart meters or load prediction systems, covering base load, peak-valley load, and random fluctuation components. The matching analysis between the dynamic power generation prediction value and the load demand uses a multi-objective optimization method, with the minimum grid power purchase cost, the lowest energy storage life loss, or the highest renewable energy consumption rate as the objective function. The constraint conditions include the SOC safety threshold, the charge-discharge power limit, and the grid scheduling instruction. The reference energy storage scheduling strategy is generated based on model predictive control (MPC) or dynamic programming, discretizing the continuous-time problem into multiple time periods and solving the optimal charge-discharge power at each decision point. The strategy output includes the charge-discharge time series curve and power distribution ratio of the energy storage, and a certain reserve capacity needs to be reserved to cope with prediction uncertainties. This step provides the energy boundary for device scheduling in Step S3, and its scheduling period and device response speed need to be coordinated to avoid instruction conflicts.

[0021] Step S3: The parameters of energy-consuming devices include rated power, start-stop characteristics, adjustable margin, and priority tags, which are collected through Internet of Things protocols or energy management systems. The device scheduling control decomposes the reference energy storage scheduling strategy into action instructions for specific devices, such as adjusting the air-conditioning temperature setpoint, limiting the power of electric vehicle chargers, or optimizing the timing of industrial production lines. The implementation process adopts a hierarchical control architecture: the upper layer generates the start-stop plans for device groups through mixed-integer linear programming (MILP) or heuristic algorithms, and the lower layer uses PID or fuzzy control to achieve precise power tracking. The device-level control instruction set needs to embed a delay compensation mechanism to avoid grid impact caused by simultaneous responses of multiple devices. The instruction verification link evaluates power quality indicators such as voltage fluctuations and harmonic distortions through digital twin simulation, and introduces demand response signals for dynamic correction when necessary.

[0022] Step S4: The grid-connected / off-grid mode collaborative analysis realizes seamless switching based on grid state monitoring (such as voltage frequency, island detection signal). In the grid-connected mode, the energy storage system follows the grid dispatching instructions to participate in frequency regulation and peak shaving, and in the off-grid mode, it switches to voltage source control (V / f mode) to maintain the stability of the microgrid. The dynamic power generation prediction value is used to predict the mode switching time, for example, starting the diesel generator standby in advance when the irradiance drops suddenly. The reference energy storage scheduling strategy focuses on economic optimization during grid connection and switches to reliability priority during off-grid operation, and reconstructs the optimization goal through dynamic weight adjustment. The device-level control instruction set triggers emergency plans during mode switching, such as hierarchically unloading non-critical loads or activating the energy storage constant voltage support. The final output of the energy storage global control strategy is a multi-time scale control matrix, including millisecond-level inverter modulation signals, minute-level SOC balancing strategies, and hour-level energy trading plans. The closed-loop feedback of each step is realized through the digital twin platform, and the interaction and iteration of real-time data and prediction models improve the system's adaptive ability.

[0023] A control method for a photovoltaic energy storage system provided by the present invention can more accurately reflect the real-time power generation capacity of the photovoltaic system through dynamic irradiance regional power generation prediction, reduce prediction errors, provide a reliable basis for energy storage scheduling, and thus improve the overall energy utilization efficiency of the system. Combining the state of charge and load demand data, dynamically adjusting the energy storage charge and discharge strategy to avoid overcharging or over-discharging problems, improve the service life of the energy storage system, enhance the grid supply-demand matching ability at the same time, and reduce the light curtailment rate. Generating device-level control instructions based on the parameters of energy-consuming devices to achieve precise load management, optimize the operation efficiency of devices, reduce unnecessary energy waste, and improve the economy of the system. Through the collaborative analysis of grid-connected and off-grid modes, ensure that the system can operate stably under different working conditions, improve power supply reliability, reduce dependence on the grid at the same time, and enhance the system's self-regulation ability. Integrating power generation prediction, energy storage scheduling, and device control to achieve multi-level collaborative optimization, while ensuring power supply stability, improve the economy and sustainability of the system.

[0024] In one embodiment, obtaining real-time irradiance data of a photovoltaic energy storage system, performing irradiance area power generation prediction, and obtaining a dynamic power generation power prediction value, including: The photovoltaic power generation array of the photovoltaic energy storage system is installed with a distributed irradiance sensor array, and the sensors collect real-time irradiance data at a fixed sampling frequency. The real-time irradiance data refers to the value of the solar radiation intensity received on the surface of the photovoltaic panel, with the unit of watts per square meter. The system converts the collected discrete irradiance data into continuous spatial distribution data through a spatial interpolation algorithm and divides it into uniform grid cells, and each grid cell corresponds to a specific physical area of the photovoltaic power generation array. This process generates an irradiance distribution matrix, and the row and column indices of the matrix are mapped one-to-one with the physical positions of the photovoltaic array, and the matrix element value is the average irradiance of the corresponding grid cell. The grid resolution of the irradiance distribution matrix is jointly determined by the physical size of the photovoltaic array and the sensor density, and the preset rule requires that the grid side length does not exceed the size of a single photovoltaic module.

[0025] Call the historical meteorological database, which stores meteorological data such as cloud cover, cloud height, wind speed and direction over a past period of time. By comparing the current irradiance distribution matrix with the cloud movement patterns under similar meteorological conditions in the historical meteorological database, the system uses an optical flow algorithm or particle image velocimetry technology to predict the movement trajectory of the clouds. The dynamic irradiance correction coefficient is a two-dimensional matrix related to time, and its dimension is the same as that of the irradiance distribution matrix. Each element value represents the proportion of irradiance attenuation caused by cloud occlusion in the corresponding grid cell in a future period of time. The calculation of the correction coefficient depends on the cloud movement speed and optical thickness, and the preset rule requires that the time window for cloud trajectory prediction does not exceed a few minutes in the future to ensure the prediction accuracy.

[0026] The dynamic irradiance correction coefficient acts on the irradiance distribution matrix, and the corrected irradiance data is obtained through element-by-element multiplication operation. The system converts the corrected irradiance value into the unit power generation efficiency according to the model parameters of the photovoltaic module. The conversion process takes into account the temperature coefficient, conversion efficiency curve and maximum power point tracking characteristics of the photovoltaic module. Each element of the unit power generation power matrix represents the theoretical output power value of the photovoltaic module in the corresponding grid cell under the corrected irradiance. The dimension of the matrix is the same as that of the irradiance distribution matrix, and the unit of the element value is watts. The preset rule requires that the influence of the electrical connection between components be ignored during conversion, and only the single-point theoretical power be reflected.

[0027] Extract target meteorological conditions from the historical meteorological database, including sunshine duration, cloud cover change cycle, and extreme weather event records in typical seasons. The target meteorological conditions are used to establish a statistical model of irradiance-power, with the input of the current unit power generation matrix and the output of the predicted results of the unit power generation in the future time period. The prediction algorithm uses time series analysis or machine learning methods, and the training data comes from the actual power generation records under the historical meteorological conditions of the same period. The predicted results of the unit power generation are a three-dimensional tensor, where the first two dimensions correspond to the spatial grid and the third dimension is the time series, and each value represents the predicted power of a specific grid unit at a specific time point. The preset rule requires that the prediction time resolution matches the cloud movement analysis window.

[0028] Dynamically integrate the predicted results of the unit power generation according to the spatial weights of the irradiance distribution matrix. The weight allocation strategy considers the topological structure of the photovoltaic array, the shadow occlusion effect, and the inverter grouping logic. The dynamic spatial integration reduces the three-dimensional prediction tensor to a one-dimensional dynamic power generation prediction value in the form of a time series through weighted summation. This prediction value represents the overall output power trend of the entire photovoltaic energy storage system in the future time period and is used for subsequent energy storage scheduling control. The preset rule requires that the integration process retains the high-frequency fluctuation characteristics of the time dimension to reflect the impact of cloud transient changes.

[0029] In this embodiment, through the spatial grid processing of real-time irradiance data, the discrete sensor data is converted into a continuous irradiance distribution matrix, enabling the system to accurately capture the irradiance differences in each area of the photovoltaic array and improving the spatial resolution of power generation prediction. Based on the analysis of the cloud movement trajectory of historical meteorological data, the irradiance distribution is dynamically corrected, effectively reducing the impact of cloud occlusion on the prediction accuracy and making the dynamic power generation prediction value more consistent with the actual weather change trend. Through the conversion of unit power generation efficiency and combining the actual performance parameters of photovoltaic modules, it is ensured that the unit power generation matrix accurately reflects the local power generation capacity and avoids prediction deviations caused by the simplification of the theoretical model.

[0030] In one embodiment, obtain the state of charge parameter of the photovoltaic energy storage system and the aggregated power load demand data of the target power consumption area, and conduct energy storage scheduling analysis in combination with the dynamic power generation prediction value to obtain a benchmark energy storage scheduling strategy, including: When the photovoltaic energy storage system starts the control process, the system controller obtains the current state of charge parameter from the energy storage battery management system. The state of charge parameter refers to the percentage value of the current remaining power of the energy storage battery to the rated capacity, reflecting the energy reserve level of the battery. The built-in charge and discharge capacity evaluation algorithm calculates the charge and discharge power constraint range based on the state of charge parameter, battery temperature parameter, and aging coefficient parameter. This range defines the maximum charge power and maximum discharge power allowed for the energy storage system in the current state, ensuring that the battery operates within a safe range. During the evaluation process, when the state of charge approaches the upper threshold, the system automatically reduces the allowed charge power; when the state of charge approaches the lower threshold, the discharge power is correspondingly restricted. The charge and discharge power constraint range serves as the basic limiting condition for subsequent scheduling strategies, ensuring that all scheduling instructions do not exceed the physical tolerance of the battery.

[0031] The prediction module equipped in the photovoltaic power generation unit outputs the predicted value of the dynamic power generation power, which is comprehensively calculated based on meteorological forecast data, historical power generation curves, and real-time irradiance measurement values, reflecting the possible output power of the photovoltaic array within a specific future time period. The intelligent meter cluster in the target power consumption area uploads the aggregated power load demand data in real time, summarizing the total power demand of all electrical equipment in the area. The system aligns the predicted value of the dynamic power generation power with the aggregated power load demand data at the same time resolution and performs real-time difference calculations point by point to generate a real-time net power difference sequence. A positive value indicates that the power generation is greater than the power consumption demand, and a negative value indicates a power deficit. This sequence accurately depicts the dynamic changes in the power supply and demand relationship, providing a quantitative basis for subsequent power distribution.

[0032] The digital signal processing technology is used to perform frequency-domain decomposition on the real-time net power difference sequence. The original signal is separated into a basic power component and an adjustable power component through a multi-level filter bank. The basic power component contains low-frequency fluctuation components, reflecting the long-term trend of power demand and the slow changes in photovoltaic power generation, and has high predictability and stability. The adjustable power component contains high-frequency fluctuation components, corresponding to the instantaneous changes in the load and the rapid fluctuations in photovoltaic output, and has greater randomness. During the decomposition process, the filter characteristics are automatically adjusted according to the preset cut-off frequency parameter to ensure that the two types of components can be effectively separated while maintaining the energy conservation of the original signal. This decomposition method lays the foundation for matching the most suitable processing strategies for different types of power fluctuations.

[0033] The scheduling engine constructs an initial charge-discharge instruction set based on the basic power component and the charge-discharge power constraint range. The processing logic follows the principle of "setting the tone with the basic power". When the basic power component is positive, a charging instruction is generated, and when it is negative, a discharging instruction is generated. The amplitude of the instruction takes the minimum value between the absolute value of the basic power component and the charge-discharge power constraint range to ensure that the instruction is executable and does not violate the battery safety limit. The initial charge-discharge instruction set forms the basic framework for the energy storage system to respond to the main trend of power supply and demand. Its time resolution is consistent with the net power difference sequence and covers the entire scheduling period. This instruction set has considered the continuous power supply capacity of the energy storage system to avoid the state of charge exceeding the reasonable range due to long-term unidirectional charge and discharge.

[0034] The optimization module receives the initial charge-discharge instruction set and the adjustable power component, and uses a multi-objective optimization algorithm to generate an optimized charge-discharge scheduling sequence. The optimization process takes into account the grid stability index, the life loss coefficient of the energy storage system, and the operation economy parameter at the same time. On the premise of meeting the charge-discharge power constraint range, the adjustable power component is processed most reasonably. The algorithm automatically identifies the key features of power fluctuations, adopts a prediction compensation strategy for the regular components in high-frequency fluctuations, and a smoothing filtering strategy for the random components.

[0035] The configured scheduling safety thresholds include multiple safety parameters such as the power change rate limit, the state of charge protection band, and the equipment overload coefficient. The boundary verification module compares the optimized charge-discharge scheduling sequence with these thresholds item by item and corrects the instructions that may cause safety risks. The power change rate verification ensures the smooth transition of the charge-discharge power between adjacent time periods and prevents equipment impact caused by power mutation; the state of charge prediction verification calculates the battery energy state after executing the complete scheduling sequence to avoid overcharge and overdischarge; the equipment capacity verification confirms that all instructions are within the rated working range of the power electronic converter. The instruction set that passes the verification becomes the final reference energy storage scheduling strategy, which not only meets the power supply and demand balance requirements but also ensures the safe and stable operation of all equipment.

[0036] The reference energy storage scheduling strategy is sent to the energy storage converter for execution, and the deviation between the actual execution effect of the strategy and the expected target is monitored in real time. When there is a significant difference between the actual photovoltaic power generation and the predicted value, or when the load demand changes suddenly, the strategy recalculation process is automatically triggered. The new state of charge parameters and the updated prediction data are input into the control algorithm to generate a scheduling strategy adapted to the latest situation. This closed-loop control mechanism ensures that the system always maintains the best performance in a complex operating environment and effectively copes with the challenges of the intermittency of photovoltaic power generation and the randomness of load demand.

[0037] In this embodiment, by obtaining the state of charge parameter of the photovoltaic energy storage system and combining it with the predicted value of the dynamic power generation, energy storage scheduling analysis is carried out, which can evaluate the charge and discharge capacity of the energy storage unit in real time, ensure that the battery operates within a safe range, and avoid overcharging or over-discharging. By calculating the real-time difference between the predicted value of the dynamic power generation and the aggregated power load demand data, the dynamic changes in power supply and demand can be accurately reflected, providing an accurate power balance basis for energy storage scheduling, and improving the response speed and stability of the system. By performing frequency-domain decomposition on the real-time net power difference sequence, the power fluctuation is separated into a basic power component and an adjustable power component, enabling the energy storage system to adopt different scheduling strategies for power fluctuations with different characteristics, and improving the overall regulation accuracy and efficiency.

[0038] In one embodiment, according to the basic power component and the charge and discharge power constraint range, a charge and discharge instruction is constructed to obtain an initial charge and discharge instruction set, including: The basic power component refers to the power curve output by the photovoltaic power generation system under ideal conditions. This component reflects the natural power generation characteristics of the photovoltaic array without energy storage intervention, and is jointly determined by the light intensity, temperature, and environmental conditions. A fixed time window is used to perform time-domain segmentation processing on the basic power component, dividing the continuous time axis into several equal-length or unequal-length time intervals. Each time interval corresponds to a basic power sub-component, which represents the average output power level of the photovoltaic system during that period.

[0039] The basis for time-domain segmentation processing is the fluctuation characteristics of photovoltaic power generation and the grid scheduling requirements. Abrupt changes in lighting conditions usually occur on the minute time scale, so the time interval length is set to the smallest unit that can capture the main change characteristics of photovoltaic output. The processing result generates a sequence of basic power sub-components arranged in chronological order, and each sub-component carries a time stamp mark and a power value attribute. This sequence serves as the input reference for subsequent charge and discharge control, ensuring that the adjustment behavior of the energy storage system is synchronized with the natural fluctuations of photovoltaic power generation.

[0040] The charge and discharge power constraint range defines the maximum charge power and maximum discharge power allowed for the energy storage system in each time interval. This range is jointly determined by the rated power of the energy storage converter, the charge and discharge rate characteristics of the battery body, and the system thermal design margin. When determining the charge and discharge feasibility of each basic power sub-component, the system compares the sub-component value with the upper and lower thresholds of the constraint range.

[0041] The charge-discharge enable flag is a binary state variable indicating whether the energy storage system has the regulation ability within a specific time interval. When the basic power sub-component exceeds the upper limit of the grid demand, if the current remaining charging power capacity is sufficient to absorb the excess, the charge enable flag is set to the effective state. When the basic power sub-component is below the lower limit of the demand, if the remaining discharge power capacity can make up the difference, the discharge enable flag becomes effective. The determination process takes into account the power change rate limit to avoid drastic jumps in the charge-discharge state between adjacent time intervals.

[0042] The output of the feasibility determination is a set of state vectors corresponding to time intervals, and each vector element contains two boolean values, namely the charge enable flag and the discharge enable flag. This vector ensures that the subsequent optimization process only explores solutions within the range allowed by the physical devices, avoiding the generation of unexecutable instructions.

[0043] The linear programming solution link combines the basic power sub-component sequence with the charge-discharge enable flags to construct a multi-objective optimization problem. The objective function considers three dimensions simultaneously: maximizing the photovoltaic power consumption, minimizing the energy storage life loss, and satisfying the grid regulation demand. The decision variables are the charge-discharge power values for each time interval, and the constraint conditions include the power balance equation, the charge-discharge efficiency model of the energy storage system, and the enable flag restrictions generated in the previous step.

[0044] The simplex method or the interior point method is used in the solution process to find the optimal solution in the continuous solution space. The candidate charge-discharge power values for each time interval must simultaneously meet the following requirements: not exceeding the technical limits of the charge-discharge power constraint range, conforming to the state constraints of the charge-discharge enable flags, and maintaining a smooth transition of the power change with adjacent intervals. The optimization result generates a set of candidate charge-discharge power values, which contains several groups of feasible power distribution schemes, and each group of schemes corresponds to different optimization weight configurations.

[0045] Each element in the set of candidate charge-discharge power values is a complete time-series solution covering all segmented intervals. The quality of the solution is measured by the objective function value, and the top N optimal solutions are retained for the next processing stage to ensure the robustness and flexibility of the decision-making. The parameter configuration of the linear programming reflects the operator's preference for different optimization objectives and can be adjusted through the policy matrix.

[0046] The state of charge safety interval defines the upper and lower limits of the battery charge that the energy storage system must maintain throughout the scheduling period. This interval is jointly determined by the battery chemical characteristics, the cycle life model, and the system reserve capacity requirements. When constructing instructions for the set of candidate charge-discharge power values, the system simulates the state of charge trajectories after the execution of each scheme and eliminates the schemes that cause over-limit at any moment.

[0047] The valid candidate solutions enter the instruction refinement stage. According to the format requirements of the power grid dispatching instructions, the system discretizes the continuous power value sequence into an executable control command set. Each command contains clear start time, end time, target power value, and execution priority attributes. Anti-jitter filtering is added to the instruction construction process to eliminate the high-frequency fluctuation components in the power instructions, ensuring that the energy storage converter can execute smoothly.

[0048] The final form of the initial charge-discharge instruction set is a time-power two-dimensional instruction table, where each row in the table corresponds to an executable instruction unit. The instruction set meets the integrity conditions: covering the entire scheduling cycle without time gaps, the power transition between instructions is natural and smooth, and the charge-discharge requirements for all time intervals are met. The system attaches an execution confidence score to the initial instruction set, which reflects the tolerance of the solution to prediction errors and the reserve regulation ability.

[0049] The basic power sub-component sequence, as the result of time-domain segmented processing, maintains the original time resolution and physical dimension when transmitted to the subsequent module. The charge-discharge permission flag vector is stored in binary encoding for easy quick retrieval and logical operations. The set of candidate charge-discharge power values retains the complete optimization path information, supporting solution backtracking and parameter sensitivity analysis.

[0050] After the initial charge-discharge instruction set is generated, a closed-loop verification simulation is carried out. In the verification process, a higher-precision device model and a finer time step are adopted to confirm that the instruction set can maintain system stability under various boundary conditions. The instruction set that passes the verification is marked as the ready state and waits to be issued for execution; the instruction set that fails returns to the reconstruction link, triggering the reconfiguration of the constraint conditions or optimization objectives. The entire processing flow forms a closed-loop optimization structure of design-verification-feedback, ensuring the reliability and adaptability of the control method.

[0051] In this embodiment, by performing time-domain segmented processing on the basic power component, the continuous power curve of photovoltaic power generation is divided into basic power sub-components in multiple time intervals, enabling the charge-discharge control of the energy storage system to accurately match the fluctuation characteristics of photovoltaic output, improving the timeliness and accuracy of power regulation. Based on the charge-discharge power constraint range, the feasibility of each time interval is determined to generate the charge-discharge permission flag, ensuring that the adjustment behavior of the energy storage system is always within the safe operating range of the equipment, avoiding equipment damage caused by overcharging, over-discharging, or power overlimit, and enhancing the reliability and lifespan of the system. The set of candidate charge-discharge power values is solved by linear programming, comprehensively considering multi-objective optimizations such as photovoltaic accommodation, energy storage lifespan, and grid demand, making the charge-discharge strategy of the energy storage system more reasonable, not only improving the utilization rate of photovoltaic power generation but also reducing battery loss and optimizing the overall operating economy.

[0052] In one embodiment, energy-consuming device parameters of a target power consumption area are obtained, and device scheduling control is performed in combination with a reference energy storage scheduling strategy to obtain a device-level control instruction set, including: The energy-consuming device parameters of the target power consumption area include basic information such as device type, rated power, operating period, and power change pattern. The system extracts these parameters from the device management database to form a structured data set of energy-consuming device parameters. The power response characteristic decomposition process analyzes the operating characteristics of each device, and decomposes the power demand pattern of the device into three parts: steady-state power component, transient power component, and adjustable power component. The steady-state power component reflects the basic energy consumption level of the device under normal operating conditions; the transient power component captures the power fluctuation characteristics when the device starts, stops, or changes operating conditions; the adjustable power component identifies the acceptable power adjustment range and response time of the device under specific conditions. Through this decomposition method, the system generates a device power characteristic set containing various power characteristics, providing a detailed characteristic description for subsequent scheduling strategy matching.

[0053] The reference energy storage scheduling strategy is a set of pre-set operating rules for energy storage devices, including core strategy elements such as charge and discharge period arrangements, power distribution priorities, and energy storage state conversion conditions. The matching process between the device power characteristic set and the reference energy storage scheduling strategy is based on a comprehensive analysis of their characteristics. Each component in the device power characteristic set is compared with the applicable conditions in the reference energy storage scheduling strategy to determine the corresponding relationship between the power demands of various devices and the response capabilities of the energy storage system. The matching process considers the time distribution characteristics, fluctuation amplitude of the device power demand, and the real-time capacity state of the energy storage system to form a device-energy storage association matrix reflecting the association strength between the device and the energy storage. The rows of this matrix represent different energy-consuming devices, the columns represent various scheduling strategies of the energy storage system, and the matrix element values represent the matching degree between a specific device and a specific scheduling strategy. A high matching degree indicates that the device is suitable for power regulation using the corresponding energy storage scheduling strategy.

[0054] The device-energy storage association matrix provides a quantitative basis for multi-device collaborative scheduling. The system performs multi-objective optimization analysis based on the association strength reflected in the matrix, combined with the current photovoltaic power generation output and the state of the energy storage system. The analysis process follows constraint conditions such as device operation priorities, the impact of power regulation on the production process, and the optimal efficiency of the energy storage system. For device-strategy combinations with high matching degrees in the association matrix, the system preferentially considers adopting the corresponding scheduling scheme; for medium matching degrees, the system makes strategy adaptability adjustments; combinations with too low matching degrees are excluded from the feasible solutions. Through this screening and optimization process, the system generates a device scheduling solution set containing multiple feasible scheduling schemes. Each scheme in the solution set details the operating power levels of each device at different times, the charge and discharge arrangements of the energy storage system, and the possible system state changes.

[0055] The device scheduling solution set provides a theoretically optimized solution, while the device-level control instruction set converts these solutions into executable specific operation commands. The system converts each solution in the device scheduling solution set according to the instruction generation rules defined in the benchmark energy storage scheduling strategy. The conversion process takes into account the standardization requirements of the device control interface, the real-time requirements of instruction transmission, and the execution feedback mechanism. The generated device-level control instruction set contains specific control parameters such as power set values for each energy-consuming device, operation period adjustment commands, and working mode switching instructions, and also contains control elements such as charge and discharge power instructions and state conversion commands for the energy storage system. The instruction set is designed with a hierarchical structure, with system-level coordination instructions at the upper layer and device-level execution instructions at the lower layer, ensuring that the response speed and operation accuracy of the control system meet the actual operation requirements.

[0056] After the device-level control instruction set is issued and executed, the actual operating status of each device and the response of the energy storage system are continuously monitored. The monitored data is compared with the expected scheduling effect, and the difference information is fed back to the strategy adjustment module. The power response characteristic decomposition link updates the device power characteristic description according to the actual operation data; the dynamic association matching process optimizes the correspondence between the device and the strategy; the multi-device scheduling analysis corrects and optimizes the algorithm parameters; the instruction construction link adjusts the generation rules of the control parameters. This closed-loop operation mechanism enables the system to adapt to changes in device characteristics, load fluctuations, and the uncertainty of photovoltaic power generation, and maintain the effectiveness and reliability of the scheduling control.

[0057] In this embodiment, by decomposing the power response characteristics of the energy-consuming device parameters, the steady-state, transient, and adjustable power characteristics of the device can be accurately identified, enabling the system to finely match the energy consumption requirements of different devices, improving the accuracy and adaptability of energy storage scheduling. Through the dynamic association and matching of the device power characteristic set and the benchmark energy storage scheduling strategy, a device-energy storage association matrix is formed to ensure that the scheduling strategy of the energy storage system highly matches the actual operating characteristics of the device, avoiding the disconnection between the strategy and the demand, and improving the overall energy efficiency. Based on the multi-device scheduling analysis of the device-energy storage association matrix, considering the device priority, the impact of power regulation, and the energy storage efficiency, an optimized device scheduling solution set is generated to achieve the optimal energy distribution under the coordinated operation of multiple devices and reduce the system operation cost.

[0058] In one embodiment, the grid-connected mode and off-grid mode are analyzed in coordination with the dynamic power generation prediction value, the benchmark energy storage scheduling strategy, and the device-level control instruction set to obtain the global energy storage control strategy, including: The dynamic power generation prediction value characterizes the power change trend of photovoltaic power generation in a future period. The system divides the dynamic power generation prediction value into a grid-connected power reference value and an off-grid power reference value according to a preset power distribution strategy. The grid-connected power reference value represents the power that can be fed into the grid, and the off-grid power reference value represents the power used for local load power supply or energy storage. The power distribution strategy is based on grid dispatching requirements and local load priorities to ensure the maximum consumption of renewable energy in the grid-connected mode and to prioritize the power supply to critical loads in the off-grid mode.

[0059] Monitor the state of charge parameter of the energy storage unit, and the state of charge parameter reflects the percentage of the remaining battery power. The grid-connected power reference value is input to the energy storage management module and is verified in combination with the upper and lower limit thresholds of the state of charge parameter. If the grid-connected power reference value requires charging but the state of charge parameter has reached the upper limit, or requires discharging but the state of charge parameter is lower than the lower limit, the system adjusts the grid-connected power reference value to generate a grid-connected power correction value. The correction rule ensures that the energy storage unit operates within a safe range and avoids overcharging or over-discharging. The grid-connected power correction value is used as the final power instruction allowed to be fed into the grid.

[0060] The device-level control instruction set includes the control logics of the photovoltaic inverter, the energy storage converter, and the load switch. The system simulates the off-grid operation scenario and evaluates the power supply continuity of critical loads according to the off-grid power reference value. The off-grid power supply reliability evaluates the matching degree of photovoltaic power generation, energy storage discharge, and load demand, and generates a load guarantee report. The load guarantee report marks the load types and durations that can be guaranteed, such as the power supply duration of important loads (such as communication devices) and the cut-off priorities of non-critical loads (such as air conditioners).

[0061] Compare the off-grid power reference value with the load demand in the load guarantee report and calculate the power matching degree. If the off-grid power reference value is continuously higher than the total load demand, it is determined that off-grid operation is feasible; if there is a power gap or the energy storage support time is insufficient, it is determined that off-grid operation is infeasible. The power matching degree analysis considers dynamic change factors, such as photovoltaic power generation fluctuations or sudden load increases, and generates a determination result of off-grid operation feasibility. This result is a binary logic output that directly determines whether the system switches to the off-grid mode.

[0062] The grid dispatching instruction includes the power adjustment requirements issued by the superior power system, such as peak shaving and valley filling or frequency support. The system compares the grid-connected power correction value with the target value of the grid dispatching instruction and calculates the power deviation. If the deviation exceeds the allowable range, a grid-connected adjustment instruction is generated. The grid-connected adjustment instruction dynamically adjusts the charge and discharge power of the energy storage unit or the output limit of the photovoltaic inverter to ensure that the power at the grid connection point meets the dispatching requirements.

[0063] Based on the comprehensive off-grid operation feasibility determination result and the grid connection regulation instruction, the reference energy storage scheduling strategy is corrected. The reference energy storage scheduling strategy is a preset charge and discharge plan, such as an economic strategy under time-of-use electricity prices. If the off-grid operation feasibility determination result is feasible and the grid dispatching instruction permits, the system switches to the off-grid mode, and the global energy storage control strategy gives priority to ensuring local loads; if the grid connection regulation instruction requires giving priority to responding to grid demands, the global energy storage control strategy aims to minimize the power deviation. The finally output global energy storage control strategy coordinates the operating states of the photovoltaic, energy storage, and load to ensure the optimal operation of the system in the grid-connected / off-grid mode.

[0064] The grid-connected power reference value, off-grid power reference value, grid-connected power correction value, load guarantee report, off-grid operation feasibility determination result, grid connection regulation instruction, and global energy storage control strategy form a closed-loop control chain. The output of each link serves as the input of the next link to ensure the dynamic adjustment of the strategy. For example, the off-grid operation feasibility determination result triggers the operation mode decision, while the grid connection regulation instruction corrects the charge and discharge behavior of the energy storage in real time. The system adapts to the random changes in photovoltaic power generation and load demands through periodic iterative calculations to achieve high-reliability and high-economy coordinated control.

[0065] In this embodiment, by performing grid-connected / off-grid power distribution calculations on the dynamic power generation prediction value, the system can flexibly adjust the flow direction of photovoltaic power generation, ensure the maximum consumption of renewable energy in the grid-connected mode, give priority to ensuring the power supply of critical loads in the off-grid mode, and improve energy utilization efficiency and power supply reliability. By performing energy storage charge and discharge constraint verification on the grid-connected power reference value through the state of charge (SOC) parameter, the system can dynamically correct the grid-connected power, avoid overcharging or over-discharging of the energy storage unit, extend the battery life, and ensure the safe and stable operation of the energy storage system. By performing off-grid power supply reliability assessment on the device-level control instruction set, the system can accurately identify the power supply requirements of critical loads and perform power matching degree analysis in combination with the off-grid power reference value to ensure the continuous power supply ability of the load in the off-grid mode and enhance the anti-disturbance ability of the system.

[0066] In one embodiment, based on the off-grid operation feasibility determination result and the grid connection regulation instruction, an operation mode decision is made on the reference energy storage scheduling strategy to obtain the global energy storage control strategy, including: During the operation of the photovoltaic energy storage system, by real-time monitoring the grid state (including grid-connected / off-grid mode and grid power supply ability), the state of charge (SOC) of the energy storage system (reflecting the current remaining charge of the energy storage), the photovoltaic power generation (determined based on the light intensity and inverter output), and the load demand (determined by the real-time power consumption of electrical equipment), the off-grid operation feasibility determination result is calculated. This result reflects the independent operation ability of the system when disconnected from the grid, and its numerical range is determined by a multi-factor comprehensive evaluation model set by the system.

[0067] The energy balance analysis layer calculates the real-time power supply-demand matching situation. If the sum of the photovoltaic power generation and the energy storage discharge power is greater than the load demand, the energy balance feasibility index is feasible; otherwise, it is infeasible. This index ensures that the system can still meet the basic electricity demand after disconnecting from the grid.

[0068] The dynamic stability assessment layer analyzes the dynamic response ability of the system in the off-grid state. The charge and discharge capabilities of the energy storage system, the volatility of photovoltaic power generation (calculated from historical irradiance data and weather forecasts), and the load mutation characteristics (modeled based on the start-stop rules of electrical equipment) jointly affect the determination result. If the charge and discharge rate of the energy storage can cover the photovoltaic power fluctuations and the energy storage can maintain voltage stability during load mutations, the dynamic stability constraint index is met; otherwise, it is not met. This index ensures that the system will not collapse due to power mutations in the off-grid mode.

[0069] The continuous operation time prediction layer estimates the sustainable power supply time of the system in the off-grid mode. The energy storage SOC, the predicted photovoltaic power generation, and the changing trend of the load demand (generated based on historical electricity consumption curves and prediction algorithms) jointly determine the determination result. If the remaining energy of the energy storage combined with the predicted photovoltaic power generation can support the load to run until the preset time threshold, the time scale assessment index is up to standard; otherwise, it is not up to standard. This index ensures that the system can maintain an adequate operation time after disconnecting from the grid.

[0070] The comprehensive decision-making layer fuses and calculates the energy balance feasibility index, the dynamic stability constraint index, and the time scale assessment index according to the preset weights. If all three indexes meet the conditions, the off-grid operation feasibility determination result is feasible; if any one does not meet, the determination result is infeasible. The weight distribution is set based on the priority of the system's safe operation. For example, the weight of dynamic stability is usually higher than that of time scale assessment to ensure the stable operation of the system.

[0071] The preset off-grid operation feasibility threshold is the critical value set by the system to judge whether the off-grid operation conditions are met. When the off-grid operation feasibility determination result is higher than the threshold, the mode switching trigger flag is set to high level, indicating that the system has the ability to operate off-grid; otherwise, it is set to low level, indicating that the system still needs to rely on the grid support. This threshold is set according to the system's fault tolerance requirements. For example, the threshold is lower when short-term off-grid operation is allowed, and higher when long-term stable operation is required.

[0072] The mode switching trigger flag serves as the trigger signal for subsequent operation mode decisions and directly affects the priority division of grid connection adjustment commands. When the flag is at high level, the system can switch to the off-grid mode and adjust the energy storage charge and discharge strategy to maintain stable operation; when the flag is at low level, the system maintains grid-connected operation, preferentially relying on grid power supply, and simultaneously optimizes the energy storage charge and discharge to cope with possible grid fluctuations.

[0073] The grid-connected regulation command is issued by the power grid dispatching system and includes power regulation requirements, frequency support requirements, or voltage regulation commands. The state of the mode switching trigger flag determines the priority of the grid-connected regulation command. When the mode switching trigger flag is at a high level, the system gives priority to ensuring the off-grid operation ability, the grid-connected regulation weight coefficient decreases, and the response degree to the power grid dispatching command weakens; when the mode switching trigger flag is at a low level, the grid-connected regulation weight coefficient increases, and the system gives priority to responding to the power grid dispatching demand. The dynamic priority division ensures the smooth transition of the system between the grid-connected and off-grid modes and avoids instability caused by frequent switching. The calculation of the grid-connected regulation weight coefficient is dynamically adjusted based on the preset weight distribution rules, combined with the urgency of the power grid dispatching command and the system operation state.

[0074] The grid-connected regulation weight coefficient and the off-grid power reference value are jointly input into the power distribution optimization module. The off-grid power reference value is comprehensively calculated from load forecasting, photovoltaic power generation forecasting, and the remaining capacity of the energy storage system, representing the power balance target that the system needs to maintain in the off-grid mode. The power distribution optimization calculation uses a multi-objective optimization method to ensure the achievement of the off-grid power reference value while meeting the grid-connected regulation requirements. The optimization objectives include minimizing the grid dependence, maximizing the life of the energy storage system, and ensuring the reliability of load power supply. The mode coordination power command, as the optimization result, specifies the charge and discharge power requirements of the energy storage system and the coordination ratio of the grid-connected / off-grid modes.

[0075] The reference energy storage dispatching strategy is the default charge and discharge plan of the system, formulated based on historical data and typical operating conditions. The mode coordination power command is used as the basis for real-time correction to dynamically adjust the power distribution parameters in the reference strategy. The correction process uses a rolling optimization method to update the strategy at regular intervals to ensure its adaptation to the real-time changes of the system state. The output of the energy storage charge and discharge optimization strategy includes detailed charge and discharge times, powers, and mode switching logics, while taking into account economy and reliability.

[0076] The load guarantee report is generated by the system's real-time monitoring and includes the power supply requirements, spare capacity, and potential failure risks of key loads. The energy storage charge and discharge optimization strategy needs to pass the off-grid power supply margin test to ensure that it can still maintain the power supply to key loads when disconnected from the grid. The test process simulates the off-grid operation scenario to evaluate the continuous power supply ability of the energy storage, the adaptability to the fluctuations of photovoltaic power generation, and the ability to respond to load mutations. The off-grid support ability information, as the test result, quantifies the longest sustainable time and power support intensity of the system in the off-grid mode. If the test fails, it triggers a strategy backtracking to readjust the power distribution or limit non-critical loads.

[0077] The off-grid support capability information and mode collaborative power command input global optimization module makes the final decision on the operation mode. The optimization objectives include the stability of mode switching, the life loss of the energy storage system, and the economy of grid interaction. The global optimization adopts a hierarchical decision-making mechanism, giving priority to ensuring the power supply safety in the off-grid mode and then optimizing the dispatching benefits in the grid-connected mode. The energy storage global control strategy, as the final output, includes charge and discharge power commands, mode switching timings, and fault response plans, and is sent to the energy storage converter and photovoltaic inverter for execution.

[0078] In this embodiment, through the binary processing of the off-grid operation feasibility determination result, the system can quickly judge whether it has the off-grid operation ability and generate a mode switching trigger flag to ensure the accuracy and timeliness of the operation mode switching and avoid power supply instability caused by misjudgment. By dynamically dividing the grid-connected regulation commands according to priorities, the system can flexibly adjust the response degree of the grid dispatching commands according to the off-grid operation requirements, taking into account the needs of grid auxiliary services while ensuring the reliability of off-grid power supply, and improving the operation adaptability and economy of the system. By generating mode collaborative power commands through power distribution optimization calculations, the system can achieve optimal power distribution between the grid-connected and off-grid modes, meeting both the grid regulation requirements and ensuring power balance in the off-grid mode, and improving the comprehensive operation efficiency of the energy storage system. By dynamically correcting and updating the benchmark energy storage dispatching strategy, the system can optimize the charge and discharge plan in real time, adapt to the fluctuations of photovoltaic power generation and the changes in load demand, extend the life of the energy storage system, and improve the power supply reliability. Through the off-grid power supply margin test, the system can evaluate the off-grid operation ability in advance, ensure the continuous power supply of critical loads, reduce the risk of sudden power outages, and improve the safety and robustness of the system.

[0079] In one embodiment, it further includes analyzing the power supply requirements of the energy-consuming equipment parameters based on the device-level control instruction set to obtain an emergency power supply demand sequence, and combining the energy storage global control strategy to perform load priority reallocation on the photovoltaic energy storage system to obtain an emergency control plan: Parse the device-level control instruction set and map the device status to extract parameters such as device type, rated power, and operating status, and generate a device operation priority matrix in combination with the preset priority rules. The device operation priority matrix is stored in the form of a two-dimensional table, where the rows represent the device numbers and the columns contain fields such as priority weight, minimum power supply duration, and interruptible flag. The priority weight is dynamically adjusted according to the device type, with the weight of medical equipment higher than that of lighting equipment, and the weight of production equipment being classified according to the process criticality.

[0080] The device operation priority matrix is input into the power continuity analysis module, which selects devices with weights above a threshold and cannot be interrupted, generating a target load identification set. This target load identification set records the device number, current power demand, and minimum power supply time limit, excluding deferrable or non-critical loads. Redundancy checks are incorporated into the analysis process to prevent misidentification of critical loads due to sensor errors.

[0081] Dynamic power generation forecasts are generated by combining historical PV array output data with a weather correction model. These forecasts are then combined with the power requirements of the target load identifier set to produce an emergency power demand sequence. This sequence is divided into time periods, with the power shortfall value and duration noted for each period. A buffer factor is incorporated into the calculation to compensate for instantaneous power fluctuations caused by forecast errors.

[0082] The global energy storage control strategy adaptively modifies the emergency power demand sequence. This modification logic includes energy storage unit charge and discharge efficiency correction, SOC safety range limits, and multi-period power smoothing. This modified demand sequence adjusts the temporal and spatial distribution of the original shortfall to avoid over-discharge or frequent switching of energy storage units. The strategy incorporates a built-in rule library, enabling the selection of different correction coefficients for scenarios such as peak shaving and valley filling, and islanding operation.

[0083] The modified demand sequence drives the load and energy storage emergency allocation module, which matches the real-time capacity and output capabilities of each unit in the PV energy storage system to generate an emergency control plan. This plan includes a storage unit schedule, load switching sequences, and PV inverter limiter instructions. The allocation process adheres to priority weights and time constraints, prioritizing high-weight loads while delaying or reducing the capacity of low-weight loads based on time tolerances. Conflict detection is performed before the plan is output to ensure that no equipment is overloaded or that energy storage backfeeds.

[0084] This embodiment, through the coordinated processing of device-level control instruction sets and the global energy storage control strategy, enables the system to quickly identify critical loads and generate emergency power supply plans, significantly improving the responsiveness and reliability of the PV energy storage system during sudden power outages or power fluctuations. Dynamic weight allocation within the device operation priority matrix ensures continuous power supply to high-priority equipment such as medical and production equipment, preventing system crashes or equipment damage caused by load dislocation. The power supply continuity analysis module, combined with the target load identification set, accurately screens non-interruptible loads, reducing the energy storage resource footprint of non-essential loads and extending the power supply duration of critical equipment. The dynamic power generation forecast and the linked calculation of the emergency power demand sequence effectively mitigate the impact of PV output fluctuations on system stability and reduce the risk of power outages caused by forecast errors. The global energy storage control strategy adaptively adjusts the demand sequence, optimizing the charge and discharge efficiency and lifespan of the energy storage units and avoiding capacity degradation caused by over-discharge or frequent switching. The resulting emergency control plan, through priority-driven load distribution, achieves efficient utilization of energy storage resources, safeguarding core loads while maximizing the overall system power supply capacity.

[0085] Referring to Figure 2 as shown, the present invention further provides a control system for a photovoltaic energy storage system, which is applied to the control method of the photovoltaic energy storage system in any one of the above, and includes: An acquisition module, which is used to acquire the real-time irradiance data of the photovoltaic energy storage system, perform irradiance area power generation prediction, and obtain a dynamic power generation power prediction value; An analysis module, which is used to acquire the state of charge parameter of the photovoltaic energy storage system and the aggregated power load demand data of the target power consumption area, and perform energy storage scheduling analysis in combination with the dynamic power generation power prediction value to obtain a reference energy storage scheduling strategy; An association module, which is used to acquire the parameters of the energy-consuming equipment in the target power consumption area, perform equipment scheduling control with the reference energy storage scheduling strategy, and obtain an equipment-level control instruction set; A processing module, which is used to perform grid-connected mode and off-grid mode collaborative analysis on the dynamic power generation power prediction value, the reference energy storage scheduling strategy, and the equipment-level control instruction set to obtain a global energy storage control strategy.

[0086] The control system for a photovoltaic energy storage system provided by the present invention can more accurately reflect the real-time power generation capacity of the photovoltaic system through dynamic irradiance area power generation prediction, reduce prediction errors, provide a reliable basis for energy storage scheduling, and thus improve the overall energy utilization efficiency of the system. Combining the state of charge and load demand data, dynamically adjusting the energy storage charge and discharge strategy, avoiding overcharging or over-discharging problems, improving the service life of the energy storage system, enhancing the grid supply-demand matching ability at the same time, and reducing the light curtailment rate. Generating equipment-level control instructions based on the parameters of energy-consuming equipment, realizing precise load management, optimizing the operation efficiency of equipment, reducing unnecessary energy waste, and enhancing the economy of the system. Through the collaborative analysis of grid-connected and off-grid modes, ensuring that the system can operate stably under different working conditions, improving power supply reliability, reducing dependence on the grid at the same time, and enhancing the autonomous regulation ability of the system. Integrating power generation prediction, energy storage scheduling, and equipment control, realizing multi-level collaborative optimization, while ensuring stable power supply, improving the economy and sustainability of the system.

[0087] It should be noted that those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described system and each module can refer to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0088] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied to other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A control method for a photovoltaic energy storage system, characterized in that, Including: Obtain the real-time irradiance data of the photovoltaic energy storage system, conduct irradiance area power generation prediction, and obtain the dynamic power generation power prediction value; Obtain the state of charge parameter of the photovoltaic energy storage system and the aggregated power load demand data of the target power consumption area, and conduct energy storage scheduling analysis in combination with the dynamic power generation power prediction value to obtain the benchmark energy storage scheduling strategy; Obtain the energy-consuming equipment parameters of the target power consumption area, perform equipment scheduling control with the benchmark energy storage scheduling strategy, and obtain the equipment-level control instruction set; Conduct grid-connected mode and off-grid mode collaborative analysis on the dynamic power generation power prediction value, the benchmark energy storage scheduling strategy, and the equipment-level control instruction set to obtain the global energy storage control strategy.

2. The control method of the photovoltaic energy storage system according to claim 1, wherein The obtaining the real-time irradiance data of the photovoltaic energy storage system, conducting irradiance area power generation prediction, and obtaining the dynamic power generation power prediction value includes: Obtain the real-time irradiance data of the photovoltaic power generation array of the photovoltaic energy storage system, and perform spatial grid processing on the real-time irradiance data to obtain an irradiance distribution matrix; Conduct cloud movement trajectory analysis on the irradiance distribution matrix according to a preset historical meteorological database to obtain a dynamic irradiance correction coefficient; Convert the unit power generation efficiency of the irradiance distribution matrix according to the dynamic irradiance correction coefficient to obtain a unit power generation power matrix; Obtain the target meteorological conditions of the historical meteorological database, and perform unit power prediction on the unit power generation power matrix to obtain a unit power generation power prediction result; Perform dynamic spatial integration on the unit power generation power prediction result according to the irradiance distribution matrix to obtain the dynamic power generation power prediction value.

3. The control method of the photovoltaic energy storage system according to claim 1, characterized in that The obtaining the state of charge parameter of the photovoltaic energy storage system and the aggregated power load demand data of the target power consumption area, and conducting energy storage scheduling analysis in combination with the dynamic power generation power prediction value to obtain the benchmark energy storage scheduling strategy includes: Perform charge and discharge capacity evaluation processing on the state of charge parameter to obtain a charge and discharge power constraint range; Perform real-time difference calculation on the dynamic power generation power prediction value and the aggregated power load demand data to obtain a real-time net power difference sequence; Perform frequency domain decomposition on the real-time net power difference sequence to obtain a basic power component and an adjustable power component; Construct charge and discharge instructions according to the basic power component and the charge and discharge power constraint range to obtain an initial charge and discharge instruction set; Perform collaborative optimization processing on the initial charge and discharge instruction set and the adjustable power component to obtain an optimized charge and discharge scheduling sequence; Perform boundary verification processing on the optimized charge and discharge scheduling sequence according to a preset scheduling safety threshold to obtain the benchmark energy storage scheduling strategy.

4. The control method of the photovoltaic energy storage system according to claim 3, characterized in that The constructing charge and discharge instructions according to the basic power component and the charge and discharge power constraint range to obtain an initial charge and discharge instruction set includes: Perform time domain segmentation processing on the basic power component to obtain basic power sub-components in multiple continuous time intervals; Perform charge and discharge feasibility determination on each basic power sub-component according to the charge and discharge power constraint range to obtain charge and discharge permission flag bits for each time interval; Perform linear programming on the charge-discharge enable flag bit according to the basic power sub-component and the charge-discharge power constraint range to obtain a set of candidate charge-discharge power values; Construct instructions for the set of candidate charge-discharge power values according to the preset state of charge safety interval to obtain the initial charge-discharge instruction set.

5. The control method of the photovoltaic energy storage system according to claim 1, wherein The obtaining the energy-consuming equipment parameters of the target power consumption area and performing equipment scheduling control with the reference energy storage scheduling strategy to obtain an equipment-level control instruction set, including: Decompose the power response characteristics of the energy-consuming equipment parameters to obtain a set of equipment power characteristics; Perform dynamic association matching on the reference energy storage scheduling strategy according to the set of equipment power characteristics to obtain an equipment energy storage association matrix; Perform multi-device scheduling analysis based on the equipment energy storage association matrix to obtain an equipment scheduling solution set; Construct equipment-level instructions for the equipment scheduling solution set according to the reference energy storage scheduling strategy to obtain the equipment-level control instruction set.

6. The control method of the photovoltaic energy storage system according to claim 1, characterized in that, The performing collaborative analysis on the dynamic power generation prediction value, the reference energy storage scheduling strategy and the equipment-level control instruction set for grid-connected mode and off-grid mode to obtain an energy storage global control strategy, including: Perform grid-connected and off-grid power distribution calculation on the dynamic power generation prediction value to obtain a grid-connected power reference value and an off-grid power reference value; Check the charge-discharge constraint of the grid-connected power reference value according to the state of charge parameter to obtain a grid-connected power correction value; Evaluate the reliability of off-grid power supply for the equipment-level control instruction set to obtain a load guarantee report; Perform power matching degree analysis according to the off-grid power reference value and the load guarantee report to obtain an off-grid operation feasibility determination result; Calculate the power deviation of the grid-connected power correction value according to the preset grid dispatching instruction to obtain a grid-connected regulation instruction; Make an operation mode decision on the reference energy storage scheduling strategy according to the off-grid operation feasibility determination result and the grid-connected regulation instruction to obtain the energy storage global control strategy.

7. The control method of the photovoltaic energy storage system according to claim 6, wherein The making an operation mode decision on the reference energy storage scheduling strategy according to the off-grid operation feasibility determination result and the grid-connected regulation instruction to obtain the energy storage global control strategy, including: Perform binary processing on the off-grid operation feasibility determination result according to the preset off-grid operation feasibility threshold to obtain a mode switching trigger flag; Perform dynamic priority division on the grid-connected regulation instruction according to the mode switching trigger flag to obtain a grid-connected regulation weight coefficient; Perform power distribution optimization calculation on the grid-connected regulation weight coefficient and the off-grid power reference value to obtain a mode collaborative power instruction; Perform dynamic correction and update on the reference energy storage scheduling strategy according to the mode collaborative power instruction to obtain an energy storage charge-discharge optimization strategy; Check the off-grid power supply margin of the energy storage charge-discharge optimization strategy according to the load guarantee report to obtain off-grid support ability information; Perform global optimization of the operation mode on the energy storage charge-discharge optimization strategy according to the off-grid support ability information and the mode collaborative power instruction to obtain the energy storage global control strategy.

8. The control method of the photovoltaic energy storage system according to claim 1, wherein It further includes analyzing the power supply demand of the energy-consuming equipment parameters based on the device-level control instruction set to obtain an emergency power supply demand sequence, and combining the global energy storage control strategy to perform load priority reallocation on the photovoltaic energy storage system to obtain an emergency control plan: Parsing the device-level control instruction set and mapping the device status to obtain a device operation priority matrix; Performing power supply continuity analysis on the energy-consuming equipment parameters according to the device operation priority matrix to obtain a target load identification set; Calculating the supply-demand gap for the target load identification set based on the dynamic power generation prediction value to obtain an emergency power supply demand sequence; Performing mode adaptability correction processing on the emergency power supply demand sequence according to the global energy storage control strategy to obtain a corrected demand sequence; Performing load energy storage emergency allocation on the photovoltaic energy storage system according to the corrected demand sequence and the global energy storage control strategy to obtain the emergency control plan.

9. A control system for a photovoltaic energy storage system, characterized in that, A control method applied to the photovoltaic energy storage system according to any one of claims 1-8 above, including: An acquisition module, which is used to obtain the real-time irradiance data of the photovoltaic energy storage system, perform irradiance area power generation prediction, and obtain a dynamic power generation prediction value; An analysis module, which is used to obtain the state of charge parameter of the photovoltaic energy storage system and the aggregated power load demand data of the target power consumption area, and perform energy storage scheduling analysis in combination with the dynamic power generation prediction value to obtain a reference energy storage scheduling strategy; An association module, which is used to obtain the energy-consuming equipment parameters of the target power consumption area, perform device scheduling control with the reference energy storage scheduling strategy, and obtain a device-level control instruction set; A processing module, which is used to perform grid-connected mode and off-grid mode collaborative analysis on the dynamic power generation prediction value, the reference energy storage scheduling strategy, and the device-level control instruction set to obtain a global energy storage control strategy.

Images